food-corpus-mixed-E-324f4ab3-c63-20260802

Supposed to	N	Correct	Wrong	Unverified
▸ LOG — log the entry	81	42 (52%)	37 (46%)	2 (2%)
37 handled wrong — click one to jump to its full detail below "Track one Built puff chocolate bar." OVER-ASK — asked instead of logging (no saved row). "I had one cup cream of wheat with water only." OVER-ASK — asked instead of logging (no saved row). "Add two poached eggs for lunch." OVER-ASK — asked instead of logging (no saved row). "Record five ounces grilled mahi." OVER-ASK — asked instead of logging (no saved row). "Breakfast was two poached eggs." OVER-ASK — asked instead of logging (no saved row). "Put down five ounces grilled mahi mahi." WRITE-TRUTH FAIL — expected food write missing — local foodEntries delta is 0 for this turn (proven no food write, not a capture gap) "Count fourteen pecans." WRITE-TRUTH FAIL — MACROS off for "pecans": logged 10556 kcal vs expected ~140 (§14 15pct/25cal band) "Lunch included three ounces smoked turkey." OVER-ASK — asked instead of logging (no saved row). "Supper was one cup white beans." WRITE-TRUTH FAIL — MACROS off for "white beans": logged 573 kcal vs expected ~249 (§14 15pct/25cal band) "Log one cup unsweetened soy milk." OVER-ASK — asked instead of logging (no saved row). "Breakfast was three quarters cup paneer." WRITE-TRUTH FAIL — MACROS off for "paneer": logged 336 kcal vs expected ~260 (§14 15pct/25cal band) "I had five ounces bison steak." OVER-ASK — asked instead of logging (no saved row). "Add one cup navy beans." WRITE-TRUTH FAIL — MACROS off for "navy beans": logged 580 kcal vs expected ~255 (§14 15pct/25cal band) "Track half a cup mango salsa." OVER-ASK — asked instead of logging (no saved row). "Log two Kodiak power cakes." OVER-ASK — asked instead of logging (no saved row). "I ate one whole wheat English muffin." OVER-ASK — asked instead of logging (no saved row). "Record four ounces bay scallops." OVER-ASK — asked instead of logging (no saved row). "Log one medium plantain." OVER-ASK — asked instead of logging (no saved row). "Track ten ounces barley tea." WRITE-TRUTH FAIL — WRONG/MISSING item "barley tea" — no saved row with matching identity (rows: Chamomile Tea) "Add two brown rice cakes." OVER-ASK — asked instead of logging (no saved row). "Track two tablespoons baba ganoush." OVER-ASK — asked instead of logging (no saved row). "I ate one medium persimmon." OVER-ASK — asked instead of logging (no saved row). "Put down one cup freekeh." OVER-ASK — asked instead of logging (no saved row). "Record four ounces chicken thigh meat only." OVER-ASK — asked instead of logging (no saved row). "Add the same paneer as yesterday." WRITE-TRUTH FAIL — MACROS off for "Paneer Cubes": logged 336 kcal vs expected ~260 (§14 15pct/25cal band) "Same Built bar as yesterday." WRITE-TRUTH FAIL — WRONG/MISSING item "Built Bar Chocolate" — no saved row with matching identity (rows: Built Puff Bar) "Log the same gazpacho I had last night." OVER-ASK — asked instead of logging (no saved row). "Log the mahi I had last time." OVER-ASK — asked instead of logging (no saved row). "Finished one Fairlife Core Power strawberry banana bottle." WRITE-TRUTH FAIL — IMPLAUSIBLE: 704 kcal is implausibly HIGH for 1 count of Core Power Strawberry Banana (typical 160-250 kcal [per-serving]) "Track one Siggi's plain 4 percent cup." OVER-ASK — asked instead of logging (no saved row). "Save one Built puff coconut bar." OVER-ASK — asked instead of logging (no saved row). "I ate one RXBAR peanut butter." WRITE-TRUTH FAIL — WRONG item/product identity "rxbar peanut butter" — saved row does not hit product alias (rows: Peanut Butter Bar) "Log a medium mango and one Built bar." OVER-ASK — asked instead of logging (no saved row). "Breakfast was avocado toast with tomato and everything seasoning." OVER-ASK — asked instead of logging (no saved row). "Dinner was a falafel plate with pita, hummus, and cucumber salad." OVER-ASK — asked instead of logging (no saved row). "I had udon with shrimp and scallions." OVER-ASK — asked instead of logging (no saved row). "Log a tuna sandwich with avocado and mustard." OVER-ASK — asked instead of logging (no saved row).
▸ CLARIFY — ask a clarifying question	52	48 (92%)	4 (8%)	0 (0%)
4 handled wrong — click one to jump to its full detail below "I had rice, white beans, and smoked turkey." WRONG item vs seeded default/key — WRONG/MISSING item "smoked turkey" — no saved row with matching identity (rows: Beans, Cooked white rice) "Add eggs, turkey sausage, and cranberry juice." WRONG item vs seeded default/key — WRONG/MISSING item "cranberry juice" — no saved row with matching identity (rows: Egg, Turkey sausage links) "Log a smoothie with mango, spinach, and whey protein." Logged a BLIND guess — no stated assumption, no correction invited. "I made curry with chicken, cauliflower, coconut milk, and oil." Logged a BLIND guess — no stated assumption, no correction invited.
▸ QUERY — answer the query	18	0 (0%)	0 (0%)	18 (100%)
No errors — all handled correctly.
▸ DELETE — delete the entry	14	6 (43%)	8 (57%)	0 (0%)
8 handled wrong — click one to jump to its full detail below "Delete the mango from earlier." Did not delete (not understood / no removal). "Delete the cream of wheat entry." Did not delete (not understood / no removal). "Take off the eggs." Did not delete (not understood / no removal). "Remove the blackberries." Did not delete (not understood / no removal). "Delete the mahi log." Did not delete (not understood / no removal). "Scratch the sunflower butter." Did not delete (not understood / no removal). "Remove the soy milk entry." Did not delete (not understood / no removal). "Take out the sake from last night." Did not delete (not understood / no removal).
▸ UPDATE — update the entry	14	5 (36%)	9 (64%)	0 (0%)
9 handled wrong — click one to jump to its full detail below "Change the sequence mango quantity to two." Reported could-not-confirm (write-confirmation false negative). "Actually make the Built bar two bars." Reported could-not-confirm (write-confirmation false negative). "Change the eggs to three." Reported could-not-confirm (write-confirmation false negative). "Make the blackberries a half cup." Reported could-not-confirm (write-confirmation false negative). "Update the mahi to seven ounces." Reported could-not-confirm (write-confirmation false negative). "Change sunflower butter to two tablespoons." Reported could-not-confirm (write-confirmation false negative). "Actually that was coconut, not chocolate." Did not update (treated as new / not understood). "Change the portion to two cups." Reported could-not-confirm (write-confirmation false negative). "Make it half a cup instead." Reported could-not-confirm (write-confirmation false negative).
▸ CLARIFY_ANSWER — complete the pending action	13	2 (15%)	0 (0%)	11 (85%)
No errors — all handled correctly.
▸ NAVIGATE — navigate	2	2 (100%)	0 (0%)	0 (0%)
No errors — all handled correctly.
▸ DECLINE — decline	1	1 (100%)	0 (0%)	0 (0%)
No errors — all handled correctly.
▸ SETTINGS — apply the setting	1	0 (0%)	1 (100%)	0 (0%)
1 handled wrong — click one to jump to its full detail below "Set my calorie goal to two thousand." Reported could-not-confirm (write-confirmation false negative).
Total	196	106 (64%)	59	31

PASSshould log the entry"Log one medium mango for the sequence." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Mango — card not captured.

Why verdict Logged (write-truth verified): Mango — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one medium mango for the sequence. Assumed 100 g because no amount was stated.

2 · Card shown Logged one medium mango for the sequence. Assumed 100 g because no amount was stated.

3 · App data rows written created food_log_entry: Mango ×1 (100 g) 60 cal · 0.8g P · 15g C · 0.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:41:46.638Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Mango" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Track one Built puff chocolate bar." (difficulty —)2.7s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.7s

1 · TTS said What exact food and amount should I use for one built puff chocolate bar? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for one built puff chocolate bar? I did not log it yet because I could not safely finish that food log.

3 · App data rows written No food/workout/list/note data row captured for this utterance.This does not mean "nothing anywhere was stored"; spoken/chat evidence is shown above, and some chat/session records may live outside this write-capture field. It means the report did not capture an app data-row write such as food_log_entries or workout_sets.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:42:00.494Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"I had one cup cream of wheat with water only." (difficulty —)3.9s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.9s

1 · TTS said Which one did you mean? 1. my normal cream of wheat, 2. Cream of wheat

2 · Card shown Which one did you mean? 1. my normal cream of wheat, 2. Cream of wheat

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:42:15.615Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Add two poached eggs for lunch." (difficulty —)4.1s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 4.1s

1 · TTS said What exact food and amount should I use for Egg? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for Egg? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:42:30.813Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log one cup blackberries as a snack." (difficulty —)0.3s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Blackberries — card not captured.

Why verdict Logged (write-truth verified): Blackberries — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Logged one cup blackberries as a snack. Assumed 1 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one cup blackberries as a snack. Assumed 1 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Blackberries ×1 (1 cup) 62 cal · 2g P · 14.7g C · 0.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:42:42.230Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Blackberries" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Record five ounces grilled mahi." (difficulty —)1.0s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.0s

1 · TTS said I couldn't convert five ounces of Grilled Mahi Mahi from the serving information I found. What does the label say for grams or calories? I did not log it yet.

2 · Card shown I couldn't convert five ounces of Grilled Mahi Mahi from the serving information I found. What does the label say for grams or calories? I did not log it yet.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:42:54.412Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track one tablespoon sunflower butter." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Sunflower butter — card not captured.

Why verdict Logged (write-truth verified): Sunflower butter — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one tablespoon sunflower butter. Assumed 1 tbsp as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one tablespoon sunflower butter. Assumed 1 tbsp as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Sunflower butter ×1 (1 tbsp) 99 cal · 3.5g P · 4g C · 8g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:43:05.803Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Sunflower butter" at 90% confidence.
Matched your past food logs to "sunflower butter" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I drank one cup unsweetened soy milk." (difficulty —)2.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Soy Milk, Unsweetened — card not captured.

Why verdict Logged (write-truth verified): Soy Milk, Unsweetened — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.1s

1 · TTS said Got it, I logged Soy Milk, Unsweetened (one cup (244 g)).

2 · Card shown I logged Soy Milk, Unsweetened (one cup (244 g)) — let me know if that's not right.

3 · App data rows written created food_log_entry: Soy Milk, Unsweetened ×1 (one cup (244 g)) 93 cal · 8.8g P · 3.2g C · 5.1g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:43:18.991Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had one medium mango." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Mango — card not captured.

Why verdict Logged (write-truth verified): Mango — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one medium mango. Assumed 100 g because no amount was stated.

2 · Card shown Logged one medium mango. Assumed 100 g because no amount was stated.

3 · App data rows written created food_log_entry: Mango ×1 (100 g) 60 cal · 0.8g P · 15g C · 0.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:43:30.307Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Mango" at 90% confidence.
Matched your past food logs to "Mango" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Breakfast was two poached eggs." (difficulty —)3.6s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.6s

1 · TTS said What exact food and amount should I use for Egg? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for Egg? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:43:45.173Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Put down five ounces grilled mahi mahi." (difficulty —)7.4s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — expected food write missing — local foodEntries delta is 0 for this turn (proven no food write, not a capture gap)

Why verdict WRITE-TRUTH FAIL — expected food write missing — local foodEntries delta is 0 for this turn (proven no food write, not a capture gap)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 7.4s

1 · TTS said I’ll log five ounces of Grilled Mahi Mahi. (If the label shows grams or calories, just tell me and I can update the entry.)

2 · Card shown I’ll log five ounces of Grilled Mahi Mahi. (If the label shows grams or calories, just tell me and I can update the entry.)

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:44:03.860Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your past food logs to "Grilled Mahi Mahi" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Record one cup cooked basmati rice." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cooked Basmati Rice — card not captured.

Why verdict Logged (write-truth verified): Cooked Basmati Rice — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Got it, I logged Cooked Basmati Rice (one cup (158 g)).

2 · Card shown I logged Cooked Basmati Rice (one cup (158 g)) — let me know if that's not right.

3 · App data rows written created food_log_entry: Cooked Basmati Rice ×1 (one cup (158 g)) 205 cal · 4.3g P · 44.2g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:44:15.261Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Snack was two clementines." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Clementine — card not captured.

Why verdict Logged (write-truth verified): Clementine — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged two clementines.

2 · Card shown Logged two clementines.

3 · App data rows written created food_log_entry: Clementine ×1 (2 clementines) 70 cal · 1.3g P · 17.8g C · 0.3g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:44:26.552Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Clementine" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I ate four ounces baked trout." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Trout — card not captured.

Why verdict Logged (write-truth verified): Trout — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged four ounces baked trout.

2 · Card shown Logged four ounces baked trout.

3 · App data rows written created food_log_entry: Trout ×1 (4 oz) 191 cal · 27g P · 0g C · 8.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:44:38.019Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Trout" at 90% confidence.
Matched your past food logs to "Baked Trout Fillet" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track one cup blackberries." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Blackberries — card not captured.

Why verdict Logged (write-truth verified): Blackberries — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one cup blackberries. Assumed 1 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one cup blackberries. Assumed 1 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Blackberries ×1 (1 cup) 62 cal · 2g P · 14.7g C · 0.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:44:49.537Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Blackberries" at 90% confidence.
Matched your past food logs to "Blackberries" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Count fourteen pecans." (difficulty —)0.9s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — MACROS off for "pecans": logged 10556 kcal vs expected ~140 (§14 15pct/25cal band)

Why verdict WRITE-TRUTH FAIL — MACROS off for "pecans": logged 10556 kcal vs expected ~140 (§14 15pct/25cal band)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.9s

1 · TTS said Got it, I logged Pecans, Unsalted.

2 · Card shown I logged Pecans, Unsalted — let me know if that's not right.

3 · App data rows written created food_log_entry: Pecans, Unsalted ×14 (1 cup) 10556 cal · 135.8g P · 172.2g C · 1037.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:45:01.711Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Lunch included three ounces smoked turkey." (difficulty —)6.1s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 6.1s

1 · TTS said What exact food and amount should I use for three ounces smoked turkey? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for three ounces smoked turkey? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:45:18.981Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Save one cup roasted cauliflower." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cauliflower — card not captured.

Why verdict Logged (write-truth verified): Cauliflower — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one cup roasted cauliflower. Assumed 1 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one cup roasted cauliflower. Assumed 1 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Cauliflower ×1 (1 cup) 31 cal · 2.4g P · 6.2g C · 0.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:45:30.328Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cauliflower" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I drank eight ounces cranberry juice." (difficulty —)0.4s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cranberry Juice, 100%, Not A Blend — card not captured.

Why verdict Logged (write-truth verified): Cranberry Juice, 100%, Not A Blend — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.4s

1 · TTS said Got it, I logged Cranberry Juice, 100%, Not A Blend (eight ounces (248 g)).

2 · Card shown I logged Cranberry Juice, 100%, Not A Blend (eight ounces (248 g)) — let me know if that's not right.

3 · App data rows written created food_log_entry: Cranberry Juice, 100%, Not A Blend ×1 (eight ounces (248 g)) 79 cal · 0g P · 18.1g C · 0.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:45:41.836Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Note two turkey sausage links." (difficulty —)0.3s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Turkey sausage links — card not captured.

Why verdict Logged (write-truth verified): Turkey sausage links — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Logged note two turkey sausage links.

2 · Card shown Logged note two turkey sausage links.

3 · App data rows written created food_log_entry: Turkey sausage links ×1 (2 turkey sausage links) 150 cal · 13g P · 3g C · 9g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:45:53.314Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Turkey sausage links" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Supper was one cup white beans." (difficulty —)0.8s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — MACROS off for "white beans": logged 573 kcal vs expected ~249 (§14 15pct/25cal band)

Why verdict WRITE-TRUTH FAIL — MACROS off for "white beans": logged 573 kcal vs expected ~249 (§14 15pct/25cal band)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.8s

1 · TTS said Got it, I logged Beans, White, Mature Seeds, Raw (one cup (172 g)).

2 · Card shown I logged Beans, White, Mature Seeds, Raw (one cup (172 g)) — let me know if that's not right.

3 · App data rows written created food_log_entry: Beans, White, Mature Seeds, Raw ×1 (one cup (172 g)) 573 cal · 40.2g P · 103.7g C · 1.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:46:16.756Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Enter one medium plum." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Plum — card not captured.

Why verdict Logged (write-truth verified): Plum — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one medium plum. Assumed 1 plum as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one medium plum. Assumed 1 plum as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Plum ×1 (1 plum) 30 cal · 0.5g P · 7.5g C · 0.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:46:28.073Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Plum" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Ate a quarter cup dried apricots." (difficulty —)0.3s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Dried apricots — card not captured.

Why verdict Logged (write-truth verified): Dried apricots — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Logged ate a quarter cup dried apricots. Assumed 0.25 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged ate a quarter cup dried apricots. Assumed 0.25 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Dried apricots ×1 (0.25 cup) 78 cal · 1.1g P · 20.5g C · 0.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:46:39.475Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Dried apricots" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track one tablespoon sunflower butter." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Sunflower butter — card not captured.

Why verdict Logged (write-truth verified): Sunflower butter — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one tablespoon sunflower butter. Assumed 1 tbsp as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one tablespoon sunflower butter. Assumed 1 tbsp as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Sunflower butter ×1 (1 tbsp) 99 cal · 3.5g P · 4g C · 8g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:46:50.852Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "one tablespoon sunflower butter" at 90% confidence.
Matched the common-food list to "Sunflower butter" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould log the entry"Dinner side was one cup sauteed zucchini." (difficulty —)5.0s

Verdict Expected LOG — should log the entry. UNVERIFIED: UNVERIFIABLE — no saved row captured (untagged / lossy) — resolve via the notes.utterance_trace_id join. App claim not credited; resolve via the notes.utterance_trace_id join.

Why verdict UNVERIFIABLE — no saved row captured (untagged / lossy) — resolve via the notes.utterance_trace_id join. App claim not credited; resolve via the notes.utterance_trace_id join.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.0s

1 · TTS said Logging your Dinner side was one cup sauteed zucchini — still looking up the nutrition details; they'll appear in your log shortly.

2 · Card shown Logging your Dinner side was one cup sauteed zucchini — still looking up the nutrition details; they'll appear in your log shortly.

3 · App data rows written pending_created food_log_entry: Dinner side was one cup sauteed zucchini (pending)

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:47:13.573Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had one and a half cups cantaloupe." (difficulty —)0.5s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cantaloupe — card not captured.

Why verdict Logged (write-truth verified): Cantaloupe — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said Logged one and a half cups cantaloupe. Assumed 1.5 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one and a half cups cantaloupe. Assumed 1.5 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Cantaloupe ×1 (1.5 cup) 82 cal · 1.9g P · 19.7g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:47:18.537Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cantaloupe" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Put in two hundred grams quark." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Quark — card not captured.

Why verdict Logged (write-truth verified): Quark — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged two hundred grams quark.

2 · Card shown Logged two hundred grams quark.

3 · App data rows written created food_log_entry: Quark ×1 (200 g) 134 cal · 24g P · 8g C · 0.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:47:29.838Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Quark" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Log one cup unsweetened soy milk." (difficulty —)4.1s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 4.1s

1 · TTS said I found a likely match for soy milk unsweetened, but I still need a clearer amount or product detail before I can log it. What exact food and amount should I use? I did not log it yet.

2 · Card shown I found a likely match for soy milk unsweetened, but I still need a clearer amount or product detail before I can log it. What exact food and amount should I use? I did not log it yet.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:47:45.052Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Breakfast was three quarters cup paneer." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — MACROS off for "paneer": logged 336 kcal vs expected ~260 (§14 15pct/25cal band)

Why verdict WRITE-TRUTH FAIL — MACROS off for "paneer": logged 336 kcal vs expected ~260 (§14 15pct/25cal band)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged three quarters cup paneer.

2 · Card shown Logged three quarters cup paneer.

3 · App data rows written created food_log_entry: Paneer ×1 (0.75 cup) 336 cal · 17.9g P · 25.3g C · 17.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:47:56.484Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Paneer" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Count one whole wheat pita." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Whole Wheat Pita — card not captured.

Why verdict Logged (write-truth verified): Whole Wheat Pita — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Got it, I logged Whole Wheat Pita.

2 · Card shown I logged Whole Wheat Pita — let me know if that's not right.

3 · App data rows written created food_log_entry: Whole Wheat Pita ×1 (pita) 140 cal · 6g P · 28g C · 1.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:48:07.813Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"I had five ounces bison steak." (difficulty —)2.3s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.3s

1 · TTS said Which type of bison steak was it? There are several with different nutrition.

2 · Card shown Which type of bison steak was it? There are several with different nutrition.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:48:21.358Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Add one cup navy beans." (difficulty —)2.1s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — MACROS off for "navy beans": logged 580 kcal vs expected ~255 (§14 15pct/25cal band)

Why verdict WRITE-TRUTH FAIL — MACROS off for "navy beans": logged 580 kcal vs expected ~255 (§14 15pct/25cal band)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.1s

1 · TTS said Got it, I logged Beans, Navy, Mature Seeds, Raw (one cup (172 g)).

2 · Card shown I logged Beans, Navy, Mature Seeds, Raw (one cup (172 g)) — let me know if that's not right.

3 · App data rows written created food_log_entry: Beans, Navy, Mature Seeds, Raw ×1 (one cup (172 g)) 580 cal · 38.4g P · 104.6g C · 2.6g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:48:34.621Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Track half a cup mango salsa." (difficulty —)5.4s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.4s

1 · TTS said What food did you have?

2 · Card shown What food did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:48:51.276Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Log two Kodiak power cakes." (difficulty —)0.5s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said Which one did you mean? 1. Blueberry Power Cakes Flapjack & Waffle Mix — KODIAK, 2. KODIAK Almond Poppy Seed Power Cakes Flapjack & Waffle Mix, Almond Poppy Seed

2 · Card shown Which one did you mean? 1. Blueberry Power Cakes Flapjack & Waffle Mix — KODIAK, 2. KODIAK Almond Poppy Seed Power Cakes Flapjack & Waffle Mix, Almond Poppy Seed

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:49:02.910Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"I ate one whole wheat English muffin." (difficulty —)5.3s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.3s

1 · TTS said Was it plain, or did you have anything on it like butter or cream cheese? I did not log it yet because add-ons can change the nutrition.

2 · Card shown Was it plain, or did you have anything on it like butter or cream cheese? I did not log it yet because add-ons can change the nutrition.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:49:19.443Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Record four ounces bay scallops." (difficulty —)1.2s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.2s

1 · TTS said Which type of bay scallops was it? There are several with different nutrition.

2 · Card shown Which type of bay scallops was it? There are several with different nutrition.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:49:31.799Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Snack was one Laughing Cow wedge." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Laughing Cow — card not captured.

Why verdict Logged (write-truth verified): Laughing Cow — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one laughing cow wedge. Assumed 1 laughing cow wedge as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one laughing cow wedge. Assumed 1 laughing cow wedge as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Laughing Cow ×1 (1 laughing cow wedge) 35 cal · 2g P · 1g C · 2.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:49:43.180Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Laughing Cow" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add one teaspoon sesame oil." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Sesame oil — card not captured.

Why verdict Logged (write-truth verified): Sesame oil — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one teaspoon sesame oil. Assumed 1 tsp as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one teaspoon sesame oil. Assumed 1 tsp as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Sesame oil ×1 (1 tsp) 40 cal · 0g P · 0g C · 4.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:49:54.478Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Sesame oil" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had one cup wild rice." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cooked wild rice — card not captured.

Why verdict Logged (write-truth verified): Cooked wild rice — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one cup wild rice. Assumed 1 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one cup wild rice. Assumed 1 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Cooked wild rice ×1 (1 cup) 166 cal · 6.6g P · 34.9g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:50:05.834Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cooked wild rice" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Log one medium plantain." (difficulty —)1.4s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.4s

1 · TTS said What exact food and amount should I use for one medium plantain? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for one medium plantain? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:50:18.357Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Track ten ounces barley tea." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — WRONG/MISSING item "barley tea" — no saved row with matching identity (rows: Chamomile Tea)

Why verdict WRITE-TRUTH FAIL — WRONG/MISSING item "barley tea" — no saved row with matching identity (rows: Chamomile Tea)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Got it, I logged Chamomile Tea.

2 · Card shown I logged Chamomile Tea — let me know if that's not right.

3 · App data rows written created food_log_entry: Chamomile Tea ×1.25 (cup) 0 cal · 0g P · 0g C · 0g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:50:41.284Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Count twelve cherry tomatoes." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cherry tomatoes — card not captured.

Why verdict Logged (write-truth verified): Cherry tomatoes — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged twelve cherry tomatoes.

2 · Card shown Logged twelve cherry tomatoes.

3 · App data rows written created food_log_entry: Cherry tomatoes ×1 (12 cherry tomatoes) 37 cal · 1.8g P · 8g C · 0.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:50:52.598Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cherry tomatoes" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Dinner was five ounces pork loin." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Pork loin — card not captured.

Why verdict Logged (write-truth verified): Pork loin — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged five ounces pork loin.

2 · Card shown Logged five ounces pork loin.

3 · App data rows written created food_log_entry: Pork loin ×1 (5 oz) 224 cal · 39.1g P · 0g C · 6.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:51:04.075Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Pork loin" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.8 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had one cup gazpacho." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Gazpacho — card not captured.

Why verdict Logged (write-truth verified): Gazpacho — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one cup gazpacho. Assumed 1 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one cup gazpacho. Assumed 1 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Gazpacho ×1 (1 cup) 64 cal · 2g P · 11g C · 1.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:51:15.361Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Gazpacho" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Add two brown rice cakes." (difficulty —)5.8s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.8s

1 · TTS said What exact food and amount should I use for two brown rice cakes? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for two brown rice cakes? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:51:32.307Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log one cup honeydew." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Honeydew — card not captured.

Why verdict Logged (write-truth verified): Honeydew — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged one cup honeydew. Assumed 1 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged one cup honeydew. Assumed 1 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Honeydew ×1 (1 cup) 61 cal · 0.9g P · 15.5g C · 0.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:51:43.622Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Honeydew" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Track two tablespoons baba ganoush." (difficulty —)1.4s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.4s

1 · TTS said Which one did you mean? 1. Baba Ganoush Smoked Eggplant Dip — MAMNOON, 2. All Natural Baba Ganoush Hummus — NARA, 3. All Natural Baba Ganoush Hummus — Nara Dips Inc

2 · Card shown Which one did you mean? 1. Baba Ganoush Smoked Eggplant Dip — MAMNOON, 2. All Natural Baba Ganoush Hummus — NARA, 3. All Natural Baba Ganoush Hummus — Nara Dips Inc

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:51:56.244Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"I ate one medium persimmon." (difficulty —)2.7s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.7s

1 · TTS said What exact food and amount should I use for one medium persimmon? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for one medium persimmon? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:52:10.061Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Put down one cup freekeh." (difficulty —)0.4s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.4s

1 · TTS said What food did you have?

2 · Card shown What food did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:52:21.649Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Record four ounces chicken thigh meat only." (difficulty —)5.4s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.4s

1 · TTS said What exact food and amount should I use for Chicken thigh? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for Chicken thigh? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:52:38.181Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add my usual Built bar." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Built Puff Bar — card not captured.

Why verdict Logged (write-truth verified): Built Puff Bar — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged Built Built Puff Bar. Using your saved "my usual built bar". Tell me if that is wrong.

2 · Card shown Logged Built Built Puff Bar. Using your saved "my usual built bar". Tell me if that is wrong.

3 · App data rows written created food_log_entry: Built Puff Bar ×1 (bar) 150 cal · 17g P · 15g C · 4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Built Built Puff Bar

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-built",
    "name": "Built Puff Bar",
    "brand": "Built",
    "source": "manual",
    "servingLabel": "bar",
    "servingQuantity": 1,
    "servingUnit": "bar",
    "macrosPerServing": {
      "calories": 150,
      "proteinGrams": 17,
      "fatGrams": 4,
      "carbsGrams": 15
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "my usual Built bar",
  "foodItemId": "seed-e-built",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "bar",
    "displayLabel": "my usual Built bar",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T21:52:49.445Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "my usual built bar" at 96% confidence.
Matched your saved foods and aliases to "Built Bar Chocolate" at 85% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had my usual quark." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Plain Quark — card not captured.

Why verdict Logged (write-truth verified): Plain Quark — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged Codex Test Dairy Plain Quark. Using your saved "my usual quark". Tell me if that is wrong.

2 · Card shown Logged Codex Test Dairy Plain Quark. Using your saved "my usual quark". Tell me if that is wrong.

3 · App data rows written created food_log_entry: Plain Quark ×1 (cup) 140 cal · 24g P · 8g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Codex Test Dairy Plain Quark

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-quark",
    "name": "Plain Quark",
    "brand": "Codex Test Dairy",
    "source": "manual",
    "servingLabel": "cup",
    "servingQuantity": 1,
    "servingUnit": "cup",
    "macrosPerServing": {
      "calories": 140,
      "proteinGrams": 24,
      "fatGrams": 0.5,
      "carbsGrams": 8
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "my usual quark",
  "foodItemId": "seed-e-quark",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "cup",
    "displayLabel": "my usual quark",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T21:53:00.824Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "my usual quark" at 96% confidence.
Matched the common-food list to "Quark" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Lunch was my usual sardine tin." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Sardines in Water — card not captured.

Why verdict Logged (write-truth verified): Sardines in Water — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged Codex Test Pantry Sardines in Water. Using your saved "my usual sardine tin". Tell me if that is wrong.

2 · Card shown Logged Codex Test Pantry Sardines in Water. Using your saved "my usual sardine tin". Tell me if that is wrong.

3 · App data rows written created food_log_entry: Sardines in Water ×1 (tin) 150 cal · 18g P · 0g C · 8g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Codex Test Pantry Sardines in Water

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-sardine",
    "name": "Sardines in Water",
    "brand": "Codex Test Pantry",
    "source": "manual",
    "servingLabel": "tin",
    "servingQuantity": 1,
    "servingUnit": "tin",
    "macrosPerServing": {
      "calories": 150,
      "proteinGrams": 18,
      "fatGrams": 8,
      "carbsGrams": 0
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "my usual sardine tin",
  "foodItemId": "seed-e-sardine",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "tin",
    "displayLabel": "my usual sardine tin",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T21:53:12.089Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "my usual sardine tin" at 96% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Dinner side was my usual rice." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Cooked Basmati Rice — card not captured.

Why verdict Logged (write-truth verified): Cooked Basmati Rice — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged Cooked Basmati Rice. Using your saved "my usual rice". Tell me if that is wrong.

2 · Card shown Logged Cooked Basmati Rice. Using your saved "my usual rice". Tell me if that is wrong.

3 · App data rows written created food_log_entry: Cooked Basmati Rice ×1 (cup) 190 cal · 4g P · 41g C · 0.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Cooked Basmati Rice

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-rice",
    "name": "Cooked Basmati Rice",
    "brand": null,
    "source": "manual",
    "servingLabel": "cup",
    "servingQuantity": 1,
    "servingUnit": "cup",
    "macrosPerServing": {
      "calories": 190,
      "proteinGrams": 4,
      "fatGrams": 0.4,
      "carbsGrams": 41
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "my usual rice",
  "foodItemId": "seed-e-rice",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "cup",
    "displayLabel": "my usual rice",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T21:53:23.304Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "my usual rice" at 96% confidence.
Matched the common-food list to "Cooked white rice" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track my usual pecans." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Pecans — card not captured.

Why verdict Logged (write-truth verified): Pecans — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged Pecans. Using your saved "my usual pecans". Tell me if that is wrong.

2 · Card shown Logged Pecans. Using your saved "my usual pecans". Tell me if that is wrong.

3 · App data rows written created food_log_entry: Pecans ×1 (ounce) 196 cal · 2.6g P · 4g C · 20g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Pecans

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-pecans",
    "name": "Pecans",
    "brand": null,
    "source": "manual",
    "servingLabel": "ounce",
    "servingQuantity": 1,
    "servingUnit": "oz",
    "macrosPerServing": {
      "calories": 196,
      "proteinGrams": 2.6,
      "fatGrams": 20,
      "carbsGrams": 4
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "my usual pecans",
  "foodItemId": "seed-e-pecans",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "ounce",
    "displayLabel": "my usual pecans",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T21:53:34.533Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "my usual pecans" at 96% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log my default soy milk." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Unsweetened Soy Milk — card not captured.

Why verdict Logged (write-truth verified): Unsweetened Soy Milk — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged Unsweetened Soy Milk. Using your saved "my default soy milk". Tell me if that is wrong.

2 · Card shown Logged Unsweetened Soy Milk. Using your saved "my default soy milk". Tell me if that is wrong.

3 · App data rows written created food_log_entry: Unsweetened Soy Milk ×1 (cup) 80 cal · 7g P · 4g C · 4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Unsweetened Soy Milk

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-soymilk",
    "name": "Unsweetened Soy Milk",
    "brand": null,
    "source": "manual",
    "servingLabel": "cup",
    "servingQuantity": 1,
    "servingUnit": "cup",
    "macrosPerServing": {
      "calories": 80,
      "proteinGrams": 7,
      "fatGrams": 4,
      "carbsGrams": 4
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "my default soy milk",
  "foodItemId": "seed-e-soymilk",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "cup",
    "displayLabel": "my default soy milk",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T21:53:45.795Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "my default soy milk" at 96% confidence.
Matched the common-food list to "Milk" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add my usual morning matcha." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Matcha Latte — card not captured.

Why verdict Logged (write-truth verified): Matcha Latte — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged Codex Test Cafe Matcha Latte. Using your saved "my usual morning matcha". Tell me if that is wrong.

2 · Card shown Logged Codex Test Cafe Matcha Latte. Using your saved "my usual morning matcha". Tell me if that is wrong.

3 · App data rows written created food_log_entry: Matcha Latte ×1 (cup) 90 cal · 4g P · 12g C · 3g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Codex Test Cafe Matcha Latte

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-matcha",
    "name": "Matcha Latte",
    "brand": "Codex Test Cafe",
    "source": "manual",
    "servingLabel": "cup",
    "servingQuantity": 1,
    "servingUnit": "cup",
    "macrosPerServing": {
      "calories": 90,
      "proteinGrams": 4,
      "fatGrams": 3,
      "carbsGrams": 12
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "my usual morning matcha",
  "foodItemId": "seed-e-matcha",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "cup",
    "displayLabel": "my usual morning matcha",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T21:53:57.044Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "my usual morning matcha" at 96% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I had my normal cream of wheat." (difficulty —)0.1s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Decisive guess with stated assumption + invited correction (good-guessing).

Why verdict Decisive guess with stated assumption + invited correction (good-guessing).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged Cream of Wheat. Using your saved "my normal cream of wheat". Tell me if that is wrong.

2 · Card shown Logged Cream of Wheat. Using your saved "my normal cream of wheat". Tell me if that is wrong.

3 · App data rows written created food_log_entry: Cream of Wheat ×1 (cup) 130 cal · 4g P · 28g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Cream of Wheat

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-cow",
    "name": "Cream of Wheat",
    "brand": null,
    "source": "manual",
    "servingLabel": "cup",
    "servingQuantity": 1,
    "servingUnit": "cup",
    "macrosPerServing": {
      "calories": 130,
      "proteinGrams": 4,
      "fatGrams": 0.5,
      "carbsGrams": 28
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "my normal cream of wheat",
  "foodItemId": "seed-e-cow",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "cup",
    "displayLabel": "my normal cream of wheat",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T21:54:08.355Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "my normal cream of wheat" at 96% confidence.
Matched the common-food list to "Cream of wheat" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Track my go-to bone broth." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Chicken Bone Broth — card not captured.

Why verdict Logged (write-truth verified): Chicken Bone Broth — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged Codex Test Kitchen Chicken Bone Broth. Using your saved "my go to bone broth". Tell me if that is wrong.

2 · Card shown Logged Codex Test Kitchen Chicken Bone Broth. Using your saved "my go to bone broth". Tell me if that is wrong.

3 · App data rows written created food_log_entry: Chicken Bone Broth ×1 (cup) 45 cal · 9g P · 1g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Codex Test Kitchen Chicken Bone Broth

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-broth",
    "name": "Chicken Bone Broth",
    "brand": "Codex Test Kitchen",
    "source": "manual",
    "servingLabel": "cup",
    "servingQuantity": 1,
    "servingUnit": "cup",
    "macrosPerServing": {
      "calories": 45,
      "proteinGrams": 9,
      "fatGrams": 0.5,
      "carbsGrams": 1
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "my go-to bone broth",
  "foodItemId": "seed-e-broth",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "cup",
    "displayLabel": "my go-to bone broth",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T21:54:19.523Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "my go to bone broth" at 96% confidence.
Matched the common-food list to "Bone broth" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Log my usual evening chamomile." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Chamomile Tea — card not captured.

Why verdict Logged (write-truth verified): Chamomile Tea — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged Chamomile Tea. Using your saved "my usual evening chamomile". Tell me if that is wrong.

2 · Card shown Logged Chamomile Tea. Using your saved "my usual evening chamomile". Tell me if that is wrong.

3 · App data rows written created food_log_entry: Chamomile Tea ×1 (cup) 0 cal · 0g P · 0g C · 0g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Chamomile Tea

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-chamomile",
    "name": "Chamomile Tea",
    "brand": null,
    "source": "manual",
    "servingLabel": "cup",
    "servingQuantity": 1,
    "servingUnit": "cup",
    "macrosPerServing": {
      "calories": 0,
      "proteinGrams": 0,
      "fatGrams": 0,
      "carbsGrams": 0
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "my usual evening chamomile",
  "foodItemId": "seed-e-chamomile",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "cup",
    "displayLabel": "my usual evening chamomile",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T21:54:31.007Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "my usual evening chamomile" at 96% confidence.
Matched the common-food list to "Chamomile tea" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Add the same paneer as yesterday." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — MACROS off for "Paneer Cubes": logged 336 kcal vs expected ~260 (§14 15pct/25cal band)

Why verdict WRITE-TRUTH FAIL — MACROS off for "Paneer Cubes": logged 336 kcal vs expected ~260 (§14 15pct/25cal band)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged Paneer. Using your recent Paneer history. Tell me if that is wrong.

2 · Card shown Logged Paneer. Using your recent Paneer history. Tell me if that is wrong.

3 · App data rows written created food_log_entry: Paneer ×1 (0.75 cup) 336 cal · 17.9g P · 25.3g C · 17.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Codex Test Dairy Paneer Cubes

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-hist-paneer",
    "name": "Paneer Cubes",
    "brand": "Codex Test Dairy",
    "source": "manual",
    "servingLabel": "cup",
    "servingQuantity": 1,
    "servingUnit": "cup",
    "macrosPerServing": {
      "calories": 260,
      "proteinGrams": 18,
      "fatGrams": 20,
      "carbsGrams": 4
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 0.75 for seed-e-hist-paneer at 2026-07-17T13:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-hist-paneer",
  "quantity": 0.75,
  "loggedAt": "2026-07-17T13:00:00.000Z"
}

snapshot captured 2026-08-02T21:54:54.380Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Paneer" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Same Built bar as yesterday." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — WRONG/MISSING item "Built Bar Chocolate" — no saved row with matching identity (rows: Built Puff Bar)

Why verdict WRITE-TRUTH FAIL — WRONG/MISSING item "Built Bar Chocolate" — no saved row with matching identity (rows: Built Puff Bar)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged Built Built Puff Bar. Using your saved "my usual built bar". Tell me if that is wrong.

2 · Card shown Logged Built Built Puff Bar. Using your saved "my usual built bar". Tell me if that is wrong.

3 · App data rows written created food_log_entry: Built Puff Bar ×1 (bar) 150 cal · 17g P · 15g C · 4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Built Built Bar Chocolate

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-hist-built",
    "name": "Built Bar Chocolate",
    "brand": "Built",
    "source": "manual",
    "servingLabel": "bar",
    "servingQuantity": 1,
    "servingUnit": "bar",
    "macrosPerServing": {
      "calories": 150,
      "proteinGrams": 17,
      "fatGrams": 4,
      "carbsGrams": 15
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-hist-built at 2026-07-17T16:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-hist-built",
  "quantity": 1,
  "loggedAt": "2026-07-17T16:00:00.000Z"
}

snapshot captured 2026-08-02T21:55:05.632Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "Built Bar Chocolate" at 85% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Log the same gazpacho I had last night." (difficulty —)0.9s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.9s

1 · TTS said I heard you want the same the same gazpacho I had last night as before, but I do not have a matching saved log yet. Which exact food and amount should I log?

2 · Card shown I heard you want the same the same gazpacho I had last night as before, but I do not have a matching saved log yet. Which exact food and amount should I log?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Codex Test Kitchen Gazpacho Bowl

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-hist-gaz",
    "name": "Gazpacho Bowl",
    "brand": "Codex Test Kitchen",
    "source": "manual",
    "servingLabel": "bowl",
    "servingQuantity": 1,
    "servingUnit": "bowl",
    "macrosPerServing": {
      "calories": 90,
      "proteinGrams": 2,
      "fatGrams": 3,
      "carbsGrams": 14
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-hist-gaz at 2026-07-17T19:30:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-hist-gaz",
  "quantity": 1,
  "loggedAt": "2026-07-17T19:30:00.000Z"
}

snapshot captured 2026-08-02T21:55:17.682Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Gazpacho" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Same oatmeal as this morning." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Oatmeal Bowl — card not captured.

Why verdict Logged (write-truth verified): Oatmeal Bowl — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged Oatmeal Bowl. Using your recent Oatmeal Bowl history. Tell me if that is wrong.

2 · Card shown Logged Oatmeal Bowl. Using your recent Oatmeal Bowl history. Tell me if that is wrong.

3 · App data rows written created food_log_entry: Oatmeal Bowl ×1 (bowl) 150 cal · 5g P · 27g C · 2.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Oatmeal Bowl

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-hist-oats",
    "name": "Oatmeal Bowl",
    "brand": null,
    "source": "manual",
    "servingLabel": "bowl",
    "servingQuantity": 1,
    "servingUnit": "bowl",
    "macrosPerServing": {
      "calories": 150,
      "proteinGrams": 5,
      "fatGrams": 2.5,
      "carbsGrams": 27
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-hist-oats at 2026-07-18T07:30:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-hist-oats",
  "quantity": 1,
  "loggedAt": "2026-07-18T07:30:00.000Z"
}

snapshot captured 2026-08-02T21:55:29.067Z

6 · Why this food The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Add the same trout I had earlier." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Trout — card not captured.

Why verdict Logged (write-truth verified): Trout — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged Trout. Using your recent Trout history. Tell me if that is wrong.

2 · Card shown Logged Trout. Using your recent Trout history. Tell me if that is wrong.

3 · App data rows written created food_log_entry: Trout ×1 (4 oz) 191 cal · 27g P · 0g C · 8.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Baked Trout Fillet

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-hist-trout",
    "name": "Baked Trout Fillet",
    "brand": null,
    "source": "manual",
    "servingLabel": "fillet",
    "servingQuantity": 1,
    "servingUnit": "fillet",
    "macrosPerServing": {
      "calories": 190,
      "proteinGrams": 28,
      "fatGrams": 8,
      "carbsGrams": 0
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-hist-trout at 2026-07-18T12:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-hist-trout",
  "quantity": 1,
  "loggedAt": "2026-07-18T12:00:00.000Z"
}

snapshot captured 2026-08-02T21:55:40.394Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Trout" at 90% confidence.
Matched your past food logs to "Trout" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Same smoothie as yesterday." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Mango Protein Smoothie — card not captured.

Why verdict Logged (write-truth verified): Mango Protein Smoothie — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged Codex Test Recipe Mango Protein Smoothie. Using your recent Codex Test Recipe Mango Protein Smoothie history. Tell me if that is wrong.

2 · Card shown Logged Codex Test Recipe Mango Protein Smoothie. Using your recent Codex Test Recipe Mango Protein Smoothie history. Tell me if that is wrong.

3 · App data rows written created food_log_entry: Mango Protein Smoothie ×1 (smoothie) 290 cal · 24g P · 36g C · 5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Codex Test Recipe Mango Protein Smoothie

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-hist-smooth",
    "name": "Mango Protein Smoothie",
    "brand": "Codex Test Recipe",
    "source": "manual",
    "servingLabel": "smoothie",
    "servingQuantity": 1,
    "servingUnit": "smoothie",
    "macrosPerServing": {
      "calories": 290,
      "proteinGrams": 24,
      "fatGrams": 5,
      "carbsGrams": 36
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-hist-smooth at 2026-07-17T10:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-hist-smooth",
  "quantity": 1,
  "loggedAt": "2026-07-17T10:00:00.000Z"
}

snapshot captured 2026-08-02T21:55:51.762Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Log the mahi I had last time." (difficulty —)0.1s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I heard you want the same the mahi I had last time as before, but I do not have a matching saved log yet. Which exact food and amount should I log?

2 · Card shown I heard you want the same the mahi I had last time as before, but I do not have a matching saved log yet. Which exact food and amount should I log?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Grilled Mahi Mahi

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-hist-mahi",
    "name": "Grilled Mahi Mahi",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 140,
      "proteinGrams": 30,
      "fatGrams": 1.5,
      "carbsGrams": 0
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-hist-mahi at 2026-07-16T18:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-hist-mahi",
  "quantity": 1,
  "loggedAt": "2026-07-16T18:00:00.000Z"
}

snapshot captured 2026-08-02T21:56:03.092Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Same pita as earlier today." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Whole Wheat Pita — card not captured.

Why verdict Logged (write-truth verified): Whole Wheat Pita — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged Whole Wheat Pita. Using your recent Whole Wheat Pita history. Tell me if that is wrong.

2 · Card shown Logged Whole Wheat Pita. Using your recent Whole Wheat Pita history. Tell me if that is wrong.

3 · App data rows written created food_log_entry: Whole Wheat Pita ×1 (pita) 140 cal · 6g P · 28g C · 1.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Whole Wheat Pita

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-hist-pita",
    "name": "Whole Wheat Pita",
    "brand": null,
    "source": "manual",
    "servingLabel": "pita",
    "servingQuantity": 1,
    "servingUnit": "pita",
    "macrosPerServing": {
      "calories": 140,
      "proteinGrams": 6,
      "fatGrams": 1.5,
      "carbsGrams": 28
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-hist-pita at 2026-07-18T07:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-hist-pita",
  "quantity": 1,
  "loggedAt": "2026-07-18T07:00:00.000Z"
}

snapshot captured 2026-08-02T21:56:14.541Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Pita bread" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Snack was one Perfect Bar peanut butter." (difficulty —)1.4s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Peanut Butter Bar — card not captured.

Why verdict Logged (write-truth verified): Peanut Butter Bar — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.4s

1 · TTS said Got it, I logged Peanut Butter Bar — Perfect Bar.

2 · Card shown I logged Peanut Butter Bar — Perfect Bar — let me know if that's not right.

3 · App data rows written created food_log_entry: Peanut Butter Bar ×1 (1 serving (65 g)) 311 cal · 15g P · 24g C · 20g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:56:27.071Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Finished one Fairlife Core Power strawberry banana bottle." (difficulty —)0.6s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — IMPLAUSIBLE: 704 kcal is implausibly HIGH for 1 count of Core Power Strawberry Banana (typical 160-250 kcal [per-serving])

Why verdict WRITE-TRUTH FAIL — IMPLAUSIBLE: 704 kcal is implausibly HIGH for 1 count of Core Power Strawberry Banana (typical 160-250 kcal [per-serving])

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.6s

1 · TTS said Logged Core Power Strawberry Banana — Fairlife.

2 · Card shown Logged Core Power Strawberry Banana — Fairlife.

3 · App data rows written created food_log_entry: Core Power Strawberry Banana ×1 (414.0g) 704 cal · 107.6g P · 29g C · 18.6g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:56:41.143Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Track one Siggi's plain 4 percent cup." (difficulty —)1.2s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.2s

1 · TTS said What food did you have?

2 · Card shown What food did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:56:51.343Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Save one Built puff coconut bar." (difficulty —)2.7s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.7s

1 · TTS said What exact food and amount should I use for one built puff coconut bar? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for one built puff coconut bar? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:57:05.249Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"I ate one RXBAR peanut butter." (difficulty —)0.2s

Verdict Expected LOG — should log the entry. FAIL: WRITE-TRUTH FAIL — WRONG item/product identity "rxbar peanut butter" — saved row does not hit product alias (rows: Peanut Butter Bar)

Why verdict WRITE-TRUTH FAIL — WRONG item/product identity "rxbar peanut butter" — saved row does not hit product alias (rows: Peanut Butter Bar)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Got it, I logged Perfect Bar Peanut Butter Bar.

2 · Card shown I logged Perfect Bar Peanut Butter Bar — let me know if that's not right.

3 · App data rows written created food_log_entry: Peanut Butter Bar ×1 (1 serving (65 g)) 311 cal · 15g P · 24g C · 20g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:57:16.603Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Log a medium mango and one Built bar." (difficulty —)0.9s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.9s

1 · TTS said I need to resolve one Built bar before I log this meal. What should I use for one Built bar?

2 · Card shown I need to resolve one Built bar before I log this meal. What should I use for one Built bar?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:57:28.706Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I had oatmeal and black coffee." (difficulty —)0.3s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Decisive guess with stated assumption + invited correction (good-guessing).

Why verdict Decisive guess with stated assumption + invited correction (good-guessing).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Logged Cooked oatmeal and Coffee. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

2 · Card shown Logged Cooked oatmeal and Coffee. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

3 · App data rows written created food_log_entry: Cooked oatmeal ×1 (100 g) 71 cal · 2.5g P · 12g C · 1.5g F
created food_log_entry: Coffee ×1 (100 g) 1 cal · 0.1g P · 0g C · 0g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:57:40.128Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"Breakfast was two poached eggs and a pita." (difficulty —)0.4s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Egg, Pita bread, Pita bread — card not captured.

Why verdict Logged (write-truth verified): Egg, Pita bread, Pita bread — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.4s

1 · TTS said Logged Egg and Pita bread. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

2 · Card shown Logged Egg and Pita bread. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

3 · App data rows written created food_log_entry: Egg ×2 (50 g) 144 cal · 12.6g P · 0.8g C · 9.6g F
created food_log_entry: Pita bread ×1 (1 pita) 165 cal · 5.5g P · 33.4g C · 0.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:57:51.742Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Lunch was a falafel wrap and a clementine." (difficulty —)3.0s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.0s

1 · TTS said I need to resolve a falafel wrap before I log this meal. What should I use for a falafel wrap?

2 · Card shown I need to resolve a falafel wrap before I log this meal. What should I use for a falafel wrap?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:58:05.935Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Snack was pecans and a plum." (difficulty —)2.0s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.0s

1 · TTS said I need to resolve pecans before I log this meal. What should I use for pecans?

2 · Card shown I need to resolve pecans before I log this meal. What should I use for pecans?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:58:19.132Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Dinner was trout and roasted cauliflower." (difficulty —)0.3s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said How much should I log for Trout and Cauliflower? I did not log them yet because the amounts were not clear.

2 · Card shown How much should I log for Trout and Cauliflower? I did not log them yet because the amounts were not clear.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:58:30.540Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cauliflower" at 90% confidence.
Matched your saved foods and aliases to "trout" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I ate quark and blackberries." (difficulty —)0.3s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said How much should I log for Quark and Blackberries? I did not log them yet because the amounts were not clear.

2 · Card shown How much should I log for Quark and Blackberries? I did not log them yet because the amounts were not clear.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:58:42.025Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Blackberries" at 90% confidence.
Matched your past food logs to "Quark" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.0 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Track mahi and basmati rice." (difficulty —)2.3s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.3s

1 · TTS said I need to resolve mahi before I log this meal. What should I use for mahi?

2 · Card shown I need to resolve mahi before I log this meal. What should I use for mahi?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:59:07.268Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Add paneer and cucumber." (difficulty —)0.4s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.4s

1 · TTS said How much should I log for Paneer and Cucumber? I did not log them yet because the amounts were not clear.

2 · Card shown How much should I log for Paneer and Cucumber? I did not log them yet because the amounts were not clear.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:59:18.886Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cucumber" at 90% confidence.
Matched your saved foods and aliases to "paneer" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had a Perfect Bar and mineral water." (difficulty —)0.8s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Water, Perfect Bar Peanut Butter Bar — card not captured.

Why verdict Logged (write-truth verified): Water, Perfect Bar Peanut Butter Bar — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.8s

1 · TTS said Logged Perfect Bar Peanut Butter Bar — Perfect Bar and Water. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

2 · Card shown Logged Perfect Bar Peanut Butter Bar — Perfect Bar and Water. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

3 · App data rows written created food_log_entry: Perfect Bar Peanut Butter Bar ×1 (1 bar (1 serving (65 g))) 311 cal · 15g P · 24g C · 20g F
created food_log_entry: Water ×1 (240 g) 0 cal · 0g P · 0g C · 0g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:59:30.871Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Log gazpacho and arugula." (difficulty —)0.5s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said How much should I log for Gazpacho and Arugula? I did not log them yet because the amounts were not clear.

2 · Card shown How much should I log for Gazpacho and Arugula? I did not log them yet because the amounts were not clear.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:59:42.525Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Gazpacho" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Breakfast was oatmeal, mango, and coffee." (difficulty —)0.5s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Decisive guess with stated assumption + invited correction (good-guessing).

Why verdict Decisive guess with stated assumption + invited correction (good-guessing).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said Logged Cooked oatmeal, Mango, and Coffee. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

2 · Card shown Logged Cooked oatmeal, Mango, and Coffee. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

3 · App data rows written created food_log_entry: Cooked oatmeal ×1 (100 g) 71 cal · 2.5g P · 12g C · 1.5g F
created food_log_entry: Mango ×1 (100 g) 60 cal · 0.8g P · 15g C · 0.4g F
created food_log_entry: Coffee ×1 (100 g) 1 cal · 0.1g P · 0g C · 0g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T21:59:54.210Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I ate a plum, Laughing Cow, and rice cakes." (difficulty —)0.5s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said How much should I log for Laughing Cow and Plain rice cakes? I did not log them yet because the amounts were not clear.

2 · Card shown How much should I log for Laughing Cow and Plain rice cakes? I did not log them yet because the amounts were not clear.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:00:05.855Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Plain rice cakes" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Dinner was pork loin, wild rice, and zucchini." (difficulty —)0.4s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.4s

1 · TTS said How much should I log for Pork loin, Cooked wild rice, and Zucchini? I did not log them yet because the amounts were not clear.

2 · Card shown How much should I log for Pork loin, Cooked wild rice, and Zucchini? I did not log them yet because the amounts were not clear.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:00:17.329Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Cooked wild rice" at 90% confidence.
Matched your past food logs to "Pork loin" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.8 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Snack was cucumber and baba ganoush." (difficulty —)0.5s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Decisive guess with stated assumption + invited correction (good-guessing).

Why verdict Decisive guess with stated assumption + invited correction (good-guessing).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said Logged Cucumber and Baba Ganoush Smoked Eggplant Dip — MAMNOON. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

2 · Card shown Logged Cucumber and Baba Ganoush Smoked Eggplant Dip — MAMNOON. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

3 · App data rows written created food_log_entry: Cucumber ×1 (100 g) 15 cal · 0.7g P · 3.6g C · 0.1g F
created food_log_entry: Baba Ganoush Smoked Eggplant Dip ×1 (2 Tbsp) 25 cal · 1g P · 2g C · 2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:00:29.084Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Lunch was a salad and whole wheat pita." (difficulty —)2.4s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.4s

1 · TTS said I need to resolve a salad before I log this meal. What should I use for a salad?

2 · Card shown I need to resolve a salad before I log this meal. What should I use for a salad?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:00:42.633Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould ask a clarifying question"I had rice, white beans, and smoked turkey." (difficulty —)0.6s

Verdict Expected CLARIFY — should ask a clarifying question. FAIL: WRONG item vs seeded default/key — WRONG/MISSING item "smoked turkey" — no saved row with matching identity (rows: Beans, Cooked white rice)

Why verdict WRONG item vs seeded default/key — WRONG/MISSING item "smoked turkey" — no saved row with matching identity (rows: Beans, Cooked white rice)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.6s

1 · TTS said Logged Cooked white rice, Beans, and Turkey breast. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

2 · Card shown Logged Cooked white rice, Beans, and Turkey breast. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

3 · App data rows written created food_log_entry: Cooked white rice ×1 (100 g) 130 cal · 2.7g P · 28g C · 0.3g F
created food_log_entry: Beans ×1 (86 g) 109 cal · 7.5g P · 19.6g C · 0.4g F
created food_log_entry: Turkey breast ×1 (100 g) 135 cal · 29g P · 0g C · 1.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:00:54.905Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Track quark, tahini, and honeydew." (difficulty —)0.7s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.7s

1 · TTS said How much should I log for Quark, Tahini, and Honeydew? I did not log them yet because the amounts were not clear.

2 · Card shown How much should I log for Quark, Tahini, and Honeydew? I did not log them yet because the amounts were not clear.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:01:06.402Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Honeydew" at 90% confidence.
Matched your past food logs to "Quark" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould ask a clarifying question"Add eggs, turkey sausage, and cranberry juice." (difficulty —)0.8s

Verdict Expected CLARIFY — should ask a clarifying question. FAIL: WRONG item vs seeded default/key — WRONG/MISSING item "cranberry juice" — no saved row with matching identity (rows: Egg, Turkey sausage links)

Why verdict WRONG item vs seeded default/key — WRONG/MISSING item "cranberry juice" — no saved row with matching identity (rows: Egg, Turkey sausage links)

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.8s

1 · TTS said Logged Egg, Turkey sausage links, and Cranberry Juice Blend, 100% Juice. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

2 · Card shown Logged Egg, Turkey sausage links, and Cranberry Juice Blend, 100% Juice. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

3 · App data rows written created food_log_entry: Egg ×1 (50 g) 72 cal · 6.3g P · 0.4g C · 4.8g F
created food_log_entry: Turkey sausage links ×1 (50 g) 75 cal · 6.5g P · 1.5g C · 4.5g F
created food_log_entry: Cranberry Juice Blend, 100% Juice ×1 (1 fl oz (no ice)) 14 cal · 0g P · 3.5g C · 0.1g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:01:19.306Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I ate pasta and a side of snow peas." (difficulty —)1.1s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Decisive guess with stated assumption + invited correction (good-guessing).

Why verdict Decisive guess with stated assumption + invited correction (good-guessing).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.1s

1 · TTS said Logged Cooked pasta and Snow peas. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

2 · Card shown Logged Cooked pasta and Snow peas. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

3 · App data rows written created food_log_entry: Cooked pasta ×1 (140 g) 220 cal · 8.1g P · 43.3g C · 1.3g F
created food_log_entry: Snow peas ×1 (98 g) 41 cal · 2.7g P · 7.4g C · 0.2g F
created food_log_entry: Cooked pasta ×1 (140 g) 220 cal · 8.1g P · 43.3g C · 1.3g F
created food_log_entry: Snow peas ×1 (98 g) 41 cal · 2.7g P · 7.4g C · 0.2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:01:30.542Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Log a fig and sunflower butter." (difficulty —)0.5s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said How much should I log for Sunflower butter? I did not log them yet because the amounts were not clear.

2 · Card shown How much should I log for Sunflower butter? I did not log them yet because the amounts were not clear.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:01:42.164Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Sunflower butter" at 90% confidence.
Matched your past food logs to "Sunflower butter" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Breakfast was cereal and soy milk." (difficulty —)1.1s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.1s

1 · TTS said I need to resolve cereal before I log this meal. What should I use for cereal?

2 · Card shown I need to resolve cereal before I log this meal. What should I use for cereal?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:01:54.450Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould log the entry"I had a shakshuka with three eggs, tomatoes, and feta" (difficulty —)4.2s

Verdict Expected LOG — should log the entry. PASS: Logged (write-truth verified): Beans and Tomatoes, Fat Added, Egg, Shakshuka, Feta cheese — card not captured.

Why verdict Logged (write-truth verified): Beans and Tomatoes, Fat Added, Egg, Shakshuka, Feta cheese — card not captured.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 4.2s

1 · TTS said Logged Shakshuka. Includes three eggs, Beans and Tomatoes, Fat Added, and Feta cheese.

2 · Card shown Logged Shakshuka. Includes three eggs, Beans and Tomatoes, Fat Added, and Feta cheese.

3 · App data rows written created food_log_entry: Shakshuka ×1 (serving) 848 cal · 49g P · 51.4g C · 50.2g F
created food_log_entry: Egg ×1 (3 eggs) 215 cal · 18.9g P · 1g C · 14.3g F
created food_log_entry: Beans and Tomatoes, Fat Added ×1 (1 cup) 369 cal · 15.9g P · 46.3g C · 14.6g F
created food_log_entry: Feta cheese ×1 (100 g) 264 cal · 14.2g P · 4.1g C · 21.3g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:02:09.774Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould ask a clarifying question"Log a smoothie with mango, spinach, and whey protein." (difficulty —)1.9s

Verdict Expected CLARIFY — should ask a clarifying question. FAIL: Logged a BLIND guess — no stated assumption, no correction invited.

Why verdict Logged a BLIND guess — no stated assumption, no correction invited.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.9s

1 · TTS said Logged Smoothie. Includes Mango, Raw, Spinach, and Whey protein powder.

2 · Card shown Logged Smoothie. Includes Mango, Raw, Spinach, and Whey protein powder.

3 · App data rows written created food_log_entry: Smoothie ×1 (serving) 216 cal · 27.1g P · 26.9g C · 1.3g F
created food_log_entry: Mango, Raw ×1 (1 cup) 99 cal · 1.3g P · 24.8g C · 0.7g F
created food_log_entry: Spinach ×1 (30 g) 7 cal · 0.9g P · 1.1g C · 0.1g F
created food_log_entry: Whey protein powder ×1 (30 g) 110 cal · 24.9g P · 1g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:02:22.830Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I had a grain bowl with freekeh, chickpeas, and tahini." (difficulty —)2.6s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.6s

1 · TTS said What should I use for freekeh?

2 · Card shown What should I use for freekeh?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:02:36.610Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Breakfast was avocado toast with tomato and everything seasoning." (difficulty —)0.8s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.8s

1 · TTS said What exact ingredient should I use for everything seasoning?

2 · Card shown What exact ingredient should I use for everything seasoning?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:02:48.544Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Toast" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 2.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould ask a clarifying question"I made curry with chicken, cauliflower, coconut milk, and oil." (difficulty —)3.6s

Verdict Expected CLARIFY — should ask a clarifying question. FAIL: Logged a BLIND guess — no stated assumption, no correction invited.

Why verdict Logged a BLIND guess — no stated assumption, no correction invited.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.6s

1 · TTS said Logging your I made curry with chicken, cauliflower, coconut milk, and oil — still looking up the nutrition details; they'll appear in your log shortly.

2 · Card shown Logging your I made curry with chicken, cauliflower, coconut milk, and oil — still looking up the nutrition details; they'll appear in your log shortly.

3 · App data rows written pending_created food_log_entry: I made curry with chicken, cauliflower, coconut milk, and oil (pending)

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:03:11.542Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Dinner was a falafel plate with pita, hummus, and cucumber salad." (difficulty —)0.9s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.9s

1 · TTS said What exact ingredient should I use for falafel plate with pita hummus?

2 · Card shown What exact ingredient should I use for falafel plate with pita hummus?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:03:26.967Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "cucumber" at 90% confidence.
Matched the common-food list to "Falafel" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"I had udon with shrimp and scallions." (difficulty —)3.0s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.0s

1 · TTS said What should I use for scallions?

2 · Card shown What should I use for scallions?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:03:41.135Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould log the entry"Log a tuna sandwich with avocado and mustard." (difficulty —)0.5s

Verdict Expected LOG — should log the entry. FAIL: OVER-ASK — asked instead of logging (no saved row).

Why verdict OVER-ASK — asked instead of logging (no saved row).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said What exact ingredient should I use for mustard?

2 · Card shown What exact ingredient should I use for mustard?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:03:52.838Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I built a yogurt bowl with freekeh, honey, and blackberries." (difficulty —)1.8s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.8s

1 · TTS said What should I use for freekeh?

2 · Card shown What should I use for freekeh?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:04:05.676Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould log the entry"Dinner was fish tacos with cabbage, mango salsa, and crema." (difficulty —)3.6s

Why verdict UNVERIFIABLE — no saved row captured (untagged / lossy) — resolve via the notes.utterance_trace_id join. App claim not credited; resolve via the notes.utterance_trace_id join.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.6s

1 · TTS said Logging your Dinner was fish tacos with cabbage, mango salsa, and crema — still looking up the nutrition details; they'll appear in your log shortly.

2 · Card shown Logging your Dinner was fish tacos with cabbage, mango salsa, and crema — still looking up the nutrition details; they'll appear in your log shortly.

3 · App data rows written pending_created food_log_entry: Dinner was fish tacos with cabbage, mango salsa, and crema (pending)

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:04:27.748Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Add some freekeh." (difficulty —)1.4s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.4s

1 · TTS said What food did you have?

2 · Card shown What food did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:04:33.014Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I had cream of wheat." (difficulty —)0.1s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said How much should I log for Cream of wheat? I did not log it yet because the amount was not clear.

2 · Card shown How much should I log for Cream of wheat? I did not log it yet because the amount was not clear.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:04:44.251Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "cream of wheat" at 90% confidence.
Matched the common-food list to "Cream of wheat" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Log basmati for dinner." (difficulty —)1.6s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.6s

1 · TTS said What food did you have?

2 · Card shown What food did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:04:57.056Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Track udon." (difficulty —)0.3s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said How much udon did you have?

2 · Card shown How much udon did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:05:08.538Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I ate plantain chips." (difficulty —)5.2s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.2s

1 · TTS said Was the potato plain, or did it have toppings like butter, sour cream, or cheese? I did not log it yet because add-ons can change the nutrition.

2 · Card shown Was the potato plain, or did it have toppings like butter, sour cream, or cheese? I did not log it yet because add-ons can change the nutrition.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:05:25.005Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Add pecans." (difficulty —)0.9s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.9s

1 · TTS said How much pecans did you have?

2 · Card shown How much pecans did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:05:36.993Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Log a bowl of gazpacho." (difficulty —)8.0s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 8.0s

1 · TTS said What exact food and amount should I use for Gazpacho? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for Gazpacho? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:05:56.173Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I had a plate of freekeh." (difficulty —)2.7s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.7s

1 · TTS said What exact food and amount should I use for a plate of freekeh? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for a plate of freekeh? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:06:10.328Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Track cranberry juice." (difficulty —)1.5s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.5s

1 · TTS said How much cranberry juice did you have?

2 · Card shown How much cranberry juice did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:06:22.701Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Add matcha." (difficulty —)0.8s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.8s

1 · TTS said How much matcha did you have?

2 · Card shown How much matcha did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:06:34.722Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I drank sake." (difficulty —)0.2s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said How much sake did you have?

2 · Card shown How much sake did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:06:46.099Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Log sorbet." (difficulty —)0.8s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.8s

1 · TTS said How much sorbet did you have?

2 · Card shown How much sorbet did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:06:58.059Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Had some sunflower seeds." (difficulty —)0.1s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said How much should I log for Sunflower seeds? I did not log it yet because the amount was not clear.

2 · Card shown How much should I log for Sunflower seeds? I did not log it yet because the amount was not clear.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:07:10.244Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Sunflower seeds" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Add rice cakes." (difficulty —)0.1s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said How much should I log for Plain rice cakes? I did not log it yet because the amount was not clear.

2 · Card shown How much should I log for Plain rice cakes? I did not log it yet because the amount was not clear.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:07:20.648Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Plain rice cakes" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I had a Built bar." (difficulty —)0.5s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said Which type of Built bar was it? There are several with different nutrition.

2 · Card shown Which type of Built bar was it? There are several with different nutrition.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:07:32.357Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Log a quark cup." (difficulty —)0.2s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Decisive guess with stated assumption + invited correction (good-guessing).

Why verdict Decisive guess with stated assumption + invited correction (good-guessing).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Logged a quark cup. Assumed 1 cup as the default serving. Tell me if you want a different amount.

2 · Card shown Logged a quark cup. Assumed 1 cup as the default serving. Tell me if you want a different amount.

3 · App data rows written created food_log_entry: Quark ×1 (1 cup) 151 cal · 27.1g P · 9g C · 0.5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:07:55.435Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Quark" at 90% confidence.
Matched your past food logs to "Codex Test Dairy Plain Quark" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Track a plant milk." (difficulty —)4.5s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 4.5s

1 · TTS said Which one did you mean? 1. my default soy milk, 2. Soy Milk, Unsweetened, 3. unsweetened soy milk

2 · Card shown Which one did you mean? 1. my default soy milk, 2. Soy Milk, Unsweetened, 3. unsweetened soy milk

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:08:11.094Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I ate a wrap." (difficulty —)0.8s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.8s

1 · TTS said Which type of wrap was it? There are several with different nutrition.

2 · Card shown Which type of wrap was it? There are several with different nutrition.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:08:23.010Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Add a recovery shake." (difficulty —)1.6s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.6s

1 · TTS said I couldn't find Add A Recovery Protein Shake in my nutrition database. What exact name is on the package? I did not log it because I'd only be guessing the nutrition. If you prefer, tell me to log a generic protein shake.

2 · Card shown I couldn't find Add A Recovery Protein Shake in my nutrition database. What exact name is on the package? I did not log it because I'd only be guessing the nutrition. If you prefer, tell me to log a generic protein shake.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:08:35.739Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I had matcha with milk." (difficulty —)5.2s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.2s

1 · TTS said How much matcha with milk did you have?

2 · Card shown How much matcha with milk did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:08:52.100Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Log pita." (difficulty —)3.0s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.0s

1 · TTS said How much pita did you have?

2 · Card shown How much pita did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:09:06.267Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Track a fish sandwich." (difficulty —)2.7s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.7s

1 · TTS said What exact food and amount should I use for a fish sandwich? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for a fish sandwich? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:09:20.220Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I had flatbread." (difficulty —)0.7s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.7s

1 · TTS said How much flatbread did you have?

2 · Card shown How much flatbread did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:09:32.099Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Add a scone with jam." (difficulty —)1.5s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Asked a clarification as expected.

Why verdict Asked a clarification as expected.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.5s

1 · TTS said What exact food and amount should I use for a scone with jam? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for a scone with jam? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:09:44.793Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Log my usual coffee size." (difficulty —)0.1s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Decisive guess with stated assumption + invited correction (good-guessing).

Why verdict Decisive guess with stated assumption + invited correction (good-guessing).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged coffee. Using your saved "log my usual coffee size". Tell me if that is wrong.

2 · Card shown Logged coffee. Using your saved "log my usual coffee size". Tell me if that is wrong.

3 · App data rows written created food_log_entry: coffee ×1 (serving) 150 cal · 10g P · 15g C · 5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: coffee

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-coffee-confirm",
    "name": "coffee",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 150,
      "proteinGrams": 10,
      "fatGrams": 5,
      "carbsGrams": 15
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "log my usual coffee size",
  "foodItemId": "seed-e-coffee-confirm",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "serving",
    "displayLabel": "log my usual coffee size",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T22:09:56.083Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "log my usual coffee size" at 96% confidence.
Matched the common-food list to "Coffee" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Add the regular breakfast portion." (difficulty —)0.1s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Decisive guess with stated assumption + invited correction (good-guessing).

Why verdict Decisive guess with stated assumption + invited correction (good-guessing).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged breakfast portion. Using your saved "add the regular breakfast portion". Tell me if that is wrong.

2 · Card shown Logged breakfast portion. Using your saved "add the regular breakfast portion". Tell me if that is wrong.

3 · App data rows written created food_log_entry: breakfast portion ×1 (serving) 150 cal · 10g P · 15g C · 5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: breakfast portion

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-bfast-portion",
    "name": "breakfast portion",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 150,
      "proteinGrams": 10,
      "fatGrams": 5,
      "carbsGrams": 15
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "add the regular breakfast portion",
  "foodItemId": "seed-e-bfast-portion",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "serving",
    "displayLabel": "add the regular breakfast portion",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T22:10:07.488Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "add the regular breakfast portion" at 96% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"Track my default snack amount." (difficulty —)0.1s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Decisive guess with stated assumption + invited correction (good-guessing).

Why verdict Decisive guess with stated assumption + invited correction (good-guessing).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged snack. Using your saved "track my default snack amount". Tell me if that is wrong.

2 · Card shown Logged snack. Using your saved "track my default snack amount". Tell me if that is wrong.

3 · App data rows written created food_log_entry: snack ×1 (serving) 150 cal · 10g P · 15g C · 5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: snack

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-snack-amt",
    "name": "snack",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 150,
      "proteinGrams": 10,
      "fatGrams": 5,
      "carbsGrams": 15
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "track my default snack amount",
  "foodItemId": "seed-e-snack-amt",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "serving",
    "displayLabel": "track my default snack amount",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T22:10:18.765Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "track my default snack amount" at 96% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould ask a clarifying question"I had the usual lunch size." (difficulty —)0.1s

Verdict Expected CLARIFY — should ask a clarifying question. PASS: Decisive guess with stated assumption + invited correction (good-guessing).

Why verdict Decisive guess with stated assumption + invited correction (good-guessing).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Logged lunch. Using your saved "i had the usual lunch size". Tell me if that is wrong.

2 · Card shown Logged lunch. Using your saved "i had the usual lunch size". Tell me if that is wrong.

3 · App data rows written created food_log_entry: lunch ×1 (serving) 150 cal · 10g P · 15g C · 5g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: lunch

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-lunch-size",
    "name": "lunch",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 150,
      "proteinGrams": 10,
      "fatGrams": 5,
      "carbsGrams": 15
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

food_default

{
  "type": "food_default",
  "phrase": "i had the usual lunch size",
  "foodItemId": "seed-e-lunch-size",
  "settings": {
    "quickLogMode": "fixed",
    "defaultQuantity": 1,
    "defaultServingLabel": "serving",
    "displayLabel": "i had the usual lunch size",
    "confirmBeforeLog": false
  }
}

snapshot captured 2026-08-02T22:10:30.027Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "i had the usual lunch size" at 96% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.1 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould complete the pending action"Half a cup." (difficulty —)0.5s

Verdict Expected CLARIFY_ANSWER — should complete the pending action. UNVERIFIED: CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

Why verdict CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.5s

1 · TTS said What food did you have?

2 · Card shown What food did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:10:41.735Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould complete the pending action"One cup, plain." (difficulty —)0.1s

Why verdict CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Should I add that as a to-do?

2 · Card shown Should I add that as a to-do?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:10:53.089Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your past food logs to "Codex Test Dairy Plain Quark" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould complete the pending action"The Built puff coconut." (difficulty —)2.4s

Why verdict CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.4s

1 · TTS said I don’t have “Built puff coconut” in my food database. Could you tell me the exact name on the package, or the brand and flavor so I can look it up correctly? Or would you like to log a generic protein bar for now?

2 · Card shown I don’t have “Built puff coconut” in my food database. Could you tell me the exact name on the package, or the brand and flavor so I can look it up correctly? Or would you like to log a generic protein bar for now?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:11:06.631Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "Built Puff Bar" at 88% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould complete the pending action"Unsweetened soy milk." (difficulty —)4.0s

Why verdict CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 4.0s

1 · TTS said How much Unsweetened soy milk did you have?

2 · Card shown How much Unsweetened soy milk did you have?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:11:21.826Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould complete the pending action"No, that's not it — the unsweetened one." (difficulty —)0.1s

Why verdict CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said What should I change about lunch?

2 · Card shown What should I change about lunch?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:11:33.018Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould complete the pending action"Wrong one. The mango, not pineapple." (difficulty —)0.6s

Verdict Expected CLARIFY_ANSWER — should complete the pending action. PASS: Completed the pending action.

Why verdict Completed the pending action.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.6s

1 · TTS said Logged Mango and Pineapple. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

2 · Card shown Logged Mango and Pineapple. Assumed catalog default servings where you did not say an amount — tell me if that is not right.

3 · App data rows written created food_log_entry: Mango ×1 (100 g) 60 cal · 0.8g P · 15g C · 0.4g F
created food_log_entry: Pineapple ×1 (100 g) 50 cal · 0.5g P · 13.1g C · 0.1g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:11:44.797Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould complete the pending action"No, not the flavored quark." (difficulty —)0.1s

Why verdict CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said What should I change about Pineapple?

2 · Card shown What should I change about Pineapple?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:11:56.008Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Quark" at 90% confidence.
Matched your past food logs to "Quark" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould decline"Cancel that." (difficulty —)0.0s

Verdict Expected DECLINE — should decline. PASS: Declined / no unwanted write.

Why verdict Declined / no unwanted write.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.0s

1 · TTS said Should I add that as a to-do?

2 · Card shown Should I add that as a to-do?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: Pending Food Item

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-corr-pending-3",
    "name": "Pending Food Item",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 150,
      "proteinGrams": 10,
      "fatGrams": 5,
      "carbsGrams": 15
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-corr-pending-3 at 2026-07-18T11:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-corr-pending-3",
  "quantity": 1,
  "loggedAt": "2026-07-18T11:00:00.000Z"
}

snapshot captured 2026-08-02T22:12:18.763Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould complete the pending action"No, that's not the bar I meant." (difficulty —)0.1s

Why verdict CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said What should I change about Mango?

2 · Card shown What should I change about Mango?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:12:41.210Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould complete the pending action"Not the large — the medium mango." (difficulty —)5.3s

Why verdict CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.3s

1 · TTS said What exact food and amount should I use for Mango? I did not log it yet because I could not safely finish that food log.

2 · Card shown What exact food and amount should I use for Mango? I did not log it yet because I could not safely finish that food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:12:57.769Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould complete the pending action"Wrong item. Pick the grilled mahi." (difficulty —)5.7s

Why verdict CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 5.7s

1 · TTS said What would you like me to do with that?

2 · Card shown What would you like me to do with that?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:13:14.558Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould complete the pending action"That's not it; the plain one." (difficulty —)2.9s

Why verdict CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.9s

1 · TTS said Got it. You’d like the plain protein bar instead of the coconut flavor. Would you like to log it now, or would you prefer to wait until after your workout?

2 · Card shown Got it. You’d like the plain protein bar instead of the coconut flavor. Would you like to log it now, or would you prefer to wait until after your workout?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:13:39.803Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould complete the pending action"No, the other pita." (difficulty —)3.8s

Verdict Expected CLARIFY_ANSWER — should complete the pending action. PASS: Completed the pending action.

Why verdict Completed the pending action.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.8s

1 · TTS said Updated to the other pita.

2 · Card shown Updated the food entry from "Mango" to "the other pita".

3 · App data rows written updated food_log_entry: Pita bread ×1 (60 g) 165 cal · 5.5g P · 33.4g C · 0.7g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:13:54.961Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould complete the pending action"Not those pecans — the raw ones." (difficulty —)3.6s

Why verdict CLARIFY_ANSWER not gradeable single-shot — no pending clarification in this replay (app re-asked / mis-parsed as fresh); needs multi-turn replay.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.6s

1 · TTS said I did not make any app changes. What would you like me to do with that?

2 · Card shown I did not make any app changes. What would you like me to do with that?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:14:20.937Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "pecans" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould delete the entry"Delete the mango from earlier." (difficulty —)0.1s

Verdict Expected DELETE — should delete the entry. FAIL: Did not delete (not understood / no removal).

Why verdict Did not delete (not understood / no removal).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I could not find a list matching earlier. Which list has mango?

2 · Card shown I could not find a list matching earlier. Which list has mango?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:14:43.437Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "mango" at 90% confidence.
Matched the common-food list to "Mango" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould delete the entry"Remove that Built bar." (difficulty —)0.2s

Verdict Expected DELETE — should delete the entry. PASS: Deleted/removed the requested food with database removal proof.

Why verdict Deleted/removed the requested food with database removal proof.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Deleted Built Bar Chocolate.

2 · Card shown Deleted Built Bar Chocolate.

3 · App data rows written deleted food_log_entry:

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:15:01.592Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "Built Puff Bar" at 85% confidence.
Matched your past food logs to "Built Bar Chocolate" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould delete the entry"Delete the cream of wheat entry." (difficulty —)0.1s

Verdict Expected DELETE — should delete the entry. FAIL: Did not delete (not understood / no removal).

Why verdict Did not delete (not understood / no removal).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Which food entry did you want to correct?

2 · Card shown Which food entry did you want to correct?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:15:06.025Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "cream of wheat" at 90% confidence.
Matched the common-food list to "Cream of wheat" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould delete the entry"Take off the eggs." (difficulty —)0.1s

Verdict Expected DELETE — should delete the entry. FAIL: Did not delete (not understood / no removal).

Why verdict Did not delete (not understood / no removal).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Which food entry did you want to correct?

2 · Card shown Which food entry did you want to correct?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:15:17.236Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "egg" at 90% confidence.
Matched the common-food list to "Egg" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.3 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould delete the entry"Remove the blackberries." (difficulty —)0.1s

Verdict Expected DELETE — should delete the entry. FAIL: Did not delete (not understood / no removal).

Why verdict Did not delete (not understood / no removal).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Which food entry did you want to correct?

2 · Card shown Which food entry did you want to correct?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:15:28.439Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Blackberries" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould delete the entry"Delete the mahi log." (difficulty —)0.0s

Verdict Expected DELETE — should delete the entry. FAIL: Did not delete (not understood / no removal).

Why verdict Did not delete (not understood / no removal).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.0s

1 · TTS said Which food entry did you want to correct?

2 · Card shown Which food entry did you want to correct?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:15:39.611Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould delete the entry"Scratch the sunflower butter." (difficulty —)0.1s

Verdict Expected DELETE — should delete the entry. FAIL: Did not delete (not understood / no removal).

Why verdict Did not delete (not understood / no removal).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Which food entry did you want to correct?

2 · Card shown Which food entry did you want to correct?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:15:50.832Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Sunflower butter" at 90% confidence.
Matched your past food logs to "Sunflower butter" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould delete the entry"Remove the soy milk entry." (difficulty —)0.1s

Verdict Expected DELETE — should delete the entry. FAIL: Did not delete (not understood / no removal).

Why verdict Did not delete (not understood / no removal).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Which food entry did you want to correct?

2 · Card shown Which food entry did you want to correct?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:16:02.043Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Milk" at 90% confidence.
Matched your past food logs to "Unsweetened Soy Milk" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould delete the entry"Delete that baklava I logged last night." (difficulty —)0.2s

Verdict Expected DELETE — should delete the entry. PASS: Deleted/removed the requested food with database removal proof.

Why verdict Deleted/removed the requested food with database removal proof.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Deleted baklava.

2 · Card shown Deleted baklava.

3 · App data rows written deleted food_log_entry:

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: baklava

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-del-0-baklava",
    "name": "baklava",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 200,
      "proteinGrams": 5,
      "fatGrams": 10,
      "carbsGrams": 20
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-del-0-baklava at 2026-07-17T20:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-del-0-baklava",
  "quantity": 1,
  "loggedAt": "2026-07-17T20:00:00.000Z"
}

snapshot captured 2026-08-02T22:16:19.926Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 2.0 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould delete the entry"Remove yesterday's gelato scoop." (difficulty —)0.2s

Verdict Expected DELETE — should delete the entry. PASS: Deleted/removed the requested food with database removal proof.

Why verdict Deleted/removed the requested food with database removal proof.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Deleted gelato.

2 · Card shown Deleted gelato.

3 · App data rows written deleted food_log_entry:

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: gelato

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-del-1-gelato",
    "name": "gelato",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 200,
      "proteinGrams": 5,
      "fatGrams": 10,
      "carbsGrams": 20
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-del-1-gelato at 2026-07-17T20:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-del-1-gelato",
  "quantity": 1,
  "loggedAt": "2026-07-17T20:00:00.000Z"
}

snapshot captured 2026-08-02T22:16:42.803Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.8 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould delete the entry"Delete the leftover gyro entry." (difficulty —)0.2s

Verdict Expected DELETE — should delete the entry. PASS: Deleted/removed the requested food with database removal proof.

Why verdict Deleted/removed the requested food with database removal proof.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Deleted gyro.

2 · Card shown Deleted gyro.

3 · App data rows written deleted food_log_entry:

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: gyro

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-del-2-gyro",
    "name": "gyro",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 200,
      "proteinGrams": 5,
      "fatGrams": 10,
      "carbsGrams": 20
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-del-2-gyro at 2026-07-17T20:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-del-2-gyro",
  "quantity": 1,
  "loggedAt": "2026-07-17T20:00:00.000Z"
}

snapshot captured 2026-08-02T22:16:54.550Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould delete the entry"Take out the sake from last night." (difficulty —)1.6s

Verdict Expected DELETE — should delete the entry. FAIL: Did not delete (not understood / no removal).

Why verdict Did not delete (not understood / no removal).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.6s

1 · TTS said I didn’t make any app changes. Would you like me to remove the sake entry from your food log?

2 · Card shown I didn’t make any app changes. Would you like me to remove the sake entry from your food log?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: sake

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-del-3-sake",
    "name": "sake",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 200,
      "proteinGrams": 5,
      "fatGrams": 10,
      "carbsGrams": 20
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-del-3-sake at 2026-07-17T20:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-del-3-sake",
  "quantity": 1,
  "loggedAt": "2026-07-17T20:00:00.000Z"
}

snapshot captured 2026-08-02T22:17:00.441Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould delete the entry"Remove the pastry log from earlier this week." (difficulty —)0.2s

Verdict Expected DELETE — should delete the entry. PASS: Deleted/removed the requested food with database removal proof.

Why verdict Deleted/removed the requested food with database removal proof.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Deleted pastry.

2 · Card shown Deleted pastry.

3 · App data rows written deleted food_log_entry:

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: pastry

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-del-4-pastry",
    "name": "pastry",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 200,
      "proteinGrams": 5,
      "fatGrams": 10,
      "carbsGrams": 20
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-del-4-pastry at 2026-07-17T20:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-del-4-pastry",
  "quantity": 1,
  "loggedAt": "2026-07-17T20:00:00.000Z"
}

snapshot captured 2026-08-02T22:17:18.712Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould delete the entry"Delete the fries I shouldn't have logged." (difficulty —)0.2s

Verdict Expected DELETE — should delete the entry. PASS: Deleted/removed the requested food with database removal proof.

Why verdict Deleted/removed the requested food with database removal proof.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Deleted fries.

2 · Card shown Deleted fries.

3 · App data rows written deleted food_log_entry:

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: fries

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-del-5-fries",
    "name": "fries",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 200,
      "proteinGrams": 5,
      "fatGrams": 10,
      "carbsGrams": 20
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-del-5-fries at 2026-07-17T20:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-del-5-fries",
  "quantity": 1,
  "loggedAt": "2026-07-17T20:00:00.000Z"
}

snapshot captured 2026-08-02T22:17:30.029Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould update the entry"Change the sequence mango quantity to two." (difficulty —)0.1s

Verdict Expected UPDATE — should update the entry. FAIL: Reported could-not-confirm (write-confirmation false negative).

Why verdict Reported could-not-confirm (write-confirmation false negative).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I could not confirm that was saved, so I did not mark it as done.

2 · Card shown I could not confirm that the change was saved, so I did not mark it as done.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:17:34.258Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "mango" at 90% confidence.
Matched the common-food list to "Mango" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould update the entry"Actually make the Built bar two bars." (difficulty —)0.1s

Verdict Expected UPDATE — should update the entry. FAIL: Reported could-not-confirm (write-confirmation false negative).

Why verdict Reported could-not-confirm (write-confirmation false negative).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I could not confirm that was saved, so I did not mark it as done.

2 · Card shown I could not confirm that the change was saved, so I did not mark it as done.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:17:45.490Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "Built Puff Bar" at 84% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould update the entry"Update the cream of wheat to one and a half cups." (difficulty —)0.3s

Verdict Expected UPDATE — should update the entry. PASS: Updated as asked.

Why verdict Updated as asked.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.3s

1 · TTS said Updated Cream of Wheat to 1.5 cup.

2 · Card shown Updated Cream of Wheat to 1.5 cup.

3 · App data rows written updated food_log_entry: Cream of Wheat ×1.5 (cup) 195 cal · 6g P · 42g C · 0.8g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:17:56.899Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "cream of wheat" at 90% confidence.
Matched the common-food list to "Cream of wheat" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould update the entry"Change the eggs to three." (difficulty —)0.1s

Verdict Expected UPDATE — should update the entry. FAIL: Reported could-not-confirm (write-confirmation false negative).

Why verdict Reported could-not-confirm (write-confirmation false negative).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I could not confirm that was saved, so I did not mark it as done.

2 · Card shown I could not confirm that the change was saved, so I did not mark it as done.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:18:08.196Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "egg" at 90% confidence.
Matched the common-food list to "Egg" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould update the entry"Make the blackberries a half cup." (difficulty —)0.1s

Verdict Expected UPDATE — should update the entry. FAIL: Reported could-not-confirm (write-confirmation false negative).

Why verdict Reported could-not-confirm (write-confirmation false negative).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I could not confirm that was saved, so I did not mark it as done.

2 · Card shown I could not confirm that the change was saved, so I did not mark it as done.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:18:19.497Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Blackberries" at 90% confidence.
Matched your past food logs to "Blackberries" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould update the entry"Update the mahi to seven ounces." (difficulty —)0.1s

Verdict Expected UPDATE — should update the entry. FAIL: Reported could-not-confirm (write-confirmation false negative).

Why verdict Reported could-not-confirm (write-confirmation false negative).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I could not confirm that was saved, so I did not mark it as done.

2 · Card shown I could not confirm that the change was saved, so I did not mark it as done.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:18:30.668Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould update the entry"Change sunflower butter to two tablespoons." (difficulty —)0.1s

Verdict Expected UPDATE — should update the entry. FAIL: Reported could-not-confirm (write-confirmation false negative).

Why verdict Reported could-not-confirm (write-confirmation false negative).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I could not confirm that was saved, so I did not mark it as done.

2 · Card shown I could not confirm that the change was saved, so I did not mark it as done.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:18:41.913Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Sunflower butter" at 90% confidence.
Matched your past food logs to "Sunflower butter" at 82% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.8 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould update the entry"Actually the soy milk was vanilla, not plain." (difficulty —)3.2s

Verdict Expected UPDATE — should update the entry. PASS: Updated as asked.

Why verdict Updated as asked.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 3.2s

1 · TTS said Updated to vanilla soy milk.

2 · Card shown Updated the food entry from "Unsweetened Soy Milk" to "vanilla soy milk".

3 · App data rows written updated food_log_entry: Milk ×1 (100 g) 50 cal · 3.3g P · 4.8g C · 2g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:18:56.212Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould update the entry"Actually that was coconut, not chocolate." (difficulty —)1.4s

Verdict Expected UPDATE — should update the entry. FAIL: Did not update (treated as new / not understood).

Why verdict Did not update (treated as new / not understood).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.4s

1 · TTS said Which exact coconut should replace Pita bread?

2 · Card shown Which exact coconut should replace Pita bread?

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: chocolate item

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-edit-0-chocolate-item",
    "name": "chocolate item",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 180,
      "proteinGrams": 8,
      "fatGrams": 6,
      "carbsGrams": 18
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-edit-0-chocolate-item at 2026-07-18T10:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-edit-0-chocolate-item",
  "quantity": 1,
  "loggedAt": "2026-07-18T10:00:00.000Z"
}

snapshot captured 2026-08-02T22:19:08.710Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould update the entry"Change the portion to two cups." (difficulty —)0.1s

Verdict Expected UPDATE — should update the entry. FAIL: Reported could-not-confirm (write-confirmation false negative).

Why verdict Reported could-not-confirm (write-confirmation false negative).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I could not confirm that was saved, so I did not mark it as done.

2 · Card shown I could not confirm that the change was saved, so I did not mark it as done.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: basmati

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-edit-1-basmati",
    "name": "basmati",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 180,
      "proteinGrams": 8,
      "fatGrams": 6,
      "carbsGrams": 18
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-edit-1-basmati at 2026-07-18T10:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-edit-1-basmati",
  "quantity": 1,
  "loggedAt": "2026-07-18T10:00:00.000Z"
}

snapshot captured 2026-08-02T22:19:20.010Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould update the entry"Make it half a cup instead." (difficulty —)0.1s

Verdict Expected UPDATE — should update the entry. FAIL: Reported could-not-confirm (write-confirmation false negative).

Why verdict Reported could-not-confirm (write-confirmation false negative).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I could not confirm that was saved, so I did not mark it as done.

2 · Card shown I could not confirm that the change was saved, so I did not mark it as done.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: cream of wheat

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-edit-2-cream-of-wheat",
    "name": "cream of wheat",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 180,
      "proteinGrams": 8,
      "fatGrams": 6,
      "carbsGrams": 18
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-edit-2-cream-of-wheat at 2026-07-18T10:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-edit-2-cream-of-wheat",
  "quantity": 1,
  "loggedAt": "2026-07-18T10:00:00.000Z"
}

snapshot captured 2026-08-02T22:19:31.216Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould update the entry"Update that to seven ounces." (difficulty —)0.2s

Verdict Expected UPDATE — should update the entry. PASS: Updated as asked.

Why verdict Updated as asked.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.2s

1 · TTS said Updated Pita bread to 7 oz.

2 · Card shown Updated Pita bread to 7 oz.

3 · App data rows written updated food_log_entry: Pita bread ×1.984465 (100 g) 327 cal · 10.9g P · 66.3g C · 1.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: mahi

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-edit-3-mahi",
    "name": "mahi",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 180,
      "proteinGrams": 8,
      "fatGrams": 6,
      "carbsGrams": 18
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-edit-3-mahi at 2026-07-18T10:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-edit-3-mahi",
  "quantity": 1,
  "loggedAt": "2026-07-18T10:00:00.000Z"
}

snapshot captured 2026-08-02T22:19:42.592Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould update the entry"Actually it was tahini, not sunflower butter." (difficulty —)1.8s

Verdict Expected UPDATE — should update the entry. PASS: Updated as asked.

Why verdict Updated as asked.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 1.8s

1 · TTS said Updated to tahini.

2 · Card shown Updated the food entry from "Pita bread" to "tahini".

3 · App data rows written updated food_log_entry: Tahini ×1 (1 tablespoon) 105 cal · 3g P · 2.1g C · 9.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: sunflower butter

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-edit-4-sunflower-butter",
    "name": "sunflower butter",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 180,
      "proteinGrams": 8,
      "fatGrams": 6,
      "carbsGrams": 18
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-edit-4-sunflower-butter at 2026-07-18T10:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-edit-4-sunflower-butter",
  "quantity": 1,
  "loggedAt": "2026-07-18T10:00:00.000Z"
}

snapshot captured 2026-08-02T22:19:55.593Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould update the entry"Change the mango to a large." (difficulty —)2.3s

Verdict Expected UPDATE — should update the entry. PASS: Updated as asked.

Why verdict Updated as asked.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 2.3s

1 · TTS said Updated Mango, Raw to 1 large.

2 · Card shown Updated Mango, Raw to 1 large.

3 · App data rows written updated food_log_entry: Mango ×1 (100 g) 60 cal · 0.8g P · 15g C · 0.4g F

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state:

Seed food: mango

{
  "type": "food_item",
  "foodItem": {
    "id": "seed-e-edit-5-mango",
    "name": "mango",
    "brand": null,
    "source": "manual",
    "servingLabel": "serving",
    "servingQuantity": 1,
    "servingUnit": "serving",
    "macrosPerServing": {
      "calories": 180,
      "proteinGrams": 8,
      "fatGrams": 6,
      "carbsGrams": 18
    },
    "aliases": [],
    "timesLogged": 0,
    "createdAt": "2026-07-16T12:00:00.000Z"
  }
}

Seed food history: 1 for seed-e-edit-5-mango at 2026-07-18T10:00:00.000Z

{
  "type": "food_history",
  "foodItemId": "seed-e-edit-5-mango",
  "quantity": 1,
  "loggedAt": "2026-07-18T10:00:00.000Z"
}

snapshot captured 2026-08-02T22:20:08.979Z

6 · Why this food No food-decision explanation was captured for this path.

7 · Lookup method No lookup method was captured for this path.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"How much protein have I logged today?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said You logged 960.3 grams of protein today.

2 · Card shown You logged 960.3 grams of protein today.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:20:20.262Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.8 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"How many calories am I at so far?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I understood the nutrition question, but I could not find a solid number for that food yet. Nothing was logged.

2 · Card shown I understood the nutrition question, but I could not find a solid number for that food yet. Nothing was logged.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:20:43.061Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"What did I eat the day before yesterday?" (difficulty —)0.0s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.0s

1 · TTS said I do not see any food logged yesterday.

2 · Card shown I do not see any food logged yesterday.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:20:54.228Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.8 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"How much fat is on my log today?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I understood the nutrition question, but I could not find a solid number for that food yet. Nothing was logged.

2 · Card shown I understood the nutrition question, but I could not find a solid number for that food yet. Nothing was logged.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:21:05.547Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"What's my remaining protein budget?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said You are about 800.3 grams of protein over today's 160 gram protein target (960.3 logged).

2 · Card shown You are about 800.3 grams of protein over today's 160 gram protein target (960.3 logged).

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:21:16.762Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"Did I log dinner yet?" (difficulty —)0.0s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.0s

1 · TTS said I do not see dinner logged today.

2 · Card shown I do not see dinner logged today.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:21:27.949Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"How many carbs have I had?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I do not see carbs in today's food log.

2 · Card shown I do not see carbs in today's food log.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:21:39.145Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"What did I eat after swimming?" (difficulty —)0.0s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.0s

1 · TTS said No — I do not see after swimming logged today.

2 · Card shown No — I do not see after swimming logged today.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:21:50.369Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"Show me today's meals." (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Opening Food Today.

2 · Card shown Opening Food Today.

4 · UI did navigated to "food" screen

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:22:01.623Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"Am I over my calorie goal?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Yes — you are about 19616 calories over today's 2200 calorie budget (21816 logged).

2 · Card shown Yes — you are about 19616 calories over today's 2200 calorie budget (21816 logged).

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:22:12.914Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"How many calories in the mango I just logged?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Mango has about 60 calories per 100 g.

2 · Card shown Mango has about 60 calories per 100 g.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:22:24.246Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "mango" at 90% confidence.
Matched the common-food list to "Mango" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.5 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"What was my second meal today?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Today you logged Mango (100 g), Blackberries (1 cup), Sunflower butter (1 tbsp), Soy Milk, Unsweetened (one cup (244 g)), Mango (100 g), plus 72 more. Total: 21816 calories, 960.3 g protein, 1387.3 g fat, 1424.9 g carbs.

2 · Card shown Today you logged Mango (100 g), Blackberries (1 cup), Sunflower butter (1 tbsp), Soy Milk, Unsweetened (one cup (244 g)), Mango (100 g), plus 72 more. Total: 21816 calories, 960.3 g protein, 1387.3 g fat, 1424.9 g carbs.

4 · UI did navigated to "food" screen

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:22:35.534Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.8 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"How much soy milk have I logged?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Today you logged 1 one cup (244 g) Soy Milk, Unsweetened (93 calories).

2 · Card shown Today you logged 1 one cup (244 g) Soy Milk, Unsweetened (93 calories).

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:22:46.891Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched the common-food list to "Milk" at 90% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"Calories left for a late snack?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said You are about 19616 calories over today's 2200 calorie budget (21816 logged).

2 · Card shown You are about 19616 calories over today's 2200 calorie budget (21816 logged).

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:22:58.063Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.7 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"Did I already log my usual Built bar?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Yes. Today you logged 1 bar Built Built Puff Bar, 1 bar Built Built Puff Bar.

2 · Card shown Yes. Today you logged 1 bar Built Built Puff Bar, 1 bar Built Built Puff Bar.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:23:09.371Z

6 · Why this food

The lookup compared against planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
Matched your saved foods and aliases to "my usual built bar" at 96% confidence.
Matched your saved foods and aliases to "Built Puff Bar" at 85% confidence.

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.2 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"What's my average fat this week?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I understood the nutrition question, but I could not find a solid number for that food yet. Nothing was logged.

2 · Card shown I understood the nutrition question, but I could not find a solid number for that food yet. Nothing was logged.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:23:20.676Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"How many food logs today?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said You have 77 food entries logged today.

2 · Card shown You have 77 food entries logged today.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:23:31.894Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.9 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

UNVERIFIEDunclassifiedshould answer the query"What did I have mid-morning?" (difficulty —)0.1s

Verdict Expected QUERY — should answer the query. UNVERIFIED: answer key has no expected values

Why verdict answer key has no expected values

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I do not see any food logged mid-morning today.

2 · Card shown I do not see any food logged mid-morning today.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:23:43.187Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.8 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould navigate"Open my food diary." (difficulty —)0.1s

Verdict Expected NAVIGATE — should navigate. PASS: Navigated.

Why verdict Navigated.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Opening Food.

2 · Card shown Opening Food.

4 · UI did navigated to "food" screen

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:23:54.408Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.6 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

PASSshould navigate"Go to calorie goal settings." (difficulty —)0.1s

Verdict Expected NAVIGATE — should navigate. PASS: Navigated.

Why verdict Navigated.

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said Opening Settings.

2 · Card shown Opening Settings.

4 · UI did navigated to "settings" screen

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:24:05.625Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 0.4 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

FAILshould apply the setting"Set my calorie goal to two thousand." (difficulty —)0.1s

Verdict Expected SETTINGS — should apply the setting. FAIL: Reported could-not-confirm (write-confirmation false negative).

Why verdict Reported could-not-confirm (write-confirmation false negative).

My override

Things I think are wrong

Use this when the report passed something that should have failed.

Comprehension

Picked wrong food Wrong quantity Should have asked (over-logged) Shouldn't have asked (over-asked) Wrong action entirely

Execution

Didn't save to DB Wrong data saved No voice/TTS Card missing/wrong No chat

Cosmetic

Wording weird

Other thing wrong (describe below)

Things the report marked wrong but I think are right

Use this when the report failed something that should have passed.

Actually right

Food/action was right Quantity/serving was right Asking was right Not asking was right Answer was right No app data row was expected App data row was right Voice/card/chat was acceptable

Report/grader issue

Report expected the wrong result Evidence was enough to judge Failure reason was not valid

Other thing actually right (describe below)

Pick every reason you think applies. Multiple boxes are allowed because one utterance can be wrong in more than one way. Use the note below for "Other" or extra detail.

Timing 0.1s

1 · TTS said I'm not sure how to help with that yet. Try asking me to log food, a workout, a note, or a to-do.

2 · Card shown I'm not sure how to help with that yet. Try asking me to log food, a workout, a note, or a to-do.

4 · UI did no screen events captured

5 · Seed/context data

Seeded starting state: none listed in the corpus row.

No DB snapshot captured for this utterance; the report cannot prove whether yesterday/personal data existed.

snapshot captured 2026-08-02T22:24:16.835Z

7 · Lookup method

The report captured the app food lookup that ran before or alongside the final decision.
It searched planned foods, your saved foods and aliases, your past food logs, the shared nutrition database, the common-food list, the exercise list, and the workout equipment list.
The lookup evidence took 1.8 seconds to collect.
This lookup is supporting evidence in the report; by itself it does not prove that food was saved.

8 · Candidates → decider no candidate list captured (this path did not run DB resolution — e.g. provisional branded log)

Why the fails happened — comprehension vs execution vs cosmetic

Handled correctly? — by expected action

Accuracy by difficulty

Clarification follow-ups — scored separately

Cosmetic only

System / infra

Latency

Filter — controls the list below

Per-utterance detail