Hard-stool follow-up: user-reported symptoms with deterministic red-flag handoff #70

2026-10-09T08:10:56Z

rockachopa commented

2026-10-09 08:10:56 +00:00

Closes #67

What changed

A hard/lumpy (Bristol 1-2) entry now asks a short follow-up before saving: hard/difficult to pass, soreness or sharp pain afterward, visible red blood, truly black/sticky/tarry versus dark brown, and dehydration/dark urine. Every question starts unanswered, is user-reported, and is never inferred from the photo.

Confirmed visible blood writes the existing blood symptom flag; confirmed truly black/tarry writes the existing blackOrDarkRed flag. Both therefore trigger the same deterministic alert manual logging already uses, and clearing the answer clears the escalation.
Answering "dark brown" (or leaving the question unanswered) never escalates: a draft color value is not evidence.
A mild report gets concise non-diagnostic self-care (hydration, no straining, gentle cleaning, warm sitz bath, external barrier ointment, food-based fiber) plus an explicit "general self-care, not a diagnosis — your photo cannot rule out bleeding or another cause" note and a seek-care line for severe/worsening pain, faintness, very low urine, or any blood.
The follow-up appears only for Type 1-2 forms, on the existing safety step, next to the unchanged red-flag checkboxes. Non-hard forms are unaffected.
The photo suggestion surface is unchanged and provably symptom-free: the acceptance test asserts the suggestion never mentions soreness, hydration, tarry, difficulty passing, or blood.

Evidence (this branch, commit `4cc5f4f`)

npm test — 104/104 pass from a clean checkout (was 92; +9 follow-up unit, +3 contract/release-demo assertions).
npm run test:followup — PASS (tests/followup.acceptance.mjs, real browser at 390x844).
npm run test:ui, test:photo, test:mobile-capture, test:sleek, test:manual-fallback, test:provisional — all PASS (no regressions).
npm run check:syntax — PASS, including src/followup.js and the new acceptance script.
scripts/record_release_demo.mjs — recorded end to end against the working app (/tmp/timmy-demo-verify.webm, 71.8s, no browser errors) exercising the follow-up block, the mild self-care output, and the black/tarry → deterministic alert handoff with synthetic data only.
Visual acceptance artifact: artifacts/followup-hard-stool-mobile.png.

RED/GREEN discipline

The unit contract was written first and observed failing (ERR_MODULE_NOT_FOUND: src/followup.js), then implemented to green. The browser acceptance was written next and observed failing (waiting for locator('[data-followup-block]') on the unwired app), then the app was wired to green. The suite also caught a real bug during that cycle: unchecking the mirrored red flag on a non-hard form left a stale alert because the refresh returned early when the follow-up block was absent.

Deliberate scope limit

Mild answers (difficulty passing, soreness, dehydration) drive the guidance but are not persisted: saving them would change the ledger schema, which #35/#61 already own. Follow-up as a separate ticket rather than a schema change on a safety PR.

Boundary

No change to the medical/non-diagnostic boundary: analysis still suggests visible form, broad color, and quality only; the follow-up adds user-reported answers, adds no new escalation keys, and reuses the canonical urgent copy from src/domain.js. No real medical images, no model weights, no infrastructure purchase, no production service change.

Closes #67 ## What changed A hard/lumpy (Bristol 1-2) entry now asks a short follow-up before saving: hard/difficult to pass, soreness or sharp pain afterward, visible red blood, truly black/sticky/tarry versus dark brown, and dehydration/dark urine. Every question starts unanswered, is user-reported, and is never inferred from the photo. - Confirmed **visible blood** writes the existing `blood` symptom flag; confirmed **truly black/tarry** writes the existing `blackOrDarkRed` flag. Both therefore trigger the same deterministic alert manual logging already uses, and clearing the answer clears the escalation. - Answering "dark brown" (or leaving the question unanswered) never escalates: a draft `color` value is not evidence. - A mild report gets concise non-diagnostic self-care (hydration, no straining, gentle cleaning, warm sitz bath, external barrier ointment, food-based fiber) plus an explicit "general self-care, not a diagnosis — your photo cannot rule out bleeding or another cause" note and a seek-care line for severe/worsening pain, faintness, very low urine, or any blood. - The follow-up appears only for Type 1-2 forms, on the existing safety step, next to the unchanged red-flag checkboxes. Non-hard forms are unaffected. - The photo suggestion surface is unchanged and provably symptom-free: the acceptance test asserts the suggestion never mentions soreness, hydration, tarry, difficulty passing, or blood. ## Evidence (this branch, commit 4cc5f4f) - `npm test` — 104/104 pass from a clean checkout (was 92; +9 follow-up unit, +3 contract/release-demo assertions). - `npm run test:followup` — PASS (`tests/followup.acceptance.mjs`, real browser at 390x844). - `npm run test:ui`, `test:photo`, `test:mobile-capture`, `test:sleek`, `test:manual-fallback`, `test:provisional` — all PASS (no regressions). - `npm run check:syntax` — PASS, including `src/followup.js` and the new acceptance script. - `scripts/record_release_demo.mjs` — recorded end to end against the working app (`/tmp/timmy-demo-verify.webm`, 71.8s, no browser errors) exercising the follow-up block, the mild self-care output, and the black/tarry → deterministic alert handoff with synthetic data only. - Visual acceptance artifact: `artifacts/followup-hard-stool-mobile.png`. ## RED/GREEN discipline The unit contract was written first and observed failing (`ERR_MODULE_NOT_FOUND: src/followup.js`), then implemented to green. The browser acceptance was written next and observed failing (`waiting for locator('[data-followup-block]')` on the unwired app), then the app was wired to green. The suite also caught a real bug during that cycle: unchecking the mirrored red flag on a non-hard form left a stale alert because the refresh returned early when the follow-up block was absent. ## Deliberate scope limit Mild answers (difficulty passing, soreness, dehydration) drive the guidance but are **not** persisted: saving them would change the ledger schema, which #35/#61 already own. Follow-up as a separate ticket rather than a schema change on a safety PR. ## Boundary No change to the medical/non-diagnostic boundary: analysis still suggests visible form, broad color, and quality only; the follow-up adds user-reported answers, adds no new escalation keys, and reuses the canonical urgent copy from `src/domain.js`. No real medical images, no model weights, no infrastructure purchase, no production service change.

rockachopa added 1 commit 2026-10-09 08:10:57 +00:00

feat: user-reported hard-stool follow-up with deterministic red-flag handoff (#67 )

Quality gates / quality (pull_request) Successful in 2m53s

Details

4cc5f4f7b1

Field finding #67: a hard/lumpy (Bristol 1-2) entry now asks a short follow-up
before saving — difficult to pass, soreness, visible red blood, truly
black/sticky/tarry versus dark brown, and dehydration. Every answer is
user-reported and starts unanswered; nothing is inferred from the photo, and
answering "dark brown" never silently escalates as black stool.

A confirmed red-flag answer writes the same symptom flag manual logging already
uses, so the existing deterministic escalation is the only escalation path.
Mild reports get concise non-diagnostic self-care that states the photo cannot
rule out bleeding, and it is withdrawn whenever a red flag is confirmed.

- src/followup.js: pure follow-up contract (9 unit tests, tests/followup.test.js)
- app.js + styles.css: follow-up block on the safety step for Type 1-2 entries only
- tests/followup.acceptance.mjs: browser acceptance incl. that the AI suggestion
  surface still contains no symptom wording
- CI + release builder gates: npm run test:followup, syntax checks, manifest gate
- scripts/record_release_demo.mjs: records the real follow-up, mild guidance, and
  the black/tarry alert handoff with synthetic data

rockachopa closed this pull request

2026-10-09 12:07:15 +00:00

Pull request closed

Please reopen this pull request to perform a merge.

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: stackchain/timmy-talking-turd#70