[BigSimmo]

User description

Requested by Josh · project thread

Summary

Before: each kind of Clinical record (DSM, Differentials, Specifiers, Formulation, Therapy, Forms, Medication, Services) had its own layout. Lists were walls of coloured pills. On phones, several things were broken: Compare sheet buttons off screen, raw tags like crisis_high on service records, the bottom box squashed at large text, and cut-off prompts.

After: every record follows one layout, modelled on the Factsheet record: name, an At a glance card in full, then cautions in full, plain sections, and sources last. Lists use grey words and status shapes instead of coloured pills. The phone faults are fixed.

Builds the "Build now" list from the Clinical pages audit (/mnt/project-files/clinical-audit-2026-10-13/README.md) that Josh approved:

• 1: Compare sheet footer is pinned on phones (all six compare sheets).
• 2: service records hide or map raw machine tags, and show the phone number once with Call.
• 3: bottom box sizing at large text. CSS only; 0 differing pixels at normal size.
• 4: Therapy Avoid/modify is no longer clamped. Safety & cautions sits right after the summary.
• 5: one Documents failure message with Retry.
• 6: shared record layout (clinical-record-panels.tsx) across the eight record types.
• 8: status shapes and grey words on Differentials, Services, Sources and Medication results. Labels, categories and ranking are unchanged.
• 10, 11: side bars removed; main buttons are Clinical blue.
• 12: compact rows on Formulation, Recommend and Calculators.
• 13: A to Z strip on Dictionary, and on Services when sorted A to Z.
• 14: Dictionary rows tidied.
• 15: Psychiatry Continue moved first; duplicate First Nations entry removed; Tools page is now "All tools".
• 17: Browse links on empty mode homes.
• 18: shorter prompts.
• 21: plain Calculators notice.
• Also fixes the medication Access panel cutting the Formulation value to "Tablets (12".

Areas touched: App experience, Design system, Clinical reference content, Delivery and assurance, Personal practice (plus 1 file not yet placed in any area)

RAG impact: no retrieval behaviour change. No ranking-protected file is touched; src/lib/medications.ts scoring and labels are unchanged.

Verification

• Checks run (local, demo mode):
  • npm run typecheck: recorded a pass for "typecheck:internal"
  • eslint on changed files: clean
  • prettier --check: All matched files use Prettier code style!
  • changed unit test files: Tests 347 passed (347)
• Browser specs that were edited (ui-tools, ui-overlap, ui-accessibility, ui-formulation-result-cards, ui-tools-show-all) have not been run locally. The full browser suite will run on this draft.
• Before and after screenshots: /mnt/project-files/clinical-audit-2026-10-13/built/{faults,records,lists}/

Risk and rollout

• Risk: clinical display only. No clinical wording, facts, urgency categories or ranking change. The 13YARN tag labels need a First Nations governance check.
• Rollback: revert this PR.
• Provider or production effects: none.

Notes

• Stays DRAFT until the full browser suite is green and Josh sends it live.
• The approved wording changes (audit items 19 and 20) are in a separate stacked draft.
• Header, band, record top bar, tabs and Back were not touched; the header thread owns them.

🤖 Generated with Claude Code

https://claude.ai/code/session_015k5m5pXngW8LaerzAGumW8

Generated by Claude Code

---

Summary by cubic

Unifies all eight Clinical record types onto one shared layout and fixes the phone faults from the 13 Oct audit: compare sheet buttons stay on screen, service records hide raw machine tags, the bottom box stops squashing at large text, and prompts are shortened.

Record layout and lists

• Every record (DSM, Differentials, Specifiers, Formulation, Therapy, Forms, Medication, Services) now follows the Factsheet model: name, At a glance card, cautions in full, plain sections, sources last.
• Lists (Differentials, Services, Sources, Medication) use grey words with a small status shape; service chip shapes are keyed to availability and confidence labels rather than tone, so acuity and confidence rows keep plain grey.
• The service Call bar is a compact phone button that keeps the visible number.
• Formulation, Recommend and Calculators lists use compact rows; Dictionary gets a phone A to Z strip and Services gets one only when sorted A to Z, each keeping its keyboard position when filtering shortens it.
• Therapy cautions and safety sections are no longer clamped; a therapy record's Format card keeps the session length beside the time required; the medication Formulation row shows the whole field, strengths and forms included.

Other fixes

• A failed documents search with nothing listed shows one message with Retry, wired to re-run the current query rather than an older failed one; the red error box stays only when earlier matches are still listed, so they cannot pass for the new query's results.
• Empty mode homes show two Browse links, keeping the idle phone home on one screen; Psychiatry's Continue moves first and the duplicate First Nations entry is removed; Tools becomes "All tools".
• Map and Related are reachable from every diagnosis tab and start the new tab at its heading so the top bar and section picker come back into view; main buttons and compare actions use the Clinical blue primary button; side bars are removed from records.

Display-only changes: no clinical wording, facts, urgency categories or ranking change. The 13YARN tag labels need a First Nations governance check.

^{Written for commit 7d320bb. Summary will update on new commits.}

---

CodeAnt-AI Description

Standardize clinical records and fix mobile search and compare flows

What Changed

• Clinical records now share a consistent layout, with full “At a glance” details, cautions near the top, and sources and related links at the end. Repeated facts and “Tap for detail” prompts are removed.
• Clinical result lists use compact rows and quieter status labels. Service records hide or translate machine-style tags and avoid repeating contact numbers; phone buttons keep the same call action.
• Compare controls stay pinned while lists scroll in phone sheets. Dictionary and alphabetically sorted service lists offer letter jumps, and composer controls remain usable with large text.
• Failed document searches show one message with the failure reason and a Retry option, instead of stacking a separate error message.
• Empty Clinical home pages offer links to browse relevant pages; the tools directory is labeled “All tools.”

Impact

✅ Full safety cautions visible without extra taps
✅ Compare controls stay reachable while phone lists scroll
✅ Failed document searches can be retried in place

💡 Usage Guide

Checking Your Pull Request

Every time you make a pull request, our system automatically looks through it. We check for security issues, mistakes in how you're setting up your infrastructure, and common code problems. We do this to make sure your changes are solid and won't cause any trouble later.

Talking to CodeAnt AI

Got a question or need a hand with something in your pull request? You can easily get in touch with CodeAnt AI right here. Just type the following in a comment on your pull request, and replace "Your question here" with whatever you want to ask:

@codeant-ai ask: Your question here

This lets you have a chat with CodeAnt AI about your pull request, making it easier to understand and improve your code.

Example

@codeant-ai ask: Can you suggest a safer alternative to storing this secret?

Preserve Org Learnings with CodeAnt

You can record team preferences so CodeAnt AI applies them in future reviews. Reply directly to the specific CodeAnt AI suggestion (in the same thread) and replace "Your feedback here" with your input:

@codeant-ai: Your feedback here

This helps CodeAnt AI learn and adapt to your team's coding style and standards.

Example

@codeant-ai: Do not flag unused imports.

Retrigger review

Ask CodeAnt AI to review the PR again, by typing:

@codeant-ai: review

Check Your Repository Health

To analyze the health of your code repository, visit our dashboard at https://app.codeant.ai. This tool helps you identify potential issues and areas for improvement in your codebase, ensuring your repository maintains high standards of code health.

[qodo-code-review]

Qodo Fixer

No findings are available for this PR yet. Findings appear here once Qodo has reviewed the PR.

[gitar-bot]

Gitar has auto-approved this PR (configure)

[gitar-bot]

Gitar has auto-approved this PR (configure)

[qodo-code-review]

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 🔗 Cross-repo conflicts (0) 📜 Skill insights (0)

No findings for this group

No findings for this group

No findings for this group

Resolved findings

1. Clinicians see closed services as urgent ✓ Resolved

Description

serviceChipShape converts every danger chip to the urgent triangle and every warning chip to
the unrecorded ring, even when the label is not urgent or unverified. This affects closed,
superseded, and low-confidence services as well as moderate-acuity, planned, temporarily
unavailable, and medium-confidence services, so their status is communicated with the wrong clinical
meaning.

Code

src/components/services/services-navigator-page.tsx[R95-97]

+  // A closed, superseded or low-confidence service must not look like a plain region word.
+  if (chip.tone === "danger") return "urgent";
+  if (chip.tone === "warning") return "unrecorded";

Relevance

●●● Strong

Tone-only mapping conflates distinct clinical statuses with urgent or unrecorded meanings,
undermining the PR’s status-shape intent.

PR-#3463

ⓘ Recommendations generated based on similar findings in past PRs

Evidence

The changed fallback assigns shapes based only on tone, while the service mapper emits those tones
for multiple unrelated meanings. The status-tag component defines the urgent shape as
overdue/emergent/crisis and the unrecorded shape as not recorded/not verified, proving that the
fallback communicates incorrect semantics for several generated service labels.

src/components/services/services-navigator-page.tsx[90-98]
src/lib/service-catalog-mapper.ts[50-88]
src/components/mode-kit/status-tag.tsx[8-20]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
`serviceChipShape` uses generic `danger` and `warning` tones to select the urgent and unrecorded shapes. Service status generation uses those tones for several meanings that do not match those shapes, including closed, superseded, low-confidence, moderate-acuity, planned, temporarily unavailable, and medium-confidence statuses.

## Fix Focus Areas
- src/components/services/services-navigator-page.tsx[90-98]
- src/lib/service-catalog-mapper.ts[50-88]

## Recommended Fix
Remove the generic tone fallbacks or map each known status label to a semantically correct shape. Reserve `urgent` for crisis/emergency/high-acuity labels and `unrecorded` for explicitly unverified states; render closed, superseded, planned, temporarily unavailable, confidence, and moderate-acuity labels as plain words unless a dedicated shape accurately represents them. Add regression coverage for representative labels from each tone.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

2. Service letter jumps can mislead readers ✗ Dismissed

Description

ServicesNavigatorPage builds alphabetEntries from displayedMatches without checking
sortValue, and alphabetStripEntries accepts lists with a few out-of-order rows. When
relevance-ranked results happen to be mostly alphabetical, the phone strip appears even though the
list is not sorted A to Z, so a letter jump can land before another result under that same letter.

Code

src/components/services/services-navigator-page.tsx[R538-541]

+  const alphabetEntries = useMemo(
+    () =>
+      alphabetStripEntries(
+        displayedMatches,

Relevance

●●● Strong

Sorting-dependent UI is expected to gate alphabetic navigation and avoid misleading jumps from
relevance-ranked results.

PR-#2407
PR-#2253

ⓘ Recommendations generated based on similar findings in past PRs

Evidence

The navigator preserves ranked search order until the user selects alphabetical sorting, but its new
strip calculation runs for either sort mode. The entry helper permits up to two out-of-order rows
(or five percent of the list), and the strip scrolls to the generated row target.

src/components/services/services-navigator-page.tsx[499-506]
src/components/services/services-navigator-page.tsx[533-546]
src/lib/result-sort.ts[7-10]
src/components/mode-kit/alphabet-strip.tsx[253-278]
src/components/mode-kit/alphabet-strip.tsx[99-109]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
The Services phone letter strip can appear for relevance-ranked results that happen to be nearly alphabetical, making its letter jumps unreliable.
## Fix Focus Areas
- src/components/services/services-navigator-page.tsx[535-546]
## Recommended Fix
Only build and render alphabet entries when the selected sort mode is A to Z. Include the sort mode in the memo dependencies; keep the existing entry helper for alphabetically sorted results.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

3. Service list cards no longer show the phone number ✓ Resolved

Description

The Call link on each navigator card now holds only a phone icon. Its accessible name is `Call
${service.title}, and the number exists only in the title` attribute. Touch users never see
tooltips and screen readers don't announce title as part of the name, so on phones the number
can't be read or written down before dialling. The old card showed Call {number} as visible text.

Code

src/components/services/services-navigator-page.tsx[R280-285]

+            aria-label={`Call ${service.title}`}
+            title={`Call ${call.display}`}
            data-testid={`service-call-${service.slug}`}
-            className="relative z-10 inline-flex min-h-12 min-w-0 flex-1 items-center justify-center gap-2 rounded-lg border border-[color:var(--clinical-accent)] bg-[color:var(--clinical-accent)] px-3 text-sm font-bold text-[color:var(--clinical-accent-contrast)] hover:bg-[color:var(--clinical-accent-hover)] focus-visible:outline focus-visible:outline-2 focus-visible:outline-offset-2 focus-visible:outline-[color:var(--focus)] sm:flex-none"
+            className="relative z-10 inline-flex min-h-12 min-w-12 shrink-0 items-center justify-center rounded-lg border border-[color:var(--border)] bg-[color:var(--surface)] text-[color:var(--clinical-accent)] hover:bg-[color:var(--surface-subtle)] focus-visible:outline focus-visible:outline-2 focus-visible:outline-offset-2 focus-visible:outline-[color:var(--focus)]"
          >
            <Phone className="h-4 w-4 shrink-0" aria-hidden />
-            <span className="truncate tabular-nums">Call {call.display}</span>

Relevance

●●● Strong

Recent phone-contact and accessibility findings favor exposing actionable contact details beyond
hover-only or tooltip presentation.

PR-#2609
PR-#3156

ⓘ Recommendations generated based on similar findings in past PRs

Evidence

The removed <span>Call {call.display}</span> was the only place the number appeared on the card.
The new link has only aria-label (no number), title (the number), and an aria-hidden icon. The
updated test checks the number only through the title attribute.

src/components/services/services-navigator-page.tsx[277-287]
tests/services-call-button.dom.test.tsx[130-142]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
The compact Call button on service cards hides the phone number. It is only in a `title` tooltip, which touch devices never show, and it is missing from the accessible name.

## Fix Focus Areas
- src/components/services/services-navigator-page.tsx[277-287]
- tests/services-call-button.dom.test.tsx[130-142]

## Recommended Fix
Put the number back in `aria-label` (`Call ${call.display}, ${service.title}`). Also show `call.display` as visible tabular text next to the icon, or on its own line in the tag row, so the button stays compact but the number can still be read. Update the test to check the visible text.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

View resolved (1)

4. Closed or unverified services look like plain region tags ✗ Dismissed

Description

serviceChipShape decides the shape by matching the chip label against regexes and ignores
chip.tone entirely. So danger- and warning-tone chips from buildStatusChips, such as "Closed",
"Superseded", "Temporarily unavailable", "Planned — not open", "Verify before use" and "Low
confidence", get none. A closed or unvetted service in the navigator list then looks exactly like
an acuity or region word, even though the PR aims to keep colour where a row differs.

Code

src/components/services/services-navigator-page.tsx[R90-95]

+function serviceChipShape(chip: ServiceStatusChip): StatusTagShape {
+  const label = chip.label?.toLowerCase() ?? "";
+  if (/\b(crisis|urgent|emergency)\b/.test(label)) return "urgent";
+  if (/unverified|not verified|unconfirmed/.test(label)) return "unrecorded";
+  if (/^active$/.test(label.trim())) return "done";
+  return "none";

Relevance

●●● Strong

Recent status-shape precedent preserves explicit warning and danger semantics instead of rendering
them as neutral labels.

PR-#3489

ⓘ Recommendations generated based on similar findings in past PRs

Evidence

The mapper creates these chips with explicit danger/warning tones. The new shape function only
matches crisis/urgent/emergency, unverified/not verified/unconfirmed, or exactly "Active", so every
other cautionary label falls through to none. The old chipTone passed danger and warning
straight through to a coloured chip.

src/lib/service-catalog-mapper.ts[75-90]
src/lib/service-catalog-mapper.ts[223-251]
src/components/services/services-navigator-page.tsx[90-96]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
`serviceChipShape` ignores `chip.tone`, so chips the catalogue marks as danger or warning (Closed, Superseded, Temporarily unavailable, Planned — not open, Verify before use, Low confidence) show as plain grey words with no status shape.

## Fix Focus Areas
- src/components/services/services-navigator-page.tsx[90-96]
- src/lib/service-catalog-mapper.ts[75-90]

## Recommended Fix
Keep the crisis label check. Then map by tone: `danger` -> "urgent", `warning` -> "soon" (or "unrecorded" for verification chips), `success` with an "Active" label -> "done". Return "none" only for `info`/`neutral` chips. Add a test for a Closed service and for a "Verify before use" chip.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

Context sources

✅ Compliance rules (platform): 9 rules

✅ Cross-repo context — repo relationships

  Explored: repo: BigSimmo/Medications (sha: f0fe3d04) — View relationship

✅ REVIEW.md

Review mode: Auto: 🧠 Deep: Multiple independent behavioral changes span authorization, navigation, UI, and RAG evidence paths.

✅ Fix-loop round: fix loop

Tip of the day

💡 Did you know, you can turn these tips off under Display > Comment parts

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗

Previous reviews

Review updated until commit 7d320bb 🧠 Deep

Results up to commit 25e404e 🧠 Deep

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0) 🎨 UX issues (0) 🔗 Cross-repo conflicts (0) 📜 Skill insights (0)

No findings for this group

No findings for this group

No findings for this group

Resolved findings

1. Closed or unverified services look like plain region tags ✗ Dismissed

Description

serviceChipShape decides the shape by matching the chip label against regexes and ignores
chip.tone entirely. So danger- and warning-tone chips from buildStatusChips, such as "Closed",
"Superseded", "Temporarily unavailable", "Planned — not open", "Verify before use" and "Low
confidence", get none. A closed or unvetted service in the navigator list then looks exactly like
an acuity or region word, even though the PR aims to keep colour where a row differs.

Code

src/components/services/services-navigator-page.tsx[R90-95]

+function serviceChipShape(chip: ServiceStatusChip): StatusTagShape {
+  const label = chip.label?.toLowerCase() ?? "";
+  if (/\b(crisis|urgent|emergency)\b/.test(label)) return "urgent";
+  if (/unverified|not verified|unconfirmed/.test(label)) return "unrecorded";
+  if (/^active$/.test(label.trim())) return "done";
+  return "none";

Relevance

●●● Strong

Recent status-shape precedent preserves explicit warning and danger semantics instead of rendering
them as neutral labels.

PR-#3489

ⓘ Recommendations generated based on similar findings in past PRs

Evidence

The mapper creates these chips with explicit danger/warning tones. The new shape function only
matches crisis/urgent/emergency, unverified/not verified/unconfirmed, or exactly "Active", so every
other cautionary label falls through to none. The old chipTone passed danger and warning
straight through to a coloured chip.

src/lib/service-catalog-mapper.ts[75-90]
src/lib/service-catalog-mapper.ts[223-251]
src/components/services/services-navigator-page.tsx[90-96]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
`serviceChipShape` ignores `chip.tone`, so chips the catalogue marks as danger or warning (Closed, Superseded, Temporarily unavailable, Planned — not open, Verify before use, Low confidence) show as plain grey words with no status shape.

## Fix Focus Areas
- src/components/services/services-navigator-page.tsx[90-96]
- src/lib/service-catalog-mapper.ts[75-90]

## Recommended Fix
Keep the crisis label check. Then map by tone: `danger` -> "urgent", `warning` -> "soon" (or "unrecorded" for verification chips), `success` with an "Active" label -> "done". Return "none" only for `info`/`neutral` chips. Add a test for a Closed service and for a "Verify before use" chip.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

2. Service list cards no longer show the phone number ✓ Resolved

Description

The Call link on each navigator card now holds only a phone icon. Its accessible name is `Call
${service.title}, and the number exists only in the title` attribute. Touch users never see
tooltips and screen readers don't announce title as part of the name, so on phones the number
can't be read or written down before dialling. The old card showed Call {number} as visible text.

Code

src/components/services/services-navigator-page.tsx[R280-285]

+            aria-label={`Call ${service.title}`}
+            title={`Call ${call.display}`}
            data-testid={`service-call-${service.slug}`}
-            className="relative z-10 inline-flex min-h-12 min-w-0 flex-1 items-center justify-center gap-2 rounded-lg border border-[color:var(--clinical-accent)] bg-[color:var(--clinical-accent)] px-3 text-sm font-bold text-[color:var(--clinical-accent-contrast)] hover:bg-[color:var(--clinical-accent-hover)] focus-visible:outline focus-visible:outline-2 focus-visible:outline-offset-2 focus-visible:outline-[color:var(--focus)] sm:flex-none"
+            className="relative z-10 inline-flex min-h-12 min-w-12 shrink-0 items-center justify-center rounded-lg border border-[color:var(--border)] bg-[color:var(--surface)] text-[color:var(--clinical-accent)] hover:bg-[color:var(--surface-subtle)] focus-visible:outline focus-visible:outline-2 focus-visible:outline-offset-2 focus-visible:outline-[color:var(--focus)]"
          >
            <Phone className="h-4 w-4 shrink-0" aria-hidden />
-            <span className="truncate tabular-nums">Call {call.display}</span>

Relevance

●●● Strong

Recent phone-contact and accessibility findings favor exposing actionable contact details beyond
hover-only or tooltip presentation.

PR-#2609
PR-#3156

ⓘ Recommendations generated based on similar findings in past PRs

Evidence

The removed <span>Call {call.display}</span> was the only place the number appeared on the card.
The new link has only aria-label (no number), title (the number), and an aria-hidden icon. The
updated test checks the number only through the title attribute.

src/components/services/services-navigator-page.tsx[277-287]
tests/services-call-button.dom.test.tsx[130-142]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
The compact Call button on service cards hides the phone number. It is only in a `title` tooltip, which touch devices never show, and it is missing from the accessible name.

## Fix Focus Areas
- src/components/services/services-navigator-page.tsx[277-287]
- tests/services-call-button.dom.test.tsx[130-142]

## Recommended Fix
Put the number back in `aria-label` (`Call ${call.display}, ${service.title}`). Also show `call.display` as visible tabular text next to the icon, or on its own line in the tag row, so the button stays compact but the number can still be read. Update the test to check the visible text.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

3. Service letter jumps can mislead readers ✗ Dismissed

Description

ServicesNavigatorPage builds alphabetEntries from displayedMatches without checking
sortValue, and alphabetStripEntries accepts lists with a few out-of-order rows. When
relevance-ranked results happen to be mostly alphabetical, the phone strip appears even though the
list is not sorted A to Z, so a letter jump can land before another result under that same letter.

Code

src/components/services/services-navigator-page.tsx[R538-541]

+  const alphabetEntries = useMemo(
+    () =>
+      alphabetStripEntries(
+        displayedMatches,

Relevance

●●● Strong

Sorting-dependent UI is expected to gate alphabetic navigation and avoid misleading jumps from
relevance-ranked results.

PR-#2407
PR-#2253

ⓘ Recommendations generated based on similar findings in past PRs

Evidence

The navigator preserves ranked search order until the user selects alphabetical sorting, but its new
strip calculation runs for either sort mode. The entry helper permits up to two out-of-order rows
(or five percent of the list), and the strip scrolls to the generated row target.

src/components/services/services-navigator-page.tsx[499-506]
src/components/services/services-navigator-page.tsx[533-546]
src/lib/result-sort.ts[7-10]
src/components/mode-kit/alphabet-strip.tsx[253-278]
src/components/mode-kit/alphabet-strip.tsx[99-109]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
The Services phone letter strip can appear for relevance-ranked results that happen to be nearly alphabetical, making its letter jumps unreliable.
## Fix Focus Areas
- src/components/services/services-navigator-page.tsx[535-546]
## Recommended Fix
Only build and render alphabet entries when the selected sort mode is A to Z. Include the sort mode in the memo dependencies; keep the existing entry helper for alphabetically sorted results.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

Results up to commit 32ac289 🚀 Fast

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0) 🎨 UX issues (0) 🔗 Cross-repo conflicts (0) 📜 Skill insights (0)

No findings for this group

No findings for this group

No findings for this group

Resolved findings

1. Clinicians see closed services as urgent ✓ Resolved

Description

serviceChipShape converts every danger chip to the urgent triangle and every warning chip to
the unrecorded ring, even when the label is not urgent or unverified. This affects closed,
superseded, and low-confidence services as well as moderate-acuity, planned, temporarily
unavailable, and medium-confidence services, so their status is communicated with the wrong clinical
meaning.

Code

src/components/services/services-navigator-page.tsx[R95-97]

+  // A closed, superseded or low-confidence service must not look like a plain region word.
+  if (chip.tone === "danger") return "urgent";
+  if (chip.tone === "warning") return "unrecorded";

Relevance

●●● Strong

Tone-only mapping conflates distinct clinical statuses with urgent or unrecorded meanings,
undermining the PR’s status-shape intent.

PR-#3463

ⓘ Recommendations generated based on similar findings in past PRs

Evidence

The changed fallback assigns shapes based only on tone, while the service mapper emits those tones
for multiple unrelated meanings. The status-tag component defines the urgent shape as
overdue/emergent/crisis and the unrecorded shape as not recorded/not verified, proving that the
fallback communicates incorrect semantics for several generated service labels.

src/components/services/services-navigator-page.tsx[90-98]
src/lib/service-catalog-mapper.ts[50-88]
src/components/mode-kit/status-tag.tsx[8-20]

Agent prompt

The issue below was found during a code review. Follow the provided context and guidance below and implement a solution

## Issue description
`serviceChipShape` uses generic `danger` and `warning` tones to select the urgent and unrecorded shapes. Service status generation uses those tones for several meanings that do not match those shapes, including closed, superseded, low-confidence, moderate-acuity, planned, temporarily unavailable, and medium-confidence statuses.

## Fix Focus Areas
- src/components/services/services-navigator-page.tsx[90-98]
- src/lib/service-catalog-mapper.ts[50-88]

## Recommended Fix
Remove the generic tone fallbacks or map each known status label to a semantically correct shape. Reserve `urgent` for crisis/emergency/high-acuity labels and `unrecorded` for explicitly unverified states; render closed, superseded, planned, temporarily unavailable, confidence, and moderate-acuity labels as plain words unless a dedicated shape accurately represents them. Add regression coverage for representative labels from each tone.

ⓘ Copy this prompt and use it to remediate the issue with your preferred AI generation tools

---

[cubic-dev-ai]

Plan monthly usage limit reached ↗
Resets Oct 16

Get more usage

[gitar-bot]

Gitar has auto-approved this PR (configure)

[qodo-code-review]

Code review by qodo was updated up to the latest commit 32ac289

[gitar-bot]

Gitar has auto-approved this PR (configure)

[qodo-code-review]

Code review by qodo was updated up to the latest commit 7d320bb

[gitar-bot]

CI failed: One Playwright UI test failed in the Care Plan mockup suite: a page-wide test-id locator matched two identity-uncertain alerts, causing a strict-mode violation. Relation to this PR is unclear, so it is not attributed to the change.

Overview

Single unique error template across 7 analyzed logs for commit 7d320bb. The only blocking failure is one Playwright test (tests/ui-care-plan-mockup.spec.ts:1612) in the chromium-mockups project; 67 tests passed, 1 skipped. Overall: a test-locator ambiguity, not a confirmed application regression.

Failures

Ambiguous test-id locator in Care Plan identity-uncertain assertion (confidence: high)

• Type: test
• Affected jobs: 115250502653
• Related to change: unclear
• Root cause: getByTestId("care-plan-identity-uncertain") resolved to two elements (one inside the Rowan Sample clinical snapshot region, one elsewhere), so the toContainText assertion failed under strict mode. The log shows a locator ambiguity; the supplied context does not show edits to this test or its implementation establishing that this PR caused the duplicate.
• Suggested fix: Scope the assertion to the intended region (e.g. the Rowan Sample clinical snapshot region) or select the specific alert whose text should contain "Return to search". If both alerts are unintended duplicates, make the test id unique in the rendering.

Summary

• Change-related failures: None confirmed. The one failure is marked unclear.
• Infrastructure/flaky failures: None identified. Chromium install retries occurred but the run proceeded to tests.
• Recommended action: Scope the Care Plan identity-uncertain assertion to its region or fix the duplicated test id, then rerun the mockup suite to confirm.

Code Review ✅ Approved 4 closed / 4 findings

🟡 Medium risk · Reworking clinical record layouts and mobile search flows could make safety or referral details harder to find if presentation changes mislead users.

Unifies Clinical record layouts and phone list behaviour across the eight record types, including service chip shapes keyed to availability and confidence labels. Resolved the stale degraded-zero Retry issue and the tone-based chip shape fallback, which now uses each chip's own label. Two A to Z strip row-skipping issues were closed without a fix in this PR. No open issues remain.

✅ 4 closed

✅ Bug: Degraded-zero Retry can re-run a stale failed query

📄 src/components/ClinicalDashboard.tsx:3499-3504
onRetrySearch is now passed whenever the mode is documents and query.trim() is set, so it also shows on degraded zero results where no error occurred. It still calls ask(lastFailedQuery ?? query). Nothing clears lastFailedQuery after a successful search or a mode switch: it is only reset at lines 1740, 2328, 2368 and 2665, and the answer-search catch sets it at 1968. Example: an answer search for "X" fails, then the user runs a documents search for "Y" that comes back degraded with zero results. Pressing Retry searches "X" again instead of "Y". The fix is to use lastFailedQuery only when the search actually failed, or to retry the current query.

✅ Bug: Tone fallback gives acuity and confidence chips the wrong status shapes

📄 src/components/services/services-navigator-page.tsx:84-98
The new chip.tone fallback in serviceChipShape covers more chips than the closed, superseded and low-confidence cases its comment names. acuityChipTone in service-catalog-mapper.ts sets "High acuity" to danger and "Moderate acuity" to warning. confidenceTone sets "Medium confidence" to warning and "Low confidence" to danger. Bundled chips such as "24/7 Acute Inpatient" (danger) and "Forensic & Acute Inpatient" (warning) are affected too. As a result, "High acuity" and "Low confidence" get the same red triangle as "Crisis / urgent". "Moderate acuity" and "Forensic & Acute Inpatient" get the dashed ring, which status-tag.tsx defines as "not recorded, not verified", so a verified service looks unverified. This also contradicts the docstring just above, which says acuity chips stay plain grey words. A fix is to match the availability and confidence labels the comment means (closed, superseded, temporarily unavailable, planned, low or medium confidence) and leave every other tone as none.

Closed: Edge Case: A to Z strip skips the first row after a stray, so a letter can vanish

📄 src/components/mode-kit/alphabet-strip.tsx:250-264
alphabetStripEntries decides whether a row is in place by comparing it with its raw neighbours (letters[index - 1] and letters[index + 1]), even when the neighbour is the stray that was just skipped. Take [G1, W(stray), H1, H2]. W is skipped as intended. H1 is then also treated as out of place because the letter before it is W > H, so the strip's H jumps to H2 and scrolls past H1. If H1 is the only H row ([G1, W, H1, I1]), H drops out of the strip altogether. The JSDoc's own case, a row like "WACHS Goldfields" filed under G, triggers this whenever the stray sits just before the next letter's first row. To fix it, compare each row with the last in-place letter instead of the raw previous item, and keep the forward check only for spotting the stray itself.

Closed: Edge Case: Low stray row still drops the previous letter's first row

📄 src/components/mode-kit/alphabet-strip.tsx:255-264
The fix replaced the raw previous row with the last kept row, but after is still the raw next row. When a stray filed too early comes right after a letter's first row, the real row fails letter <= after and gets skipped. Example: ["Goldfields", "Hedland", "Albany", "Inner City"] (G H A I). Hedland fails because H > A, then Albany fails because G > A, so the strip shows only G and I, H is lost, and the list is charged with two strays instead of one. This is the mirror image of the case this commit fixed. Suggested fix: when picking after, skip upcoming rows whose letter is below the last kept letter, since those are strays too. Add a test for this case.

Verification

Rules No rules evaluated

Functional validation Not enabled

Set up

Auto-approval Approved

Tip

Comment Gitar fix CI or enable auto-apply: gitar auto-apply:on

Options

Auto-apply is off → Gitar will not commit updates to this branch.
Display: compact → Counting what did not apply, without listing it.

Comment with these commands to change the behavior for this request:

Auto-apply	Compact
gitar auto-apply:on	gitar display:verbose

_{Was this helpful? React with 👍 / 👎 | Gitar}

[BigSimmo]

Advisory UI (non-blocking) failed on 7d320bb in tests/ui-care-plan-mockup.spec.ts:1599, the Care Plan mockup's degraded-specimen test. This PR touches no Care Plan or mockup files, and the same job passed on the previous head before main was merged in. So it is not this PR's change. It most likely comes from the recent main work on the role switcher and audit hardening, or it is a flake. I'm not fixing it here, per the owner's instruction to report failures rather than widen this PR, and I haven't re-run it. It does not block merge.

---

Generated by Claude Code