Search results bar: delete the inert scope system, rebuild the filter shelf, and land the bar anatomy

[BigSimmo]

Summary

Three changes from the Results bar — perfected study (artifact 007a83f4), applied in the shared results band so all twelve search modes take them at once.

• Deletes the inert command-scope system. universal-search-command-surface.tsx received commandScopes and onCommandScopesChange and discarded both with void, and every other write in the repo passed an empty array — so the value was permanently [] in production, on main, before this branch existed. Everything built on it was therefore unreachable: the scopes config for six modes, the three matcher helpers, the four call sites that filtered against them (services, forms, prescribing, favourites), and the band's original scope shelf. Behaviour-preserving by construction rather than by argument — each matcher early-returned true on an empty array, so every call site was already a no-op. SearchCommandContext survives with query and modeId, which applications-launcher-page.tsx still consumes.
• Rebuilds the applied-filter shelf on live facet data. The band once carried a shelf, but it read filter state from a context nothing populated, so it rendered for nobody and passed a DOM test that constructed that context by hand. The new one is prop-driven: the page supplies appliedFilters (id, label, onRemove) and the band renders them. Filtered by labels the row as state rather than a second bank of buttons; every chip removes in one tap with the count following immediately; Clear appears past one chip. It survives loading and a zero result — nothing matching is exactly when you want to relax a filter, and dropping it mid-search would flicker the chips out and back on every keystroke. Only a fault removes it.
• Bar anatomy. The state tile already existed but carried only two of its four states; it now shows a spinner while a search runs and a funnel once the result set is narrowed, the funnel driven by appliedFilters.length so it tracks the same live data as the shelf. Filter moves hard against the ribbon's right edge and Sort moves inboard: Sort is set about once a session, Filter is the only control carrying state and the one a thumb reaches for.

The shelf is scoped to documents and therapy-compass — the two modes whose filters are multi-valued and hidden behind a panel. Differentials, prescribing, specifiers, formulation, services and factsheets each have a single-select dimension whose control is already visible in the bar, so a shelf there would restate what is on screen.

One study step deliberately not done

The study's step 6 removes the library button from the results bar on the grounds that corpus browsing belongs in nav. It explicitly declines to check where nav puts it. In this repo the documents action menu routes through onSearchModeChange, which calls setQuery("") and setModeSearchSubmitted(false), so reaching the library that way discards the search being read. The bar button is the only in-context route to it, and was renamed Library / Open source library instead. docs/handoff-search-bar.md records this so it is not "finished" by mistake.

Verification

• npm run verify:pr-local — exit 0. Test Files 460 passed (460), Tests 4796 passed | 4 skipped (4800), production build clean, client-bundle secret surface check passed, offline RAG fixture validation passed (36 golden cases, 23 suites).
• npm run verify:ui scope covered by direct Chromium spec runs: tests/ui-tools.spec.ts 87 passed; tests/ui-smoke.spec.ts + tests/ui-accessibility.spec.ts 108 passed, 1 failed.
• ui-tools was run first and separately because it was the real risk: it carried two assertions on the Sort/Filter pair that this work deliberately separates, and it runs expectNoPageHorizontalOverflow at 390 px immediately after them. Both clear. Those assertions are rewritten to pin what still has to hold — the page filter is the utility rail's last child, and both controls keep a matched phone tap height of at least 43 px — rather than the old adjacency.
• The single failure is document viewer puts the PDF preview first with pinned evidence after it on mobile, at pdfScroller.locator("canvas"). It fails identically with these changes stashed: this environment has Chromium 1194 against the project's pinned 1228, run through the PLAYWRIGHT_CHROMIUM_EXECUTABLE_PATH hook the config already supports.
• Mutation-tested. Widening the shelf's render condition from !faulted to !countUntrusted fails the pending-search case, so the "survives a search in flight" guard is real rather than vacuous.
• npm run eval:retrieval:quality, eval:rag, eval:quality, check:supabase-project — not run and not applicable. No retrieval, ranking, selection, chunking, scoring, ingestion, or provider-backed surface is touched.

RAG impact: no retrieval behaviour change — this changes only which client-side controls render, and how already-returned result sets are filtered and labelled. No file under src/lib/rag/**, clinical-search, retrieval-selection, ranking-config, answer-ranking, the eval harness, or the golden fixture is touched, and no retrieval RPC or comparator ordering is altered. src/lib/search-command-surface.ts is edited, but only to remove the dead scope config and matchers described above; its retrieval-adjacent exports are untouched.

Risk and rollout

• Risk: UI-only. The largest change by line count is a deletion of code that provably could not execute — every matcher early-returned true on a permanently-empty array. The behavioural change a reader will notice is the shelf appearing when filters are applied, and Filter sitting at the right edge instead of beside Sort. The services, forms, prescribing and favourites pages lose filtering calls that were no-ops, so their result sets are unchanged.
• Rollback: four independently revertible commits — the scope deletion, the shelf, the bar anatomy, and docs. A squash merge folds them, so a post-merge revert of one item means reverting its hunks by hand.
• Provider or production effects: None. No provider-backed gate was run and no Supabase, OpenAI, or Railway surface is touched.

Clinical Governance Preflight

classifyPullRequestFiles returns clinicalRisk: true for this file set, checked against the full origin/main...HEAD diff rather than the tip commit. The trigger is path-based — src/lib/search-command-surface.ts plus the services and documents result pages — and the edits to all three are removals of unreachable filtering code and changes to which client-side controls render. No ingestion, answer generation, retrieval, source rendering, or document-access path is altered.

• Source-backed claims still require linked source verification before clinical use
• No patient-identifiable document workflow was introduced or expanded without explicit governance approval
• Supabase target remains Clinical KB Database (sjrfecxgysukkwxsowpy)
• Service-role keys and private document access remain server-only
• Demo/synthetic content remains clearly separated from real clinical sources
• Source metadata, review status, and outdated/unknown-source behavior remain conservative
• Deployment classification/TGA SaMD impact was checked when clinical decision-support behavior changed

Notes

• The wip(search) commit message calls itself UNVERIFIED and points at docs/handoff-search-bar.md. That caveat is superseded by the gate results above; history was not rewritten to correct it in place because the branch was already pushed and a force-push is not worth a stale adjective. The doc records the correction.
• Two traps worth keeping out of the next change on this surface, both recorded in the handoff doc. Count what a control does, not how many there are — formulation's "Pattern" and factsheets' "Category" look like filters and are navigation, and services' "quick filter" rewrites the query. A shared component must not read filter state from context — that is precisely how the previous shelf shipped dead while its test passed.
• Still open from the study and not attempted here: moving Sort into the sheet on phone, with the sheet retitled Filter and sort. It is per-page work, because only documents and therapy-compass have sheets; doing it in the shared band would strip Sort from the other six modes on phones, which is a defect the study itself records fixing in an earlier round.

---

Generated by Claude Code

[coderabbitai]

Warning

Review limit reached

You’ve reached a temporary PR review limit under our Fair Usage Limits Policy.

Your recent review volume is higher than typical usage, so adaptive limits are currently applied.

Next review available in: 10 minutes

Your organization has reached its usage spending cap. Adjust your spending cap in the billing tab.

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: 0c7efb29-8d53-4cf6-9db1-c712bde771dc

📥 Commits

Reviewing files that changed from the base of the PR and between 6aa9427 and 3f08fa0.

📒 Files selected for processing (20)

• docs/branch-review-ledger.md
• docs/handoff-search-bar.md
• docs/outstanding-issues.md
• src/components/ClinicalDashboard.tsx
• src/components/clinical-dashboard/document-search-results.tsx
• src/components/clinical-dashboard/favourites-command-library-page.tsx
• src/components/clinical-dashboard/global-search-shell.tsx
• src/components/clinical-dashboard/master-search-header.tsx
• src/components/clinical-dashboard/medication-prescribing-workspace.tsx
• src/components/clinical-dashboard/search-command-context.tsx
• src/components/clinical-dashboard/search-results-header-band.tsx
• src/components/clinical-dashboard/universal-search-command-surface.tsx
• src/components/forms/forms-search-results-page.tsx
• src/components/services/services-navigator-page.tsx
• src/components/therapy-compass/screens/search-screen.tsx
• src/lib/search-command-surface.ts
• tests/document-filter-panel.dom.test.tsx
• tests/search-command-surface.test.ts
• tests/search-results-header-band.dom.test.tsx
• tests/ui-tools.spec.ts

---

_{Comment @coderabbitai help to get the list of available commands.}

[supabase]

This pull request has been ignored for the connected project sjrfecxgysukkwxsowpy because there are no changes detected in supabase directory. You can change this behaviour in Project Integrations Settings ↗︎.

---

Preview Branches by Supabase.
Learn more about Supabase Branching ↗︎.

[ecc-tools]

🔒 Upgrade Required

Private repository analysis requires Pro or Enterprise.

Upgrade: https://ecc.tools/pricing?plan=pro

---

_{ECC Tools keeps the core app open, and puts private repos, team features, and enterprise controls behind paid tiers.}

[ecc-tools]

🔒 Upgrade Required

Private repository analysis requires Pro or Enterprise.

Upgrade: https://ecc.tools/pricing?plan=pro

---

_{ECC Tools keeps the core app open, and puts private repos, team features, and enterprise controls behind paid tiers.}

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.

[ecc-tools]

🔒 Upgrade Required

Private repository analysis requires Pro or Enterprise.

Upgrade: https://ecc.tools/pricing?plan=pro

---

_{ECC Tools keeps the core app open, and puts private repos, team features, and enterprise controls behind paid tiers.}

[qodo-code-review]

PR Summary by Qodo

Remove inert command scopes; add prop-driven applied-filter shelf and results bar states

✨ Enhancement 🐞 Bug fix 🧪 Tests 📝 Documentation 🕐 40+ Minutes

AI Description

• Delete the unreachable command-scope system and its no-op consumers across search modes.
• Rebuild the applied-filter shelf as prop-driven chips sourced from live facet state.
• Update results bar anatomy: spinner/funnel tile states and mobile Sort/Filter layout.

Diagram

graph TD
  A["Documents results page"] --> B["Build appliedFilters[]"] --> D["SearchResultsHeaderBand"] --> E["Applied-filter shelf"]
  C["Therapy Compass search"] --> B
  D --> F["Tile states (spinner/funnel)"]
  G["SearchCommandContext (query, modeId)"] --> H["UniversalSearchCommandSurface"]

Loading

High-Level Assessment

The following are alternative approaches to this PR:

1. Wire command scopes end-to-end instead of deleting

• ➕ Preserves the conceptual 'scope chips' feature if it’s still desired
• ➕ Could unify certain filtering affordances across modes
• ➖ Requires UX decisions and real interaction design (chips vs dropdown vs panel)
• ➖ Higher risk: would change production filtering behavior that was previously inert
• ➖ Adds ongoing maintenance surface across many modes and tests

2. Keep shelf context-driven (populate shared context from pages)

• ➕ Centralizes applied filter state and avoids per-page prop plumbing
• ➕ Could enable cross-mode persistence if ever needed
• ➖ Easy to regress into the prior failure mode (context exists but no producer)
• ➖ Harder to keep mode-specific semantics correct (facet keys, labels, remove handlers)

3. Derive applied filters inside SearchResultsHeaderBand from standard interfaces

• ➕ Would reduce duplication if all modes implement a common filter model
• ➕ Could standardize labels and chip behaviors
• ➖ Not currently feasible: filter state is heterogeneous (facets vs result-type vs toggles)
• ➖ Pushes mode-specific knowledge into a shared component

Recommendation: Keep the PR’s approach: deleting the inert scope system is the safest and clearest outcome given it was behaviorally unreachable, and moving the shelf to prop-driven appliedFilters makes ownership explicit and prevents the prior ‘context with no producer’ regression. If scope chips are ever reintroduced, treat it as a separate feature with explicit UX + end-to-end tests rather than reviving the previous scaffolding.

Files changed (20) +358 / -479

Enhancement (3) +120 / -92

document-search-results.tsxDerive applied filter chips from facet + result-type state and pass to header band +36/-4

Derive applied filter chips from facet + result-type state and pass to header band

• Builds an appliedFilters array from active facet keys and the result-type tab (when not 'all'), each with an onRemove handler. Adds a clearAllFilters helper and wires appliedFilters/onClearFilters into SearchResultsHeaderBand and the filter panel's Clear action.

src/components/clinical-dashboard/document-search-results.tsx

search-results-header-band.tsxAdd prop-driven applied-filter shelf and update results bar states/layout +74/-88

Add prop-driven applied-filter shelf and update results bar states/layout

• Introduces AppliedFilterChip prop API and renders a labelled 'Filtered by' shelf with one-tap chip removal and optional Clear. Updates the tile icon logic to show spinner while busy and funnel when filtered, and changes mobile layout so Sort is inboard and page filter controls sit last at the right edge; removes scope-based empty-state affordances.

src/components/clinical-dashboard/search-results-header-band.tsx

search-screen.tsxProvide appliedFilters chips for Therapy Compass shelf +10/-0

Provide appliedFilters chips for Therapy Compass shelf

• Builds an appliedFilters array from active topics and availability toggles and passes it to SearchResultsHeaderBand, along with a clear-all handler.

src/components/therapy-compass/screens/search-screen.tsx

Refactor (10) +16 / -240

ClinicalDashboard.tsxRemove commandScopes state from ClinicalDashboard context and mode changes +1/-14

Remove commandScopes state from ClinicalDashboard context and mode changes

• Deletes commandScopes state and clear/remove callbacks from the dashboard and from SearchCommandContext value construction. Also removes mode-change resets that previously (and ineffectively) cleared scopes.

src/components/ClinicalDashboard.tsx

favourites-command-library-page.tsxRemove command-scope filtering paths from favourites mode +6/-19

Remove command-scope filtering paths from favourites mode

• Eliminates useSearchCommand and favouriteMatchesCommandScopes usage, removing scope-driven filtering and empty-state 'clear scopes' behavior. Match counts and empty states now reflect the actual filteredItems only.

src/components/clinical-dashboard/favourites-command-library-page.tsx

global-search-shell.tsxRemove commandScopes state from global search shell +1/-15

Remove commandScopes state from global search shell

• Deletes commandScopes state and callbacks and removes scope resets on mode changes. SearchCommandContext now provides only query and modeId, and scope props are no longer passed to the command surface.

src/components/clinical-dashboard/global-search-shell.tsx

master-search-header.tsxDrop commandScopes props from shared master search header wiring +0/-6

Drop commandScopes props from shared master search header wiring

• Removes commandScopes and onCommandScopesChange from the header’s props and stops passing them into UniversalSearchCommandSurface. This aligns the header API with the now-simplified command surface and context.

src/components/clinical-dashboard/master-search-header.tsx

medication-prescribing-workspace.tsxRemove command-scope filtering from prescribing results +2/-8

Remove command-scope filtering from prescribing results

• Eliminates useSearchCommand and medicationMatchesCommandScopes usage, making filter counts and displayed rows reflect the underlying search results only.

src/components/clinical-dashboard/medication-prescribing-workspace.tsx

search-command-context.tsxSimplify SearchCommandContextValue to query + modeId only +0/-3

Simplify SearchCommandContextValue to query + modeId only

• Removes commandScopes and scope mutation callbacks from the context type, reflecting the deletion of the entire scope system.

src/components/clinical-dashboard/search-command-context.tsx

universal-search-command-surface.tsxRemove unused commandScopes props from command surface +0/-6

Remove unused commandScopes props from command surface

• Deletes commandScopes and onCommandScopesChange props (and the prior 'void' discards), making the component’s interface match real behavior.

src/components/clinical-dashboard/universal-search-command-surface.tsx

forms-search-results-page.tsxRemove command-scope filtering from forms search results +2/-11

Remove command-scope filtering from forms search results

• Eliminates useSearchCommand and recordMatchesCommandScopes usage, and updates sorting to operate directly on matches. Removes empty-state 'clear scopes' hook.

src/components/forms/forms-search-results-page.tsx

services-navigator-page.tsxRemove command-scope filtering from services navigator results +4/-13

Remove command-scope filtering from services navigator results

• Deletes useSearchCommand and recordMatchesCommandScopes usage. Displayed matches, rank map, and empty state now use the unscoped matches set.

src/components/services/services-navigator-page.tsx

search-command-surface.tsDelete scope chip config and all scope matcher helpers +0/-145

Delete scope chip config and all scope matcher helpers

• Removes CommandScopeChip and per-mode scopes configuration, along with record/favourite/medication scope matching functions and helpers. Leaves suggestions/examples and cross-mode config intact.

src/lib/search-command-surface.ts

Tests (4) +110 / -146

document-filter-panel.dom.test.tsxAdd DOM tests for applied-filter shelf behavior in documents +66/-0

Add DOM tests for applied-filter shelf behavior in documents

• Introduces tests asserting shelf appearance, one-tap chip removal, Clear behavior with multiple chips (facet + source type), and shelf persistence through loading vs removal on fault.

tests/document-filter-panel.dom.test.tsx

search-command-surface.test.tsRemove tests for deleted scope matchers and other unrelated assertions +0/-90

Remove tests for deleted scope matchers and other unrelated assertions

• Deletes unit tests that exercised record/favourite/medication scope matchers and removes assertions that depended on reading universal-search-command-surface source. Retains command dropdown environment checks and suggestion filtering tests.

tests/search-command-surface.test.ts

search-results-header-band.dom.test.tsxUpdate header band tests for removed scopes and new mobile layout +10/-34

Update header band tests for removed scopes and new mobile layout

• Removes scope-related utility strip tests and updates wiring tests to only cover sort/view/save controls. Adjusts the mobile control test to validate Sort in utilities and page filter controls rendered last at the right edge.

tests/search-results-header-band.dom.test.tsx

ui-tools.spec.tsUpdate Playwright layout assertions for Sort/Filter separation on mobile +34/-22

Update Playwright layout assertions for Sort/Filter separation on mobile

• Replaces the prior 'mobile control pair' assertions with checks that the page filter group is the last child in the utilities rail and that Sort and Filter maintain matched tap height.

tests/ui-tools.spec.ts

Documentation (3) +112 / -1

branch-review-ledger.mdRecord PR handoff and verification status in ledger +1/-0

Record PR handoff and verification status in ledger

• Adds a new ledger row capturing the handoff summary, scope deletion rationale, shelf rebuild, bar anatomy changes, and verification results for this branch/PR.

docs/branch-review-ledger.md

handoff-search-bar.mdAdd detailed handoff doc for 'Results bar — perfected' changes +110/-0

Add detailed handoff doc for 'Results bar — perfected' changes

• Introduces a handoff document describing which study steps were implemented, which were intentionally declined, test gate status, and scoping decisions (notably shelf limited to documents and therapy-compass).

docs/handoff-search-bar.md

outstanding-issues.mdArchive issue #182 as deleted and summarize outcome +1/-1

Archive issue #182 as deleted and summarize outcome

• Moves the command-scope system issue into the resolved section with a deletion summary and verification notes, reflecting the completed removal across the codebase.

docs/outstanding-issues.md

[qodo-code-review]

Code Review by Qodo

🐞 Bugs (2) 📘 Rule violations (0) 📜 Skill insights (0)

1. Clear wipes query too 🐞 Bug ≡ Correctness

Description

Therapy Compass passes clearSearch to SearchResultsHeaderBand.onClearFilters, but clearSearch
resets EMPTY_SEARCH including query: "". This makes the shelf’s trailing Clear unexpectedly
delete the user’s search term even though the shelf explicitly models only non-query filters.

Code

src/components/therapy-compass/screens/search-screen.tsx[R53-55]

+        appliedFilters={appliedFilters}
+        onClearFilters={b.clearSearch}
        filterLabel="Filter therapy results"

Evidence

The shelf Clear callback is b.clearSearch, and clearSearch resets EMPTY_SEARCH, whose query
is ""; meanwhile the shelf’s applied filters intentionally exclude the query, and the band’s prop
contract describes clearing filters (not query).

src/components/therapy-compass/screens/search-screen.tsx[24-55]
src/components/therapy-compass/bindings.tsx[390-409]
src/components/therapy-compass/data/select.ts[57-71]
src/components/clinical-dashboard/search-results-header-band.tsx[173-179]

Agent prompt

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

### Issue description
Therapy Compass’s applied-filter shelf Clear button is wired to `b.clearSearch`, which resets the whole search state including `query`. The shelf intentionally excludes the query from `appliedFilters`, so Clear should clear only the applied filters represented by the shelf (tags + availability toggles), or the UI/contract should be updated to make “Clear” explicitly mean “clear the entire search including query”.

### Issue Context
- `SearchResultsHeaderBand` documents `onClearFilters` as clearing applied filters.
- Therapy Compass’s `clearSearch` resets `EMPTY_SEARCH` where `query` is empty.

### Fix Focus Areas
- src/components/therapy-compass/screens/search-screen.tsx[32-55]
- src/components/therapy-compass/bindings.tsx[390-409]
- src/components/therapy-compass/data/select.ts[57-71]
- src/components/clinical-dashboard/search-results-header-band.tsx[173-179]

### Suggested fix
Implement a filter-only clear handler (e.g., `clearFiltersOnly`) that preserves the current query while resetting `tags`, `reviewedOnly`, and `briefOnly`, and pass that to `onClearFilters` instead of `b.clearSearch`.

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

---

2. Memo deps lint suppressed 🐞 Bug ⚙ Maintainability

Description

DocumentSearchResultsPanelImpl suppresses react-hooks/exhaustive-deps for the appliedFilters
memo while capturing toggleTagFacet, which is declared inline (identity changes every render).
This under-specified dependency list makes the chip callbacks fragile for future refactors and
contradicts the in-code comment claiming a “stable closure”.

Code

src/components/clinical-dashboard/document-search-results.tsx[R1162-1187]

+  const appliedFilters = useMemo(() => {
+    const selected = new Set(activeFacetKeys);
+    const chips = tagFacetGroups.flatMap((group) =>
+      group.facets
+        .filter((facet) => selected.has(facet.key))
+        .map((facet) => ({
+          id: facet.key,
+          label: facet.label,
+          onRemove: () => toggleTagFacet(facet),
+        })),
+    );
+    if (effectiveResultType !== "all") {
+      const tab = resultTabs.find((entry) => entry.key === effectiveResultType);
+      if (tab) {
+        chips.push({
+          id: `result-type-${tab.key}`,
+          label: tab.label,
+          onRemove: () => setActiveResultType("all"),
+        });
+      }
+    }
+    return chips;
+    // `toggleTagFacet` is a stable closure over `query`, which is already a
+    // dependency of `activeFacetKeys`.
+    // eslint-disable-next-line react-hooks/exhaustive-deps
+  }, [tagFacetGroups, activeFacetKeys, effectiveResultType, resultTabs]);

Evidence

The chip onRemove closures call toggleTagFacet, but the memo dependency list excludes it and the
lint rule is explicitly disabled; toggleTagFacet is defined as a normal function inside the
component (not useCallback).

src/components/clinical-dashboard/document-search-results.tsx[1113-1121]
src/components/clinical-dashboard/document-search-results.tsx[1159-1187]

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 `appliedFilters` `useMemo` intentionally omits `toggleTagFacet` (and suppresses the lint rule), but `toggleTagFacet` is an inline function recreated each render. This increases the risk of subtle bugs if `toggleTagFacet` later starts capturing additional mutable values or its implementation changes.

### Issue Context
Current behavior likely works because `activeFacetKeys/tagFacetGroups` change on the meaningful transitions, but the lint suppression makes this easy to regress.

### Fix Focus Areas
- src/components/clinical-dashboard/document-search-results.tsx[1113-1121]
- src/components/clinical-dashboard/document-search-results.tsx[1159-1187]

### Suggested fix
Either:
1) Wrap `toggleTagFacet` in `useCallback` with correct deps and include it in the `useMemo` dependency array, and remove the eslint disable; or
2) Avoid memoizing `appliedFilters` and just compute it inline (if perf impact is negligible), removing the lint suppression.

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

---

_{To customize comments, go to the Qodo configuration screen, or learn more in the docs.}

[BigSimmo]

@cursoragent fox comments and CI failing

[cursor]

❌ Unable to add follow-up after multiple attempts. The agent may be experiencing issues.

Please try again later or Open in Web to view the status.

  

_{Learn more about Cursor Agents}