Detail ready

\n ,\n );\n \n await waitFor(() => expect(fetchMock).toHaveBeenCalledTimes(1));\n expect(String(fetchMock.mock.calls[0]?.[0])).toMatch(/\\/therapies\\.json$/);\n });\n });\n","external_storage_diff",{"_340":341,"_342":-5},"files_modified",2,"lines_added",73,"lines_removed",17,"commit_message","fix(therapy): scope retry settlement to request","additional_parent_commit_shas",[],"file_id","file_00000000b5e0820c85e5d2c4f2a106ec","ttl","partial_repo_snapshot","files",[346,347],{"_348":485,"_350":-5,"_351":486,"_353":-5},{"_348":349,"_350":-5,"_351":352,"_353":-5},"path","tests/therapy-compass-data-recovery.dom.test.tsx","contents","line_range_contents",[354],"missing_reason",{"_355":356,"_357":358,"_359":360,"_361":62},"line_range_start",1,"line_range_end",209,"content",[362,363,181,364,365,366,181,367,181,368,369,370,371,372,181,373,374,375,376,377,378,379,380,181,381,382,383,384,385,386,387,181,388,389,390,391,392,181,393,394,395,396,397,398,399,400,401,402,403,404,405,398,406,181,407,408,409,410,411,181,412,413,414,415,181,416,181,417,418,413,414,419,181,420,421,399,400,401,422,404,405,398,406,181,407,408,409,410,411,181,423,414,424,181,425,415,181,426,427,181,428,429,430,431,419,181,432,421,433,434,435,398,399,400,401,436,437,403,404,405,398,406,181,407,408,409,410,411,181,425,426,438,439,181,440,441,415,181,442,181,428,430,419,181,443,444,445,446,447,398,448,449,450,451,452,453,454,455,456,457,458,459,460,411,181,461,462,463,464,411,465,181,466,467,468,469,470,471,398,472,181,473,474,475,181,476,477,478,419,181,479,480,399,400,481,405,398,406,181,407,408,482,410,411,181,483,484,419,392],"contains_end_of_file","import { act, fireEvent, render, renderHook, screen, waitFor } from \"@testing-library/react\";","import { afterEach, describe, expect, it, vi } from \"vitest\";","import { TherapyCompassWorkspace } from \"@/components/therapy-compass\";","import { clearTherapyDataCache, useTherapyData } from \"@/components/therapy-compass/data/use-therapy-data\";","import { HomeScreen } from \"@/components/therapy-compass/screens/home-screen\";","const navigation = vi.hoisted(() => ({ pathname: \"/therapy-compass\" }));","vi.mock(\"next/navigation\", () => ({"," usePathname: () => navigation.pathname,"," useSearchParams: () => new URLSearchParams(),"," useRouter: () => ({ push: vi.fn(), replace: vi.fn(), prefetch: vi.fn() }),","}));","const therapy = {"," slug: \"test-therapy\","," name: \"Test therapy\","," category: \"Skills based\","," tags: [],"," aliases: [],"," reviewStatus: \"reviewed\",","};","function response(body: unknown, ok = true, status = 200) {"," return {"," ok,"," status,"," json: vi.fn().mockResolvedValue(body),"," };","}","afterEach(() => {"," navigation.pathname = \"/therapy-compass\";"," clearTherapyDataCache();"," vi.unstubAllGlobals();","});","describe(\"Therapy Compass required data recovery\", () => {"," it(\"does not advertise a zero therapy count while the catalogue is still loading\", async () => {"," let release!: (value: unknown) => void;"," const therapiesGate = new Promise((resolve) => {"," release = resolve;"," });"," const fetchMock = vi.fn(async (input: RequestInfo | URL) => {"," const path = String(input);"," if (path.endsWith(\"/therapies-index.json\")) {"," await therapiesGate;"," return response([therapy]);"," }"," throw new Error(`Unexpected fetch: ${path}`);"," vi.stubGlobal(\"fetch\", fetchMock);"," render("," "," "," ,"," );"," expect(await screen.findByRole(\"status\")).toHaveTextContent(\"Loading therapy library…\");"," expect(screen.getAllByRole(\"main\")).toHaveLength(1);"," expect(screen.queryByText(/Source-grounded therapy records\\./)).not.toBeInTheDocument();"," expect(screen.queryByRole(\"heading\", { name: \"Therapy\" })).not.toBeInTheDocument();"," release(undefined);"," await waitFor(() => expect(screen.getByText(/1 source-grounded therapy record\\./)).toBeInTheDocument());"," expect(screen.getByRole(\"heading\", { name: \"Therapy\" })).toBeInTheDocument();"," });"," it(\"shows an honest load error, retries all required files, and recovers\", async () => {"," let failTherapies = true;"," return failTherapies ? response(null, false, 503) : response([therapy]);"," expect(screen.getByRole(\"status\")).toHaveTextContent(\"Loading therapy library\");"," expect(screen.queryByRole(\"heading\", { name: \"Frequently used therapies\" })).not.toBeInTheDocument();"," expect(await screen.findByRole(\"alert\")).toHaveTextContent(\"Therapy could not load\");"," failTherapies = false;"," fireEvent.click(screen.getByRole(\"button\", { name: \"Retry\" }));"," expect(await screen.findByRole(\"heading\", { name: \"Therapy\" })).toBeInTheDocument();"," expect(screen.getByText(/1 source-grounded therapy record\\./)).toBeInTheDocument();"," expect(screen.queryByRole(\"alert\")).not.toBeInTheDocument();"," await waitFor(() => expect(fetchMock).toHaveBeenCalledTimes(2));"," it(\"keeps Retry busy until the replacement catalogue request settles\", async () => {"," let releaseRetry!: (value: unknown) => void;"," const retryGate = new Promise((resolve) => {"," releaseRetry = resolve;"," if (failTherapies) return response(null, false, 503);"," await retryGate;"," const retry = screen.getByRole(\"button\", { name: \"Retry\" });"," fireEvent.click(retry);"," await waitFor(() => expect(screen.getByRole(\"button\", { name: \"Retrying…\" })).toBeDisabled());"," expect(screen.getByRole(\"alert\")).toBeInTheDocument();"," releaseRetry(undefined);"," it(\"settles a retry only when its request key and attempt finish\", async () => {"," let indexRequests = 0;"," let releaseIndexRetry!: () => void;"," const indexRetryGate = new Promise((resolve) => {"," releaseIndexRetry = resolve;"," vi.stubGlobal("," \"fetch\","," vi.fn(async (input: RequestInfo | URL) => {"," const path = String(input);"," if (path.endsWith(\"/therapies-index.json\")) {"," indexRequests += 1;"," if (indexRequests === 1) return response(null, false, 503);"," await indexRetryGate;"," return response([therapy]);"," }"," if (path.endsWith(\"/therapies.json\")) return response([therapy]);"," throw new Error(`Unexpected fetch: ${path}`);"," }),"," const { result, rerender } = renderHook("," ({ catalogue }: { catalogue: \"index\" | \"full\" }) =>"," useTherapyData({ catalogue, includePathways: false, includeReference: false }),"," { initialProps: { catalogue: \"index\" as const } },"," await waitFor(() => expect(result.current.error).toMatch(/503/));"," let retrySettled = false;"," let retryPromise!: Promise;"," act(() => {"," retryPromise = result.current.retry().then(() => {"," retrySettled = true;"," });"," await waitFor(() => expect(indexRequests).toBe(2));"," rerender({ catalogue: \"full\" });"," await waitFor(() => expect(result.current.data?.therapies).toEqual([therapy]));"," expect(retrySettled).toBe(false);"," releaseIndexRetry();"," await act(async () => retryPromise);"," expect(retrySettled).toBe(true);"," it(\"loads full therapy records only on a record-rich route\", async () => {"," navigation.pathname = \"/therapy-compass/test-therapy\";"," if (path.endsWith(\"/therapies.json\")) return response([therapy]);","

Detail ready

"," await waitFor(() => expect(fetchMock).toHaveBeenCalledTimes(1));"," expect(String(fetchMock.mock.calls[0]?.[0])).toMatch(/\\/therapies\\.json$/);","src/components/therapy-compass/data/use-therapy-data.ts",[487],{"_355":356,"_357":488,"_359":489,"_361":62},141,[490,181,491,181,492,181,493,494,495,496,181,497,498,499,500,380,181,501,502,503,181,504,505,506,387,181,507,508,509,510,387,181,511,512,513,514,515,516,517,518,519,387,181,520,521,522,523,524,525,380,181,526,527,528,529,530,531,532,533,534,535,536,537,538,539,540,541,542,543,544,181,545,546,547,548,549,411,550,551,552,553,554,555,556,539,557,558,181,559,560,561,562,563,564,565,566,567,568,569,398,570,571,572,181,573,574,575,576,577,578,579,580,581,582,583,584,585,586,579,580,587,588,589,590,591,592,471,593,594,595,596,181,573,593,597,595,598,181,599,600,387],"\"use client\";","import { useCallback, useEffect, useMemo, useRef, useState } from \"react\";","import type { Pathway, ReferenceData, Therapy, TherapyDataset } from \"./types\";","// Served as static public assets. Kept outside /mockups so the production","// `/therapy-compass` route can load them — `proxy.ts` 404s every /mockups path","// in production, which would otherwise starve the tool of its dataset.","const BASE = \"/therapy-compass-data\";","type TherapyDataOptions = {"," catalogue?: \"index\" | \"full\";"," includePathways?: boolean;"," includeReference?: boolean;","// Cache each route-sized payload once per session. Catalogue routes use the compact","// index; detail/compare/recommend/artifact routes opt into the full records.","const cache = new Map>();","/** Test helper: drop the session-scoped dataset cache between cases. */","export function clearTherapyDataCache() {"," cache.clear();","async function fetchJson(path: string): Promise {"," const res = await fetch(path);"," if (!res.ok) throw new Error(`Failed to load ${path}: ${res.status}`);"," return res.json() as Promise;","async function loadDataset(options: Required): Promise {"," const [therapies, pathways, reference] = await Promise.all(["," fetchJson(`${BASE}/${options.catalogue === \"full\" ? \"therapies.json\" : \"therapies-index.json\"}`),"," options.includePathways ? fetchJson(`${BASE}/pathways.json`) : Promise.resolve([]),"," options.includeReference"," ? fetchJson(`${BASE}/reference.json`)"," : Promise.resolve({ categories: [], tags: [], measures: [] }),"," ]);"," return { therapies, pathways, reference };","export type TherapyDataState = {"," data: TherapyDataset | null;"," loading: boolean;"," error: string | null;"," /** Settles when the replacement request finishes so Retry can stay busy. */"," retry: () => Promise;","export function useTherapyData(options: TherapyDataOptions = {}): TherapyDataState {"," const resolved = useMemo>("," () => ({"," catalogue: options.catalogue ?? \"full\","," includePathways: options.includePathways ?? true,"," includeReference: options.includeReference ?? true,"," }),"," [options.catalogue, options.includePathways, options.includeReference],"," );"," const requestKey = `${resolved.catalogue}:${resolved.includePathways ? \"pathways\" : \"none\"}:${resolved.includeReference ? \"reference\" : \"none\"}`;"," const [state, setState] = useState<"," Omit & {"," requestKey: string | null;"," }"," >({ requestKey: null, data: null, loading: true, error: null });"," const [attempt, setAttempt] = useState(0);"," const retryTokenRef = useRef(0);"," const retryWaitersRef = useRef void }>>([]);"," const inFlightRetryRef = useRef<{ requestKey: string; attempt: number; promise: Promise } | null>(null);"," const settleRetryWaiters = useCallback((settledRequestKey?: string, settledAttempt?: number) => {"," const matching = retryWaitersRef.current.filter("," (waiter) =>"," settledRequestKey === undefined ||"," (waiter.requestKey === settledRequestKey && waiter.attempt === settledAttempt),"," retryWaitersRef.current = retryWaitersRef.current.filter((waiter) => !matching.includes(waiter));"," if ("," settledRequestKey === undefined ||"," (inFlightRetryRef.current?.requestKey === settledRequestKey &&"," inFlightRetryRef.current.attempt === settledAttempt)"," ) {"," inFlightRetryRef.current = null;"," for (const waiter of matching) waiter.settle();"," }, []);"," const retry = useCallback((): Promise => {"," // Coalesce repeated clicks onto the same in-flight reload."," if (inFlightRetryRef.current?.requestKey === requestKey) return inFlightRetryRef.current.promise;"," cache.delete(requestKey);"," // Keep the prior error message mounted so Retry focus is not lost while the"," // next attempt is in flight; success/failure handlers replace it below."," setState((prev) => ({ ...prev, loading: true }));"," const retryAttempt = ++retryTokenRef.current;"," const promise = new Promise((resolve) => {"," retryWaitersRef.current.push({ requestKey, attempt: retryAttempt, settle: resolve });"," setAttempt(retryAttempt);"," inFlightRetryRef.current = { requestKey, attempt: retryAttempt, promise };"," return promise;"," }, [requestKey]);"," useEffect(() => {"," let active = true;"," if (!cache.has(requestKey)) cache.set(requestKey, loadDataset(resolved));"," const request = cache.get(requestKey)!;"," request"," .then((data) => {"," settleRetryWaiters(requestKey, attempt);"," if (!active) return;"," setState({ requestKey, data, loading: false, error: null });"," })"," .catch((err: unknown) => {"," // Only clear the shared cache when this request is still the active one."," // A newer retry may have already replaced `cache` with a fresh promise."," if (cache.get(requestKey) === request) cache.delete(requestKey);"," setState({"," requestKey,"," data: null,"," loading: false,"," error: err instanceof Error ? err.message : \"Failed to load\","," });"," return () => {"," active = false;"," };"," }, [attempt, requestKey, resolved, settleRetryWaiters]);"," settleRetryWaiters();"," }, [settleRetryWaiters]);"," if (state.requestKey !== requestKey) return { data: null, loading: true, error: null, retry };"," return { data: state.data, loading: state.loading, error: state.error, retry };","message",[603,604,605,606,607,608,609],{"_610":49,"_49":621},{"_610":612,"_348":485,"_355":619,"_357":620},{"_610":49,"_49":618},{"_610":612,"_348":485,"_355":616,"_357":617},{"_610":49,"_49":615},{"_610":612,"_348":349,"_355":613,"_357":614},{"_610":49,"_49":611},"content_type","\n* Committed the fix as `e544d0d` (`fix(therapy): scope retry settlement to request`) and created the required follow-up PR metadata titled **“Follow-up: scope Therapy Compass retry settlement.”**\n\n**Testing**\n\n* ✅ `git diff --check`\n* ✅ `npx prettier --check src/components/therapy-compass/data/use-therapy-data.ts tests/therapy-compass-data-recovery.dom.test.tsx` — all files passed Prettier validation.\n* ✅ `git status --short --branch` — worktree clean after commit.\n* ⚠️ `npm run test -- tests/therapy-compass-data-recovery.dom.test.tsx` — not run because repository dependencies are absent and the environment provides Node `v20.20.2`, while the project requires Node 24.","repo_file_citation",145,189,"\n* Added a route-change regression test proving that completion of a newer full-catalogue request does not settle the older index-catalogue retry; the retry settles only after its original replacement request completes. ",105,137,"\n* Updated request settlement so superseded effects may settle only their matching retry while remaining unable to update the active route’s state. Unmount cleanup still safely releases remaining waiters. ",67,103,"### Summary\n\n* Scoped Therapy Compass retry waiters and in-flight retry coalescing to both the originating request key and retry-attempt token. A newer route request can no longer resolve a retry promise belonging to an older request. ",{"_96":297,"_45":88,"_100":297,"_624":625,"_626":627,"_628":629,"_135":-5,"_122":630},[87],"input_items",[649,650,651,652,653,654,655,656,657,658,659,660,661,662,663,664,665,666,667,668,669,670,671,672,673,674,675,676,677,678,679,680,681,682,683,684,685,686,687,688,689,690,691,692,693,694,695,696,697,698,699,700,701,702,703,704,705,706,707,708,709,710,711,712,713,714,715,716,717,718,719,720,721,722,723,724,725,726,727,728,729,730,731,732,733,734,735,736,737,738,739,740,741,742,743,744,745,746,747,748,749,750,751,752,753,754,755,756,757,758,759,760],"external_post_config",{"_631":632,"_234":167,"_633":277,"_245":-5,"_324":228,"_634":-5,"_635":-5,"_636":637,"_638":639,"_640":-5,"_641":-5,"_642":643,"_644":18,"_645":62,"_646":57,"_647":18,"_648":-5},"user_model_slug","slug",1785241265.936681,"installation_id",131602480,"pr_number","gitlab_service_account_project_id","review_trigger","comment_id",5104066531,"comment_type","issue","pr_write_guard_token","pending_pr_reaction_id","pending_comment_reaction_id",388961507,"comment_if_no_findings","comment_on_error","share_url","is_parallel_review","review_summary_id",{"_96":601,"_100":297,"_359":901},{"_284":285,"_286":287,"_898":291,"_899":132,"_293":-5,"_96":900},{"_96":761,"_286":897,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":896,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":895,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":894,"_348":887,"_764":765,"_766":-5},{"_96":761,"_286":893,"_348":842,"_764":765,"_766":-5},{"_96":761,"_286":892,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":891,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":890,"_348":848,"_764":765,"_766":849},{"_96":761,"_286":889,"_348":763,"_764":765,"_766":-5},{"_96":761,"_286":888,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":886,"_348":887,"_764":765,"_766":-5},{"_96":761,"_286":885,"_348":848,"_764":765,"_766":849},{"_96":761,"_286":884,"_348":763,"_764":765,"_766":-5},{"_96":761,"_286":883,"_348":842,"_764":765,"_766":-5},{"_96":761,"_286":882,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":881,"_348":842,"_764":765,"_766":879},{"_96":761,"_286":880,"_348":840,"_764":765,"_766":841},{"_96":761,"_286":878,"_348":842,"_764":765,"_766":879},{"_96":761,"_286":877,"_348":840,"_764":765,"_766":841},{"_96":761,"_286":876,"_348":843,"_764":765,"_766":-5},{"_96":761,"_286":875,"_348":842,"_764":765,"_766":-5},{"_96":761,"_286":873,"_348":842,"_764":765,"_766":874},{"_96":761,"_286":871,"_348":763,"_764":765,"_766":872},{"_96":761,"_286":870,"_348":840,"_764":765,"_766":841},{"_96":761,"_286":869,"_348":794,"_764":765,"_766":-5},{"_96":761,"_286":868,"_348":822,"_764":765,"_766":844},{"_96":761,"_286":867,"_348":797,"_764":765,"_766":798},{"_96":761,"_286":866,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":863,"_348":864,"_764":765,"_766":865},{"_96":761,"_286":862,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":861,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":860,"_348":848,"_764":765,"_766":849},{"_96":761,"_286":859,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":858,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":807,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":857,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":856,"_348":822,"_764":765,"_766":838},{"_96":761,"_286":855,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":854,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":853,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":852,"_348":842,"_764":765,"_766":-5},{"_96":761,"_286":851,"_348":843,"_764":765,"_766":-5},{"_96":761,"_286":850,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":847,"_348":848,"_764":765,"_766":849},{"_96":761,"_286":846,"_348":822,"_764":765,"_766":844},{"_96":761,"_286":845,"_348":840,"_764":765,"_766":841},{"_96":761,"_286":791,"_348":822,"_764":765,"_766":844},{"_96":761,"_286":791,"_348":843,"_764":765,"_766":-5},{"_96":761,"_286":791,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":791,"_348":842,"_764":765,"_766":-5},{"_96":761,"_286":839,"_348":840,"_764":765,"_766":841},{"_96":761,"_286":837,"_348":822,"_764":765,"_766":838},{"_96":761,"_286":836,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":835,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":834,"_348":822,"_764":765,"_766":-5},{"_96":761,"_286":833,"_348":820,"_764":765,"_766":-5},{"_96":761,"_286":832,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":831,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":830,"_348":763,"_764":765,"_766":824},{"_96":761,"_286":829,"_348":771,"_764":765,"_766":826},{"_96":761,"_286":828,"_348":771,"_764":765,"_766":826},{"_96":761,"_286":827,"_348":763,"_764":765,"_766":824},{"_96":761,"_286":825,"_348":771,"_764":765,"_766":826},{"_96":761,"_286":823,"_348":763,"_764":765,"_766":824},{"_96":761,"_286":821,"_348":822,"_764":765,"_766":-5},{"_96":761,"_286":819,"_348":820,"_764":765,"_766":-5},{"_96":761,"_286":818,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":817,"_348":792,"_764":765,"_766":-5},{"_96":761,"_286":816,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":815,"_348":794,"_764":765,"_766":795},{"_96":761,"_286":814,"_348":797,"_764":765,"_766":798},{"_96":761,"_286":813,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":812,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":811,"_348":794,"_764":765,"_766":795},{"_96":761,"_286":810,"_348":797,"_764":765,"_766":798},{"_96":761,"_286":809,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":808,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":807,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":806,"_348":794,"_764":765,"_766":795},{"_96":761,"_286":805,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":804,"_348":797,"_764":765,"_766":798},{"_96":761,"_286":803,"_348":788,"_764":765,"_766":-5},{"_96":761,"_286":802,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":801,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":800,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":799,"_348":792,"_764":765,"_766":-5},{"_96":761,"_286":796,"_348":797,"_764":765,"_766":798},{"_96":761,"_286":793,"_348":794,"_764":765,"_766":795},{"_96":761,"_286":791,"_348":792,"_764":765,"_766":-5},{"_96":761,"_286":790,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":789,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":787,"_348":788,"_764":765,"_766":-5},{"_96":761,"_286":786,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":784,"_348":485,"_764":765,"_766":785},{"_96":761,"_286":783,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":782,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":781,"_348":771,"_764":765,"_766":772},{"_96":761,"_286":780,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":779,"_348":763,"_764":765,"_766":767},{"_96":761,"_286":778,"_348":763,"_764":765,"_766":767},{"_96":761,"_286":777,"_348":773,"_764":765,"_766":774},{"_96":761,"_286":776,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":775,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":770,"_348":773,"_764":765,"_766":774},{"_96":761,"_286":770,"_348":763,"_764":765,"_766":767},{"_96":761,"_286":770,"_348":763,"_764":765,"_766":767},{"_96":761,"_286":770,"_348":771,"_764":765,"_766":772},{"_96":761,"_286":769,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":768,"_348":763,"_764":765,"_766":767},{"_96":761,"_286":762,"_348":763,"_764":765,"_766":767},"pull_request_comment","### Summary\n\n* The requested fix is already present in commit `861a351`: a `401` from `GET /api/account/favourites` invokes `markSessionExpired()`, while transient failures continue through the normal recoverable error path. [src/components/account-data-provider.tsxL109-L138](https://github.com/BigSimmo/Database/blob/5911c2ed46714fb1e7099f77603a724870db162a/src/components/account-data-provider.tsx#L109-L138)\n* Regression coverage confirms that `401` responses expire the session, while a `503` does not incorrectly sign the user out. [tests/favourites-account-retry.dom.test.tsxL113-L150](https://github.com/BigSimmo/Database/blob/5911c2ed46714fb1e7099f77603a724870db162a/tests/favourites-account-retry.dom.test.tsx#L113-L150)\n* The current branch already contains that commit and the worktree is clean, so no duplicate code change, commit, or follow-up PR was created.\n\n**Testing**\n\n* ⚠️ `npx vitest run tests/favourites-account-retry.dom.test.tsx` — could not execute because repository dependencies are absent; `npx` fetched a temporary Vitest installation that could not resolve `jsdom`.\n* ✅ `git diff --check`\n* ✅ `git status --short --branch` — clean worktree on branch `work`.\n\n [View task →](https://chatgpt.com/s/cd_6a689e587d048191b1cad5ba2de5edac)","src/components/clinical-dashboard/document-search-results.tsx","side","RIGHT","line",906,"### Summary\n\n* The requested fix is already present on the current branch in commit `861a351` (`fix(account): treat a 401 favourites load as session expiry`). The favourites loader now invokes `markSessionExpired()` for HTTP 401 responses, preventing Retry from repeatedly sending an expired authorization token. Transient non-authentication errors retain the normal retry path. [src/components/account-data-provider.tsxL109-L138](https://github.com/BigSimmo/Database/blob/5911c2ed46714fb1e7099f77603a724870db162a/src/components/account-data-provider.tsx#L109-L138)\n* Regression coverage verifies that a 401 expires the session, while a 503 does not incorrectly sign the user out. [tests/favourites-account-retry.dom.test.tsxL113-L150](https://github.com/BigSimmo/Database/blob/5911c2ed46714fb1e7099f77603a724870db162a/tests/favourites-account-retry.dom.test.tsx#L113-L150)\n* No additional changes were necessary, and the worktree remains clean. Therefore, no redundant commit or pull request was created.\n\n**Testing**\n\n* ✅ `git show --stat --oneline 861a351`\n* ✅ `git status --short --branch`\n* ⚠️ `npm run test:focused -- tests/favourites-account-retry.dom.test.tsx` (the repository coordinator declined focused execution because there are no changed files and requires the full suite)\n* ⚠️ `npm run test -- tests/favourites-account-retry.dom.test.tsx` (dependencies are unavailable: `node_modules/vitest/vitest.mjs` is missing, and the environment has Node `20.20.2` rather than the repository’s required Node 24 runtime)\n\n [View task →](https://chatgpt.com/s/cd_6a689e6043348191b810cbaca8a7ba99)","\n> [!NOTE]\n> Autofix is a beta feature. Expect some limitations and changes as we gather feedback and continue to improve it.\n\nAutofix skipped. No unresolved CodeRabbit review comments with fix instructions found.\n","@Codex fix","tests/search-results-band-adoption.test.ts",101,"src/components/account-data-provider.tsx",130,"**Re: `account-data-provider.tsx` — \"Treat favourites 401 responses as session expiry\"** — fixed in `861a3518`.\n\nConfirmed the asymmetry before changing anything: `setFavourite` (`:177`) and `clearFavourites` (`:200`) both call `markSessionExpired()` on a 401; the load path at `:122` threw a generic message with no status inspection at all.\n\nThe finding is right that this branch created the harm. Before `reload` existed there was no Retry on that path, so the missing check was inert — wiring Retry to it turned it into a button that re-sends the same rejected token forever. Same rule as the earlier finding here: a control that advertises an action must perform one.\n\nImplementation note: I depend on `markSessionExpired` by identity rather than reaching through `auth` inside the effect. `react-hooks/exhaustive-deps` demands the whole `auth` object once you call a method on it, and depending on that would re-run the favourites load on every auth-object render. It's a `useCallback` in `client.tsx:366`, so the destructured reference is stable.\n\nTwo tests, both checked against the unfixed code rather than assumed:\n\n- **401 → `markSessionExpired()` called.** Removing the guard fails this case specifically.\n- **503 → not called.** A transient server fault is exactly what Retry is for; flipping the session there would bounce a signed-in reader to a sign-in screen over a blip. That's the over-fire I wanted pinned, not just the under-fire reported.\n\nI also had to clear the hoisted `vi.fn()` call history in `beforeEach` — it survives `restoreAllMocks`, so the \"not called\" assertion was inheriting the previous test's call and passing for the wrong reason.\n\n**Verification on `861a3518`:** `prettier --check .` clean; `npm run typecheck` and `npm run lint` exit 0 with no `exhaustive-deps` warning; `npm run test` — **417 files, 4237 passed, 4 skipped**.\n\nFor context on the run before this one: `54b302f3` was fully green in CI, including **`Production UI` at 325 passed (12.1m)** — the first complete browser run this PR has had, and the `ui-overlap` duplicate-header failure did not recur, consistent with the intermittent clone tracked as `#093`.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `7abc91f19d`\n \n\n

ℹ️ About Codex in GitHub

\n
\n\n[Your team has set up Codex to review pull requests in this repo](https://chatgpt.com/codex/cloud/settings/general). Reviews are triggered when you\n- Open a pull request for review\n- Mark a draft as ready\n- Comment \"@codex review\".\n\nIf Codex has suggestions, it will comment; otherwise it will react with 👍.\n\n\n\n\nCodex can also answer questions or update the PR. Try commenting \"@codex address that feedback\".\n \n

","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Treat favourites 401 responses as session expiry**\n\nWhen `GET /api/account/favourites` returns 401 for an expired token, this stores only a generic `loadError` and leaves the auth state authenticated. The new favourites fault surfaces therefore offer Retry, but every retry reuses the same rejected authorization header and cannot recover or direct the user to sign in; the PUT/DELETE paths already call `markSessionExpired()` for this case. Preserve the response status, mark the session expired on 401, and add a focused test that a 401 transitions the favourites UI to its sign-in state rather than a retry loop.\n\nAGENTS.md reference: [AGENTS.md:L180-L182](https://github.com/BigSimmo/Database/blob/7abc91f19d7f799e586f130be8769caa9ef0d0a9/AGENTS.md#L180-L182)\n\nUseful? React with 👍 / 👎.","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Remove the duplicate registry failure notice**\n\nWhen a dashboard services/forms search reaches `error` or `unauthorized`, this status now mounts the band's fault panel, but the same `showRecordMatches` branch still renders `RecordRegistryNotice` at line 974. Users therefore see two adjacent failure panels for one registry request, unlike the standalone services/forms pages where the legacy notice was removed; suppress that notice while the band owns the failure, with a DOM assertion that the record-search path renders one fault message.\n\nUseful? React with 👍 / 👎.","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Preserve document matches through registry failures**\n\nWhen `showRecordMatches` is true and the services/forms registry fails while the document request succeeds, this branch faults the entire band and hides its count even though document results still render below. Fresh evidence at this head is that `matchCount` remains `recordMatchCount + sortedMatches.length`, so the ribbon is counting both sources while its status reflects only the registry; represent the partial failure separately or derive trust from both requests, and test a failed registry alongside successful document matches.\n\nUseful? React with 👍 / 👎.","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `54b302f38d`\n \n\n

ℹ️ About Codex in GitHub

","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Include query-backed dashboard modes in route inventory**\n\nFresh evidence after the prior top-level-route fix is that this early return drops `/?mode=prescribing` because its path is `/`, while Documents has no `href` and is omitted for the same root dashboard route. If `(search-app)/page.tsx` stops reaching `ClinicalDashboard` or its mode workspace, the global component scan still passes while the real result route has no band. Resolve query-backed root modes to that page and add a negative fixture that disconnects the root route.\n\nUseful? React with 👍 / 👎.","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `9662d9b2cc`\n \n\n

ℹ️ About Codex in GitHub

","**Actionable comments posted: 1**\n\n

🤖 Prompt for all review comments with AI agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nInline comments:\nIn `@src/components/therapy-compass/data/use-therapy-data.ts`:\n- Around line 68-90: Scope retry waiters to the request key and retry attempt in\nretry and settleRetryWaiters, so a pending retry promise is resolved only by the\nmatching request effect. Prevent a route change from allowing the new request to\nsettle the old retry, while ensuring the original retry is settled when its\nmatching request completes; add a test covering navigation during an in-flight\nretry.\n```\n\n

\n\n

🪄 Autofix (Beta)

\n\n❌ Autofix failed (check again to retry)\n\nFix all unresolved CodeRabbit comments on this PR:\n\n- [ ] Push a commit to this branch (recommended)\n- [ ] Create a new PR with the fixes\n\n

\n\n---\n\n

ℹ️ Review info

\n\n

⚙️ Run configuration

\n\n**Configuration used**: Path: .coderabbit.yaml\n\n**Review profile**: CHILL\n\n**Plan**: Pro\n\n**Run ID**: `7995cdbc-521f-4be3-b2b9-de2e58ecbf9e`\n\n

\n\n

📥 Commits

\n\nReviewing files that changed from the base of the PR and between 14e16a75c18fff90861a700e1391c8b6afe61d30 and 1b7c0238d62dae35f6b0d55a2f92078f8e8fc0a1.\n\n

\n\n

📒 Files selected for processing (13)

\n\n* `docs/branch-review-ledger.md`\n* `docs/outstanding-issues.md`\n* `src/components/account-data-provider.tsx`\n* `src/components/clinical-dashboard/favourites-hub.tsx`\n* `src/components/clinical-dashboard/saved-registry-favourites-status.ts`\n* `src/components/clinical-dashboard/use-saved-registry-favourites.ts`\n* `src/components/therapy-compass/bindings.tsx`\n* `src/components/therapy-compass/data/use-therapy-data.ts`\n* `tests/favourites-account-retry.dom.test.tsx`\n* `tests/favourites-hub-unavailable-controls.dom.test.tsx`\n* `tests/saved-registry-favourites-status.test.ts`\n* `tests/search-results-band-adoption.test.ts`\n* `tests/therapy-compass-data-recovery.dom.test.tsx`\n\n

\n\n

🚧 Files skipped from review as they are similar to previous changes (5)

\n\n* tests/saved-registry-favourites-status.test.ts\n* tests/favourites-account-retry.dom.test.tsx\n* tests/search-results-band-adoption.test.ts\n* docs/outstanding-issues.md\n* src/components/clinical-dashboard/use-saved-registry-favourites.ts\n\n

\n\n

\n\n","_🎯 Functional Correctness_ | _🟠 Major_ | _⚡ Quick win_\n\n**Scope retry settlement to its request key and attempt.**\n\n`retryWaitersRef` is global to the hook instance. If navigation changes `requestKey` while a retry is pending, the old effect becomes inactive, but the new route’s request calls `settleRetryWaiters()` and resolves the old retry promise. The original replacement request can therefore still be running—or the promise can remain pending until an unrelated request completes. Associate waiters with a retry token/request key and settle only from the matching effect; add a route-change-during-retry test.\n\n\n\n\n\n\nAlso applies to: 99-127\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@src/components/therapy-compass/data/use-therapy-data.ts` around lines 68 -\n90, Scope retry waiters to the request key and retry attempt in retry and\nsettleRetryWaiters, so a pending retry promise is resolved only by the matching\nrequest effect. Prevent a route change from allowing the new request to settle\nthe old retry, while ensuring the original retry is settled when its matching\nrequest completes; add a test covering navigation during an in-flight retry.\n```\n\n

\n\n\n\n\n\n\n\n\n\n✅ Addressed in commits b92c272 to 5911c2e",90,"**Heads-up for anyone else pushing to this branch: run `npm run format:check` before you push.**\n\nThe same failure has now broken CI twice — `77f6b0a9` fixed `tests/favourites-hub-unavailable-controls.dom.test.tsx`, and `9662d9b2` fixed `tests/search-results-band-adoption.test.ts`. Both were Prettier-only, no behaviour change.\n\nIt keeps recurring because **`format:check` is not part of `npm run test`, `npm run typecheck`, or `npm run lint`.** Both preceding closeouts reported a green unit suite, a clean typecheck and clean eslint — all accurate — and CI was still red, because `Static PR checks` and the new `ci/circleci: verify` job each run `prettier --check .` as a separate step. A green suite is not evidence of a green static gate.\n\n`npm run verify:cheap` does cover it, as does `npm run format:changed` if you want something quicker.\n\nWorth knowing about the CircleCI check specifically: it is **new**. `5bcdc1dd` on `main` replaced the CircleCI stub with a real verify job, so this PR is among the first to actually run it. It isn't misconfigured — it caught a genuine defect both times.\n\n**Verified on the current tip `9662d9b2`:**\n\n- `prettier --check .` — *All matched files use Prettier code style!*\n- `npm run typecheck`, `npm run lint` — exit 0\n- `npm run test` — **416 files, 4232 passed, 4 skipped**\n\nThat last one also closes the gap the previous closeout flagged, which noted the suite was green *before* the tip merge rather than after it. It is now green after.\n\nOne red still outstanding and unaddressed: **Gitleaks**. It is not reporting a secret — the log says `no leaks found in partial scan` — but it scanned `~0 bytes` because its commit range was invalid on the runner, so the secret scan did not actually execute on this PR. That is a genuine gap rather than noise, and it is not fixed by any of the above.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","\nFixed in `05f07d60` (tip includes merge with formatter). Confirmed the failure mode: `setFavourite` wrote mutation failures into the same `error` field that `foldSavedFavouritesStatus` treated as an unread library, so a failed first-save on an empty successful GET suppressed the truthful empty state and offered GET Retry.\n\n`AccountDataProvider` now tracks `loadError` (GET/reload only) separately from `error` (save/clear). The favourites fold/band reads only `loadError`. Added `tests/favourites-account-retry.dom.test.tsx` coverage for a successful empty GET followed by a failed PUT — status stays `ready`, `loadError` stays empty, action error is set.\n\n**Verification:** focused DOM + status tests pass; full unit suite was green before the tip merge (4232 passed / 4 skipped).","src/components/clinical-dashboard/saved-registry-favourites-status.ts","\nFixed in `05f07d60`. The route inventory now unions nested `/search/` pages with App Router pages resolved from every `resultsSurface: \"results-band\"` mode `href` (so `/services`, `/forms`, `/favourites`, `/formulation`, `/tools`, etc. are checked). Reachability follows two import hops into `@/components/**`, and landing-only homes (`/dsm`, `/factsheets`, `/therapy-compass`) are documented in `BAND_ROUTE_ALLOWLIST`. Added a negative fixture for a root-level page that never reaches the band.\n\n**Verification:** `tests/search-results-band-adoption.test.ts` 5/5; typecheck clean.","### Summary\n\n* Replaced substring-based adoption detection with TypeScript TSX parsing, ensuring imports, dead references, and commented-out `` markup cannot satisfy the structural gate. Mode IDs are now extracted from actual rendered band elements. [tests/search-results-band-adoption.test.tsL23-L45](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/tests/search-results-band-adoption.test.ts#L23-L45) [tests/search-results-band-adoption.test.tsL66-L87](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/tests/search-results-band-adoption.test.ts#L66-L87)\n* Added the exact commented-out JSX regression fixture alongside positive rendered-element coverage. [tests/search-results-band-adoption.test.tsL200-L214](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/tests/search-results-band-adoption.test.ts#L200-L214)\n* Threaded active document-search errors from `ClinicalDashboard` into the document results panel without affecting registry-backed services or forms. [src/components/ClinicalDashboard.tsxL3736-L3748](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/src/components/ClinicalDashboard.tsx#L3736-L3748)\n* Updated document result-band derivation so failed requests render an error state and message rather than exposing zero or stale counts. Normal result and empty-state content is suppressed while the request is faulted. [src/components/clinical-dashboard/document-search-results.tsxL876-L907](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/src/components/clinical-dashboard/document-search-results.tsx#L876-L907) [src/components/clinical-dashboard/document-search-results.tsxL974-L984](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/src/components/clinical-dashboard/document-search-results.tsx#L974-L984)\n* Added DOM coverage proving a failed document request displays the fault, suppresses a zero count, and does not make a false “No matching documents” claim. [tests/document-search-results-band.dom.test.tsxL16-L43](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/tests/document-search-results-band.dom.test.tsx#L16-L43)\n* Committed the changes as `bc0310b` (`fix(search): land results-band follow-up safeguards`) and created PR metadata titled **“Follow-up: harden search-band adoption and document failures.”**\n\n**Testing**\n\n* ✅ `npx vitest run tests/document-search-results-band.dom.test.tsx tests/search-results-band-adoption.test.ts` — **2 files passed, 5 tests passed**.\n* ✅ `npm run typecheck`\n* ✅ `npm run lint`\n* ✅ `npm run test` — full unit suite completed successfully after the focused regression was corrected.\n* ✅ `npm run check:runtime` — Node **24.18.0** and npm **11.16.0** satisfy repository requirements.\n* ✅ `npm run check:installed-lock-parity`\n* ✅ `npm run format:changed`\n* ✅ `npm run build` — production build compiled successfully, generated all **1,682** static pages, and passed the client-bundle secret scan.\n* ✅ `npm run check:rag:fixtures` — **36 golden cases across 21 suites passed**.\n* ✅ `git diff --check`\n* ✅ `git status --short --branch` — worktree clean on branch `work` after commit.\n\n [View task →](https://chatgpt.com/s/cd_6a686a9e6f4c8191a50bc13f19f08a64)","> Skipped: comment is from another GitHub bot.\n\n","docs/outstanding-issues.md","\nFixed in `9bace1d1`. Library count surfaces now wait on `libraryCountsTrusted` (`status === \"ready\"` or concrete items already present): Items/Sets show `—`, the tab trigger drops `· N`, listbox count badges hide, and the list-header badge is omitted. Filters may still show `0` — that is local UI state, not library inventory.\n\nCovered in `tests/favourites-hub-unavailable-controls.dom.test.tsx` for both pending and failed registry responses (demoMode off, empty items).\n\n**Verification:** focused DOM tests pass; full unit suite 4229 passed / 4 skipped.","src/components/clinical-dashboard/favourites-hub.tsx",54,"\nFixed in `9bace1d1`. Kept the error-first band status mapping so the fault panel stays mounted, and made `useTherapyData().retry` return a Promise that settles only when the **active** reload finishes (concurrent clicks coalesce on the in-flight promise). That feeds the band’s `await onRetry()` busy contract; the workspace Retry path already gates on `b.loading` and now also waits through the same settlement.\n\nAdded `tests/therapy-compass-data-recovery.dom.test.tsx` coverage: after Retry, the control stays on disabled `Retrying…` until the delayed catalogue response releases.\n\n**Verification:** focused recovery tests pass; full unit suite 4229 passed / 4 skipped.","src/components/therapy-compass/screens/search-screen.tsx",33,"\nFixed in `9bace1d1`. Confirmed the contradiction: the scheduled-comparison note claimed `#024` was independently resolved while the Open items table (and recommended queue) still list it as open Safari/`_rsc` reproduction work.\n\nI did **not** archive `#024` — that would have been wrong. The open-items row (Added 2026-07-28) is still valid. The prose now says `#024` remains open for native Safari/`_rsc` reproduction; release-matrix debt stays `#055`.\n\n**Verification:** `npm run check:branch-review-ledger` pass; full unit suite 4229 passed / 4 skipped.","\n> [!NOTE]\n> Autofix is a beta feature. Expect some limitations and changes as we gather feedback and continue to improve it.\n\n❌ **Cannot run autofix: This PR has merge conflicts.**\n\nPlease resolve the conflicts with the base branch and try again.\n\nAlternatively, use `@coderabbitai resolve merge conflict` to automatically resolve the conflicts.\n","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `4c6524a593`\n \n\n

ℹ️ About Codex in GitHub

","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Include top-level search pages in the adoption gate**\n\nThis route inventory only includes paths containing `/search/`, so production search surfaces such as `/services/page.tsx`, `/forms/page.tsx`, `/favourites/page.tsx`, `/formulation/page.tsx`, and `/tools/page.tsx` are never checked. The separate mode scan can still pass when a disconnected component contains a band mount, allowing one of those real routes to drop the band while the advertised reachability gate remains green. Discover search routes from the mode hrefs or explicitly include these top-level pages, with a negative fixture for a root-level search route.\n\nUseful? React with 👍 / 👎.","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Separate account-load errors from mutation failures**\n\nWhen an authenticated user with no existing favourites attempts to save their first item and the `PUT /api/account/favourites` request fails, `AccountDataProvider.setFavourite` restores the valid empty list but stores the mutation error in this same `accountError` field. This branch then reports the already-loaded library as unavailable, suppressing the truthful empty state and offering a GET retry for a failed write. Track initial-load failure separately from update/clear errors, and cover a failed first-save request where the favourites read previously succeeded.\n\nUseful? React with 👍 / 👎.","### Summary\n\n* Added a caller-controlled `retrying` state to `SearchResultsHeaderBand`, combining it with the component’s internal async state so Retry remains busy and disabled for callbacks that cannot return their request promise. [src/components/clinical-dashboard/search-results-header-band.tsxL82-L122](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/src/components/clinical-dashboard/search-results-header-band.tsx#L82-L122) [src/components/clinical-dashboard/search-results-header-band.tsxL423-L439](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/src/components/clinical-dashboard/search-results-header-band.tsx#L423-L439)\n* Connected Therapy Compass’s retained-error loading state to that busy state. Retry now remains disabled until the replacement dataset request settles, preventing repeated requests. [src/components/therapy-compass/screens/search-screen.tsxL27-L35](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/src/components/therapy-compass/screens/search-screen.tsx#L27-L35)\n* Added focused regression coverage proving the externally controlled Retry changes from disabled “Retrying…” back to enabled “Retry” only after the reload completes. [tests/search-results-header-band.dom.test.tsxL127-L153](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/tests/search-results-header-band.dom.test.tsx#L127-L153)\n* Committed the fix as `81ffb86` (`fix(therapy): keep Retry busy during reload`) and created the follow-up PR metadata titled **“Follow-up: keep Therapy Compass Retry busy during reload.”**\n\n**Testing**\n\n* ✅ `npm run workflow:design-sweep -- --write-evidence` — completed and generated the UI verification plan.\n* ✅ `git diff --check`\n* ✅ `git status --short --branch` — worktree clean after commit.\n* ⚠️ `npm run ensure` — environment has Node `20.20.2`, while the repository requires Node `24.x`; the local application and browser verification could not be started.\n* ⚠️ `npm run test -- tests/search-results-header-band.dom.test.tsx` — not runnable because `node_modules` is absent and installation is blocked by the Node runtime mismatch.\n\n [View task →](https://chatgpt.com/s/cd_6a686e0331348191bb9e4b71188849eb)","CI/review closeout for [#1316](https://github.com/BigSimmo/Database/pull/1316) is done on tip `1b7c0238`.\n\n### Root causes\n| Issue | Cause |\n|---|---|\n| CodeRabbit `#024` | Prose said “independently resolved” while the open-items row stayed open |\n| Codex favourites counts | Hub still rendered `0` / `· 0` while status was loading/error |\n| Codex therapy Retry | `retry()` returned `void`, so band busy ended immediately |\n| GitHub `CONFLICTING` | Stale vs `main` (merge-tree was clean); synced |\n| New Codex load/mutation | PUT failures reused the same `error` as GET failures → false “couldn’t load” |\n| New Codex adoption gate | Inventory only matched `/search/`, so mode homes like `/services` were unchecked |\n\n### Changes made\n- Corrected `#024` prose (item stays open; release-matrix debt remains `#055`)\n- Suppressed favourites library counts until trusted\n- Therapy `retry()` returns a Promise that settles with the active reload\n- Split `loadError` vs mutation `error` in `AccountDataProvider`\n- Expanded band adoption to mode-href pages + 2-hop reachability\n- Merged latest `main`; resolved all review threads addressed by these fixes\n\n### Validation\n- Full unit suite: **4232 passed / 4 skipped** (pre-final tip)\n- Focused recheck on tip: favourites + adoption **10/10**\n- `typecheck`, eslint on touched files, `check:branch-review-ledger` — pass\n- Bugbot (`pr-bugbot`): no P0/P1 on the WIP scope\n- **Unresolved review threads: 0**\n- Merge: `MERGEABLE`, merge-tree clean, **0 behind** `main`\n\n### Remaining blockers\n1. Hosted CI still settling on `1b7c0238` (checks queued) — wait for Production UI / PR required\n2. Residual product debt `#091` (partial multi-source status) — not this PR’s binary-status redesign\n3. Therapy search-band `onRetry` remains behind the workspace error gate; busy UX that matters is the workspace `loading` Retry path\n\n\n\n

\n\n","### Summary\n\n* Added a single readiness guard so favourites counts are considered trustworthy only after saved registry data reaches `ready`. [src/components/clinical-dashboard/favourites-hub.tsxL108-L115](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/src/components/clinical-dashboard/favourites-hub.tsx#L108-L115)\n* Suppressed the summary cards, selected-tab count, per-tab counts, and list-header count during loading and failed states, preventing the hub from asserting a misleading zero while data is unavailable. [src/components/clinical-dashboard/favourites-hub.tsxL186-L210](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/src/components/clinical-dashboard/favourites-hub.tsx#L186-L210) [src/components/clinical-dashboard/favourites-hub.tsxL230-L237](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/src/components/clinical-dashboard/favourites-hub.tsx#L230-L237) [src/components/clinical-dashboard/favourites-hub.tsxL293-L305](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/src/components/clinical-dashboard/favourites-hub.tsx#L293-L305) [src/components/clinical-dashboard/favourites-hub.tsxL397-L414](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/src/components/clinical-dashboard/favourites-hub.tsx#L397-L414)\n* Added hub-level regression coverage for loading and error states, including assertions across the summary, selected tab, list header, and opened type menu. [tests/favourites-hub-unavailable-controls.dom.test.tsxL37-L53](https://github.com/BigSimmo/Database/blob/14e16a75c18fff90861a700e1391c8b6afe61d30/tests/favourites-hub-unavailable-controls.dom.test.tsx#L37-L53)\n* Committed the change as `9840ed9` (`fix(favourites): hide untrusted hub counts`) and created the required follow-up PR metadata titled **“Follow-up: hide untrusted favourites hub counts.”**\n\n**Testing**\n\n* ✅ `git diff --check`\n* ⚠️ `npm run ensure` (environment uses Node `20.20.2`; the repository requires Node `24.x`, so the application could not be started for a screenshot)\n* ⚠️ `npm run test:focused -- tests/favourites-hub-unavailable-controls.dom.test.tsx` (the repository coordinator requires the full unit suite whenever a test file changes)\n* ⚠️ `npm run typecheck` (repository dependencies are not installed; `node_modules/typescript/bin/tsc` is unavailable)\n* ⚠️ `npm run lint` (repository dependencies are not installed; `node_modules/eslint/bin/eslint.js` is unavailable)\n* ✅ `git status --short --branch` (worktree clean after commit)\n\n [View task →](https://chatgpt.com/s/cd_6a686deeb5688191bc814968566503e6)","@Cursoragent Inspect the PR and repository state, identify the root cause of every failing CI check, unresolved review comment, merge conflict, and related defect. Apply the smallest robust fix using existing project patterns and best practices, without unrelated changes. Run Bugbot and all relevant lint, type-check, test, build, and CI-equivalent commands, fix any issues found, then resolve only the review comments that are fully addressed. Summarise the root causes, changes made, validation results, and any remaining blockers.","\n> [!NOTE]\n> Autofix is a beta feature. Expect some limitations and changes as we gather feedback and continue to improve it.\n\n❌ **Cannot run autofix: This PR has merge conflicts.**\n\nPlease resolve the conflicts with the base branch and try again.\n\nAlternatively, use `@coderabbitai resolve merge conflict` to automatically resolve the conflicts.\n","\n> [!NOTE]\n> Autofix is a beta feature. Expect some limitations and changes as we gather feedback and continue to improve it.\n\n❌ **Cannot run autofix: This PR has merge conflicts.**\n\nPlease resolve the conflicts with the base branch and try again.\n\nAlternatively, use `@coderabbitai resolve merge conflict` to automatically resolve the conflicts.\n","@Codex fix and resolve ","@Codex fix and resolve","@Curseragent Inspect the PR and repository state, identify the root cause of every failing CI check, unresolved review comment, merge conflict, and related defect. Apply the smallest robust fix using existing project patterns and best practices, without unrelated changes. Run Bugbot and all relevant lint, type-check, test, build, and CI-equivalent commands, fix any issues found, then resolve only the review comments that are fully addressed. Summarise the root causes, changes made, validation results, and any remaining blockers.","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `14e16a75c1`\n \n\n

ℹ️ About Codex in GitHub

","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Keep the therapy Retry busy until the reload settles**\n\nAfter a therapy dataset failure, `retryData()` deliberately leaves the prior error set while changing `loading` to true, but this error-first status mapping keeps the band faulted throughout the retry. Because `retryData` returns `void`, the band's `await onRetry()` also finishes immediately, so Retry never remains busy or disabled and repeated clicks restart the cached dataset request while the first retry is still running. Tie the button's busy state to `b.loading` or return the reload promise, with a test that Retry stays disabled until the replacement request settles.\n\nUseful? React with 👍 / 👎.","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Hide favourites counts until the saved data is trustworthy**\n\nWhen the account or registry request is loading or fails with no demo items, this newly retained status correctly changes the empty-state copy, but the hub still unconditionally renders `Items 0`, `All · 0`, per-tab zeroes, and the list-header zero at lines 185-203, 228-233, and 404-405. Fresh evidence after the claimed hook fix is that `savedRegistryStatus` is consumed only by the empty-state text and Retry control, so the page still asserts an empty library while saying it could not load one; suppress these count-bearing surfaces until the status is `ready`, and cover the whole hub for pending and failed account responses.\n\nUseful? React with 👍 / 👎.","**Actionable comments posted: 1**\n\n

🤖 Prompt for all review comments with AI agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nInline comments:\nIn `@docs/outstanding-issues.md`:\n- Around line 93-131: Reconcile the `#024` entry in the outstanding-issues\ndocument: it is marked independently resolved in the update but remains in the\nopen-items table. Move `#024` to the resolved archive or update its status\nconsistently so the /issues view reports only one authoritative state.\n```\n\n

\n\n

🪄 Autofix (Beta)

\n\n❌ Autofix failed (check again to retry)\n\nFix all unresolved CodeRabbit comments on this PR:\n\n- [ ] Push a commit to this branch (recommended)\n- [ ] Create a new PR with the fixes\n\n

\n\n---\n\n

ℹ️ Review info

\n\n

⚙️ Run configuration

\n\n**Configuration used**: Path: .coderabbit.yaml\n\n**Review profile**: CHILL\n\n**Plan**: Pro\n\n**Run ID**: `bf8cbd34-3397-46d2-be0a-e5415404e38a`\n\n

\n\n

📥 Commits

\n\nReviewing files that changed from the base of the PR and between 8ae374acd57abc3926aaea367c7d76de654c32c7 and 14e16a75c18fff90861a700e1391c8b6afe61d30.\n\n

\n\n

📒 Files selected for processing (9)

\n\n* `docs/outstanding-issues.md`\n* `src/components/account-data-provider.tsx`\n* `src/components/clinical-dashboard/document-search-results.tsx`\n* `src/components/clinical-dashboard/document-search-unavailable-status.ts`\n* `src/components/clinical-dashboard/search-results-header-band.tsx`\n* `src/components/clinical-dashboard/use-saved-registry-favourites.ts`\n* `tests/document-search-unavailable-status.test.ts`\n* `tests/favourites-account-retry.dom.test.tsx`\n* `tests/ui-accessibility.spec.ts`\n\n

\n\n

🚧 Files skipped from review as they are similar to previous changes (2)

\n\n* src/components/clinical-dashboard/use-saved-registry-favourites.ts\n* src/components/clinical-dashboard/search-results-header-band.tsx\n\n

\n\n

\n\n","_🗄️ Data Integrity & Integration_ | _🟡 Minor_ | _⚡ Quick win_\n\n**Reconcile the contradictory `#024` status.**\n\nThe update above says `#024` is independently resolved, but the Open items table still lists it as an unresolved P2 issue. Move it to the resolved archive or correct the status so `/issues` cannot report contradictory state.\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@docs/outstanding-issues.md` around lines 93 - 131, Reconcile the `#024` entry\nin the outstanding-issues document: it is marked independently resolved in the\nupdate but remains in the open-items table. Move `#024` to the resolved archive or\nupdate its status consistently so the /issues view reports only one\nauthoritative state.\n```\n\n

\n\n\n\n\n\n\n\n","## Full `verify:ui` result — first complete run this PR has had\n\n`Production UI` in CI has been cancelled on every attempt (eleven times; each red `PR required` was `COVERAGE_RESULT: cancelled` / `UI_RESULT: cancelled` from a superseding push, never a test failure), so I ran the gate locally to end the guessing.\n\n**`npm run verify:ui` — 320 passed, 4 failed (9.9m).** Against an isolated *production* build. Note this corrects something I wrote earlier: `run-playwright.mjs` does a real `next build`, it does **not** use `next dev`.\n\nEnvironment caveat, unchanged: this box has Chromium `1194`, the project pins `1228`, so the run used the config's `PLAYWRIGHT_CHROMIUM_EXECUTABLE_PATH` hook. A reviewer should still re-run on the pinned build.\n\n### The 4 failures, triaged\n\n| Test | Cause | This PR? |\n| --- | --- | --- |\n| `ui-tools.spec.ts:563` forms home | strict mode: `forms-home` resolved to 2 | **No** |\n| `ui-smoke.spec.ts:3001` favourites library | strict mode: favourite row button resolved to 2 | **No** |\n| `ui-smoke.spec.ts:3526` document viewer | PDF.js canvas not rendering | **No** — browser-build mismatch |\n| `ui-pwa.spec.ts:143` manifest/worker | installability reports `in-incognito` | **No** — environmental |\n\n**I verified the first two rather than assuming.** Both pass when run alone:\n\n```\n✓ 1 … › header mode switches open clean services and forms home pages (2.3s) 1 passed\n✓ 1 … › favourites command library exposes truthful item details … (1.4s) 1 passed\n```\n\nThey only fail inside the full suite. In both, copy 1 of the duplicate is nested under `mobile-composer-reserve-pad` — the identical signature to the `differentials-search-results` failure CI hit earlier. So it is the known duplicate-page-root condition, load/order-dependent, and **not** caused by this diff. Recorded as `#093` in `docs/outstanding-issues.md`, now with a local reproduction, which is new — I could not reproduce it before today.\n\nTo be explicit about what that does **not** prove: I have not shown the duplicate pre-exists on `main`, only that it is order-dependent and that neither failing surface is touched by this diff in the relevant structure. If a reviewer wants that baseline it's a `main` run of the same two specs.\n\n## Two Codex commits are not on this branch\n\n`@codex` was asked to fix the adoption-gate and document-status findings and reports committing `ff5b682` and `77cfe12` — but into **separate follow-up PRs**, not here. Its own testing notes say the suite could not run in that environment (`node_modules` missing; Node 20 against a repo requiring 24), so neither change has been executed anywhere yet.\n\nFlagging rather than acting: those two need to land here and be run before this merges, and `77cfe12` touches the same status derivation I changed in `a7fadf55`, so it will need reconciling rather than a clean apply.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","\nFixed in `9f055e42`. Added `flex-wrap` to the fault action row.\n\nConfirmed the mechanism you described: differentials passes Retry plus `Browse presentations` and `Browse diagnoses`, and the band root is `overflow-hidden` (`search-results-header-band.tsx:193`), so the trailing link was cut off rather than moving to a second line.\n\nCovered at 390px in `tests/ui-accessibility.spec.ts` — the case asserts every action sits inside the fault panel's horizontal bounds **and** that the three occupy more than one line, so a future change that merely lets them overflow silently cannot pass.\n\n**I checked the case actually catches it.** Reverting to `flex justify-center gap-2` fails on `expect(received).toBeGreaterThan(expected) / Expected: > 1 / Received: 1` — all three sharing one line, which is the clipped state. With `flex-wrap` it passes.\n\nI used 390px rather than 320px because that is the width the rest of this repo's phone specs use, and it already reproduces — the three controls exceed the panel there.\n\n**Verification:** `npm run typecheck`, `npm run lint` exit 0; `npm run test` — **412 files, 4164 passed, 4 skipped**; the Playwright case run directly against the isolated production build, both with and without the fix.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","src/components/clinical-dashboard/search-results-header-band.tsx","\nFixed in `9f055e42`. Confirmed against the code first — you were right, and this is the repo's own wiring rule broken by my change: a control that advertises an action must perform one.\n\n`AccountDataContextValue` exposed no reload at all, so `refetch()` could only call two registry refetches that a failed account load had already disabled. Added `reload` to the provider. It bumps an attempt counter and re-runs the **existing** load effect rather than introducing a second fetch path, so an explicit Retry gets exactly the same clearing and abort semantics as an auth transition — the identity-clearing invariant in `use-differential-catalog.ts:122` is untouched.\n\nThe hook reissues the account request first, since when that is what failed the registries hold nothing left to re-request.\n\nTest is the one you specified — `tests/favourites-account-retry.dom.test.tsx`: the account endpoint fails once, Retry is clicked, the second request succeeds and the band returns to `ready` with the recovered item.\n\n**I checked it isn't a tautology.** With the `reloadAccount()` line removed it fails on `expected \"vi.fn()\" to be called 2 times, but got 1 times` — the exact failure mode you described.\n\n**Verification:** `npm run typecheck`, `npm run lint` exit 0; `npm run test` — **412 files, 4164 passed, 4 skipped**.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","src/components/clinical-dashboard/use-saved-registry-favourites.ts","### Summary\n\n* Passed the active document-search error into `DocumentSearchResultsPanel`, scoped only to document searches so registry-backed services and forms retain their independent status handling. [src/components/ClinicalDashboard.tsxL3731-L3744](https://github.com/BigSimmo/Database/blob/abc8eb57179177d7237339e5f0a38d58f911593a/src/components/ClinicalDashboard.tsx#L3731-L3744)\n* Updated the results-band status derivation so a rejected active document request renders the fault state and suppresses zero or stale match counts. The request error is also displayed as the fault detail when no higher-priority availability message exists. [src/components/clinical-dashboard/document-search-results.tsxL888-L907](https://github.com/BigSimmo/Database/blob/abc8eb57179177d7237339e5f0a38d58f911593a/src/components/clinical-dashboard/document-search-results.tsx#L888-L907)\n* Added a focused DOM regression test confirming a failed document search displays “Couldn't search,” surfaces the request error, and does not expose a `0` count. [tests/document-search-results-band.dom.test.tsxL6-L33](https://github.com/BigSimmo/Database/blob/abc8eb57179177d7237339e5f0a38d58f911593a/tests/document-search-results-band.dom.test.tsx#L6-L33)\n* Committed the changes as `77cfe12` (`fix(documents): mark failed searches in results band`) and created the required follow-up PR metadata with the title **“Fix document result-band status after failed searches.”**\n\n**Testing**\n\n* ✅ `git diff --check`\n* ✅ `npm run test:focused -- tests/document-search-results-band.dom.test.tsx` (the repository coordinator correctly rejected focused execution because a test/configuration path changed and requires the full suite)\n* ⚠️ `npx vitest run tests/document-search-results-band.dom.test.tsx` (could not run because the repository dependencies were not installed; transient Vitest lacked `jsdom`)\n* ⚠️ `npm ci --ignore-scripts` (environment limitation: installed Node is `v20.20.2`, while the project requires Node `24.x`; therefore the full unit suite could not be run)\n* ✅ `git status --short --branch` (worktree clean after commit)\n\n [View task →](https://chatgpt.com/s/cd_6a6867c98f4c81919e1b24c59bfdf9ac)",910,"### Summary\n\n* Replaced raw substring matching with TypeScript AST traversal, so imports, prose comments, and commented-out JSX no longer satisfy the search-results-band adoption gate. [tests/search-results-band-adoption.test.tsL23-L48](https://github.com/BigSimmo/Database/blob/abc8eb57179177d7237339e5f0a38d58f911593a/tests/search-results-band-adoption.test.ts#L23-L48)\n* Updated `modeId` extraction to inspect attributes on parsed `SearchResultsHeaderBand` JSX elements rather than scanning nearby source text. [tests/search-results-band-adoption.test.tsL68-L90](https://github.com/BigSimmo/Database/blob/abc8eb57179177d7237339e5f0a38d58f911593a/tests/search-results-band-adoption.test.ts#L68-L90)\n* Added regression coverage for the exact reported case: `{/* */}` is rejected, while a rendered JSX element is accepted. [tests/search-results-band-adoption.test.tsL203-L218](https://github.com/BigSimmo/Database/blob/abc8eb57179177d7237339e5f0a38d58f911593a/tests/search-results-band-adoption.test.ts#L203-L218)\n* Committed the change as `ff5b682` (`test(search): parse JSX for band adoption`) and created the required follow-up PR metadata titled **“Follow-up: parse rendered JSX in search-band adoption gate.”**\n\n**Testing**\n\n* ✅ `git diff --check`\n* ⚠️ `npm run test -- tests/search-results-band-adoption.test.ts` (environment limitation: dependencies are not installed; `node_modules/vitest/vitest.mjs` is missing)\n* ✅ `git status --short --branch` (worktree clean after commit)\n\n [View task →](https://chatgpt.com/s/cd_6a6867bdf2588191b08ca19276c24435)",28,"@Codex fox and resolve comment ","@Codex fix and resolve comment ","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Parse rendered JSX before accepting band adoption**\n\nWhen a route comments out its mount as `{/* */}`, this raw substring check still reports adoption, allowing the production route to lose the band while the gate stays green. Fresh evidence at this head is that the added negative fixture uses only a prose comment without the `_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)} Derive status from the active document request**\n\nWhen a document search request rejects while setup and background API health remain available, `executeSearch` sets `error` and ends `loading` without changing `apiUnavailable`; this branch consequently resolves to `ready` and asserts zero or stale matches for the failed query beneath the separate error panel. Pass the active search failure into this status derivation so only a successful request can expose its count.\n\nUseful? React with 👍 / 👎.","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `abc8eb5717`\n \n\n

ℹ️ About Codex in GitHub

","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `a7fadf5528`\n \n\n

ℹ️ About Codex in GitHub

","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Wrap fault recovery actions on phones**\n\nOn a narrow phone, the differential error passes Retry plus two long browse links into this non-wrapping flex row; their combined width exceeds the fault panel, and the band root uses `overflow-hidden`, so the trailing recovery action is clipped rather than wrapping onto another line. Add wrapping here and cover the three-action fault state at the smallest supported viewport.\n\nAGENTS.md reference: [AGENTS.md:L180-L182](https://github.com/BigSimmo/Database/blob/a7fadf55288342289bc7782fe4b09fe28125f16a/AGENTS.md#L180-L182)\n\nUseful? React with 👍 / 👎.","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Refetch account favourites from Retry**\n\nWhen the authenticated `/api/account/favourites` request fails, the folded status becomes `error` and both favourites views offer Retry, but the failed request has cleared every saved slug, so this callback invokes neither registry refetch and cannot reissue the account request. The button therefore leaves the page faulted indefinitely; expose an account-data reload and add a test where the account endpoint fails once and succeeds after Retry.\n\nAGENTS.md reference: [AGENTS.md:L180-L182](https://github.com/BigSimmo/Database/blob/a7fadf55288342289bc7782fe4b09fe28125f16a/AGENTS.md#L180-L182)\n\nUseful? React with 👍 / 👎.","Fixed in `a7fadf55`.\n\nThe finding is correct — `authUnavailable` produced the right message (\"Sign in again to view private indexed documents\") under the wrong headline. The band rendered its generic **\"Couldn't search\"**, which tells the reader the search broke when in fact it never ran for want of a session. That is the same class of untruthful failure state this branch exists to remove, so thank you for catching it on a path I'd migrated.\n\n**I did not apply the proposed diff as written.** Putting the `authUnavailable` arm ahead of `unavailableMessage` inverts the message chain's own precedence: `apiUnavailable` is checked first when building the message, so with both flags true the diff would pair *\"Clinical KB could not be reached\"* with an `unauthorized` status — telling the user to sign in when the fix is unreachable infrastructure.\n\nThe root cause is that the message and the status were derived from two independent chains, so they could disagree. Both now come from one function, `deriveDocumentSearchUnavailable`, which returns `{ message, status }` together and cannot drift. Precedence is explicit and pinned by a test: an unreachable API outranks an expired session, because signing back in cannot fix an API that is down.\n\nTest added as requested — `tests/document-search-unavailable-status.test.ts`, 7 cases including the both-flags-true precedence case. I unit-tested the derivation rather than mounting `DocumentSearchResults`, following the `saved-registry-favourites-status.ts` pattern already in that directory.\n\n**Verification:** `npm run typecheck` and `npm run lint` exit 0; `npm run test` — **411 files, 4163 passed, 4 skipped**. Browser gates are unchanged by this commit and not re-run here.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","\n\n> [!CAUTION]\n> Some comments are outside the diff and can’t be posted inline due to platform limitations.\n> \n> \n> \n>

\n>

⚠️ Outside diff range comments (1)

\n> \n>
\n>
src/components/clinical-dashboard/document-search-results.tsx (1)
\n> \n> `894-909`: _🎯 Functional Correctness_ | _🟠 Major_ | _⚡ Quick win_\n> \n> **Map expired document sessions to `unauthorized`, not `error`.**\n> \n> `authUnavailable` creates `unavailableMessage` at Lines 868-869, but this branch maps it to `\"error\"` at Lines 903-907. The band consequently presents a broken-search state instead of its distinct sign-in-required status.\n> \n>
\n>
Proposed fix
\n> \n> ```diff\n> status={\n> showRecordMatches\n> ? recordStatus === \"unauthorized\"\n> ? \"unauthorized\"\n> : recordStatus === \"error\" || recordStatus === \"not_found\"\n> ? \"error\"\n> : recordStatus === \"loading\"\n> ? \"loading\"\n> : \"ready\"\n> - : unavailableMessage\n> + : authUnavailable\n> + ? \"unauthorized\"\n> + : unavailableMessage\n> ? \"error\"\n> : loading\n> ? \"loading\"\n> : \"ready\"\n> }\n> ```\n>
\n> \n> Add a documents-path test for `authUnavailable` asserting the band announces “Sign in to search.”\n> \n>
\n>
🤖 Prompt for AI Agents
\n> \n> ```\n> Verify each finding against current code. Fix only still-valid issues, skip the\n> rest with a brief reason, keep changes minimal, and validate.\n> \n> In `@src/components/clinical-dashboard/document-search-results.tsx` around lines\n> 894 - 909, Update the documents-path status mapping in the band component around\n> the showRecordMatches/unavailableMessage branch so authUnavailable-derived\n> unavailableMessage produces the unauthorized status instead of error, preserving\n> existing loading and generic error behavior. Add a documents-path test covering\n> authUnavailable and assert the band announces “Sign in to search.”\n> ```\n> \n>
\n> \n> \n> \n>
\n> \n>

\n\n

🤖 Prompt for all review comments with AI agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nOutside diff comments:\nIn `@src/components/clinical-dashboard/document-search-results.tsx`:\n- Around line 894-909: Update the documents-path status mapping in the band\ncomponent around the showRecordMatches/unavailableMessage branch so\nauthUnavailable-derived unavailableMessage produces the unauthorized status\ninstead of error, preserving existing loading and generic error behavior. Add a\ndocuments-path test covering authUnavailable and assert the band announces “Sign\nin to search.”\n```\n\n

\n\n---\n\n

ℹ️ Review info

\n\n

⚙️ Run configuration

\n\n**Configuration used**: Path: .coderabbit.yaml\n\n**Review profile**: CHILL\n\n**Plan**: Pro\n\n**Run ID**: `94fa6f0e-330d-4dcd-a328-23c5e98d7d44`\n\n

\n\n

📥 Commits

\n\nReviewing files that changed from the base of the PR and between 4cc9f21706e0e541e3bf780957a65212ab158b0b and 8ae374acd57abc3926aaea367c7d76de654c32c7.\n\n

\n\n

📒 Files selected for processing (17)

\n\n* `docs/branch-review-ledger.md`\n* `docs/search-chrome-behaviour.md`\n* `src/components/clinical-dashboard/differentials-home.tsx`\n* `src/components/clinical-dashboard/document-search-results.tsx`\n* `src/components/clinical-dashboard/favourites-command-library-page.tsx`\n* `src/components/clinical-dashboard/favourites-hub.tsx`\n* `src/components/clinical-dashboard/saved-registry-favourites-status.ts`\n* `src/components/clinical-dashboard/search-results-header-band.tsx`\n* `src/components/clinical-dashboard/use-saved-registry-favourites.ts`\n* `src/components/therapy-compass/screens/search-screen.tsx`\n* `tests/favourites-auth-gate.dom.test.tsx`\n* `tests/favourites-hub-unavailable-controls.dom.test.tsx`\n* `tests/mode-menu-prefetch.dom.test.tsx`\n* `tests/saved-registry-favourites-status.test.ts`\n* `tests/search-results-band-adoption.test.ts`\n* `tests/search-results-header-band.dom.test.tsx`\n* `tests/ui-accessibility.spec.ts`\n\n

\n\n

🚧 Files skipped from review as they are similar to previous changes (9)

\n\n* tests/favourites-hub-unavailable-controls.dom.test.tsx\n* tests/mode-menu-prefetch.dom.test.tsx\n* src/components/therapy-compass/screens/search-screen.tsx\n* tests/search-results-band-adoption.test.ts\n* src/components/clinical-dashboard/favourites-command-library-page.tsx\n* tests/ui-accessibility.spec.ts\n* docs/search-chrome-behaviour.md\n* src/components/clinical-dashboard/differentials-home.tsx\n* src/components/clinical-dashboard/search-results-header-band.tsx\n\n

\n\n

\n\n","Fixed in the follow-up tip. `foldSavedFavouritesStatus` now treats an authenticated unsettled/failed `/api/account/favourites` response as loading/error before registry fold, with unit coverage for pending and failed account responses so a ready zero/empty-library claim cannot appear.",96,"**Reopening this concern — the fix in `8ccd7f48` may have inverted the safety direction.** Flagging rather than changing anything, since Cursor owns the branch.\n\nThe hook now does:\n\n```js\nconst status = items.length > 0 && registryStatus !== \"ready\" ? \"ready\" : registryStatus;\n```\n\nand the page consumes the masked value:\n\n```js\n1037: status: favouritesRegistryStatus, // the masked status, not registryStatus\n1207: status={favouritesRegistryStatus}\n1208: onRetry={favouritesRegistryStatus === \"error\" ? refetchFavouritesRegistry : undefined}\n1250: … && favouritesRegistryStatus === \"ready\" ? \n```\n\nSo for a user holding any local favourite while the services/forms registry fails:\n\n- the band reports `ready` and shows a **confident count that silently omits every item from the failed registry**\n- no fault panel — `faulted` is false\n- no Retry — the guard tests `=== \"error\"`, which the mask has just turned into `\"ready\"`\n- and if filtering empties the list, it says **\"No matches\"**\n\n`registryStatus` carries the true fold, but I can't find a consumer for it, so the failure appears nowhere.\n\nThat is the defect this PR exists to remove: a number presented as complete when the underlying read did not finish. It is the original \"0 matches\" bug in a different costume, now on favourites.\n\nMy earlier note on this thread argued the pre-existing behaviour **over**-reported failure (hid a valid count), which is the conservative direction for a clinical surface, and that the correct fix needs a `partial` state or an explicit incompleteness marker — not a mask that makes an incomplete count indistinguishable from a complete one.\n\n**I may be wrong** — if a notice renders from `registryStatus` somewhere I haven't found, this concern dissolves, and I'd welcome the pointer.\n\nOtherwise the smallest safe options are: surface `registryStatus` on the page as a notice plus Retry so the count is visibly partial; or revert this masking and leave the thread open for the `partial`-state design.\n\nPlease don't treat this thread as settled on the strength of the count staying visible — visibility was never the problem.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","src/components/clinical-dashboard/favourites-command-library-page.tsx",1211,"src/components/clinical-dashboard/differentials-home.tsx","docs/search-chrome-behaviour.md",43,"Fixed in `8ccd7f48`. `useSavedRegistryFavourites` now keeps band `status` at `ready` when unaffected sources already produced items (e.g. local differentials while a service/form registry failed), so a valid nonzero count stays visible; `registryStatus` exposes the raw fold for callers that need it.","Fixed in `8ccd7f48`. The hook now returns `refetch` from the requested registries; favourites command-library wires it to the band’s `onRetry`, and the hub empty-fault state exposes a Retry button.","Addressed on tip (landed in `db592b26`, covered by a DOM test in `8ccd7f48`). The band treats `loading` like a fault for count-bearing page controls (`countUntrusted`), so ResultTabs/`Forms 0` no longer render under “Searching…”.","src/components/forms/forms-search-results-page.tsx",636,"Fixed in `8ccd7f48`. The adoption test now fails if `.search-band` is absent from the final forced-colors block and asserts the selector sits between that block’s opener and closer.","Fixed in `8ccd7f48`. Clarified that the query is the only heading element, with weights 560 (query) and 580 (figure) as nearby steps on the same scale rather than conflicting “heading-weight” claims.","Fixed in `8ccd7f48`. Unauthorized fault body now says “Sign in again…”, matching the omitted Retry affordance; error still promises retry.","Addressed on tip (`db592b26` + `8ccd7f48`). Adoption now requires a rendered `\n> [!NOTE]\n> Autofix is a beta feature. Expect some limitations and changes as we gather feedback and continue to improve it.\n\n⚠️ **Branch updated during autofix.**\n\nThe branch was updated while autofix was in progress. Please try again.\n","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `db592b26dc`\n \n\n

ℹ️ About Codex in GitHub

","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Fold account-data failures into favourites status**\n\nWhen an authenticated user’s `/api/account/favourites` request is still pending or fails, `AccountDataProvider` leaves `favourites` empty and exposes the condition through `ready`/`error`, but this hook discards both and folds only the downstream registry statuses. With no slugs yet, `requested` contains only nulls and reports `ready`, so the favourites band and empty state confidently show zero/no matches even though the saved-items request never completed. Include the account-data state in this status and add a pending/500 account-response test asserting that neither a count nor an empty-library claim appears.\n\nAGENTS.md reference: [AGENTS.md:L233-L237](https://github.com/BigSimmo/Database/blob/db592b26dc0663ebc919cc92f396d866c4110df1/AGENTS.md#L233-L237)\n\nUseful? React with 👍 / 👎.","CI/review closeout for [#1316](https://github.com/BigSimmo/Database/pull/1316) is done on tip `8ae374ac`.\n\n### Root causes\n| Issue | Cause |\n|---|---|\n| Prior Production UI red (`6a4810c`) | Suspense duplicate `global-search-input` strict-mode (already hardened earlier; not on current tip) |\n| Codex/CodeRabbit threads | Favourites status incomplete; unauthorized copy promised Retry; forced-colors gate soft-passed; docs wording conflict |\n| New Codex P2 | Account `/api/account/favourites` pending/failure ignored → false “ready / 0 matches” |\n| Merge conflicts | None (`merge-tree` clean) |\n\n### Changes made\n- Differentials unauthorized `faultBody` no longer promises Retry\n- Favourites: account readiness folded with registry status; partial/demo items keep a truthful nonzero band count; `refetch` wired to Retry\n- Adoption gate: real JSX mount + forced-colors membership in the final block\n- Band DOM test for loading suppressing count controls\n- Docs typography clarified\n\n### Validation\n- Focused vitest **43/43**, `tsc` + eslint on touched files **PASS**\n- All **19** review threads **resolved**\n- Bugbot via `pr-bugbot` (no Bugbot MCP here): residual note that a binary band status can’t express true multi-source “partial” without a dedicated shape\n\n### Remaining blockers\n1. **Hosted CI still settling** on the latest push — wait for Production UI / PR required\n2. Branch is **1 commit behind `main`** (clean merge-tree) — sync before merge\n3. Full `verify:ui` / `verify:pr-local` not re-run in this pass (focused gates only)\n\n\n\n

\n\n","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `65ba242837`\n \n\n

ℹ️ About Codex in GitHub

","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Require an actual band render in the adoption gate**\n\nThis shortcut treats any occurrence of `SearchResultsHeaderBand` in a route as adoption, including an import, comment, dead branch, or unused reference, so a production search route can omit the band while this structural gate remains green. The one-hop component check has the same substring weakness; parse rendered JSX or otherwise verify a reachable element, with a negative fixture that imports the identifier without mounting it.\n\nUseful? React with 👍 / 👎.","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Suppress count controls during initial loading**\n\nUnlike the resolved fault-state case, an initial forms-registry load still mounts these controls: `displayedMatches` is forced to `[]` until `registryReady`, so `ResultTabs` visibly asserts “Forms 0” while the band's status says “Searching…”. Loading explicitly means no trustworthy count exists, so suppress count-bearing controls for `loading` as well as faults; a delayed registry-response DOM test should verify that no zero appears before readiness.\n\nUseful? React with 👍 / 👎.","\n> [!NOTE]\n> Autofix is a beta feature. Expect some limitations and changes as we gather feedback and continue to improve it.\n\n⚠️ **Branch updated during autofix.**\n\nThe branch was updated while autofix was in progress. Please try again.\n","**Actionable comments posted: 10**\n\n

🧹 Nitpick comments (2)

\n\n
\n
src/components/clinical-dashboard/search-results-header-band.tsx (1)
\n\n`410-435`: _🩺 Stability & Availability_ | _🔵 Trivial_ | _⚡ Quick win_\n\n**Guard the retry against a rejected `onRetry`.**\n\n`await onRetry()` has no `catch`, so a rejecting recovery callback (several call sites pass `registry.refetch` / `b.retryData` directly) escapes as an unhandled rejection while the panel silently returns to the idle \"Retry\" label with no feedback. `finally` only resets the spinner.\n\n
\n
🛡️ Proposed fix
\n\n```diff\n onClick={async () => {\n setRetrying(true);\n try {\n await onRetry();\n+ } catch {\n+ // The fault panel is already the failure surface; a failed\n+ // retry simply leaves it in place rather than crashing.\n } finally {\n setRetrying(false);\n }\n }}\n```\n
\n\n
\n
🤖 Prompt for AI Agents
\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@src/components/clinical-dashboard/search-results-header-band.tsx` around\nlines 410 - 435, Update the retry handler on the AsyncButton in the retry action\nblock to catch rejected onRetry calls, preserving the existing\nsetRetrying(false) cleanup while surfacing the failure through the component’s\nestablished error or feedback mechanism instead of allowing an unhandled\nrejection.\n```\n\n
\n\n\n\n
\n
\n
src/components/clinical-dashboard/medication-prescribing-workspace.tsx (1)
\n\n`471-472`: _🩺 Stability & Availability_ | _🔵 Trivial_ | _⚡ Quick win_\n\n**Add a retry action to the medication catalogue fault panel.** `useMedicationCatalog` only returns `data/loading/error`, so this panel has no recovery path. Expose a `refetch`/`retry` from the hook and pass it into `SearchResultsHeaderBand` as `onRetry` (or `faultAction`).\n\n
\n
🤖 Prompt for AI Agents
\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@src/components/clinical-dashboard/medication-prescribing-workspace.tsx`\naround lines 471 - 472, Expose a refetch or retry callback from\nuseMedicationCatalog, then pass it as onRetry or faultAction to the\nSearchResultsHeaderBand rendering the medication catalogue fault panel. Preserve\nthe existing catalog error, loading, and ready status behavior while enabling\nusers to retry failed requests.\n```\n\n
\n\n\n\n
\n\n

\n\n

🤖 Prompt for all review comments with AI agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nInline comments:\nIn `@docs/search-chrome-behaviour.md`:\n- Around line 53-60: Update the typography guidance in the section describing\nheading weights to consistently state whether the query and figure both use\nweight 580, or only the query is emphasized; align the opening claim with the\nexplicit weight values and clarify the intended hierarchy for implementers.\n\nIn `@src/components/clinical-dashboard/differentials-home.tsx`:\n- Around line 919-929: Update the faultBody expression in the catalog fault\npanel to branch on catalog.status like faultTitle, omitting retry guidance for\nunauthorized responses while retaining the retry and browse guidance for other\nfailures; keep onRetry behavior unchanged.\n- Around line 992-1007: Update the main catalogue-content branching around the\n`!best && !catalogFailed` empty-state condition so `catalogFailed` suppresses\nthe entire results body, preventing the results grid from rendering when `best`\nis null. Preserve the existing fault panel as the sole failure UI and leave the\nguarded `DifferentialsMobileCompareBar` behavior unchanged.\n\nIn `@src/components/clinical-dashboard/document-search-results.tsx`:\n- Around line 894-909: Update the record-mode handling in the status/faultBody\nlogic and the related standalone alert near the corresponding duplicate block so\nan unavailableMessage remains surfaced when the ribbon does not carry it.\nPreserve recordStatus-based status and valid ribbon counts, but avoid setting\nfaultBody and alert visibility in a way that drops auth, API, or setup notices\non services and forms record paths.\n\nIn `@src/components/clinical-dashboard/favourites-command-library-page.tsx`:\n- Around line 1198-1201: Update the empty-results rendering in the favourites\ncommand library page so SearchResultsEmptyState is shown only when\nfavouritesRegistryStatus is ready. For loading, error, or unauthorized statuses,\nsuppress the normal “No matches” branch and use the existing recovery action for\nfaulted states; retain the current empty-state behavior for successfully loaded\nresults.\n\nIn `@src/components/clinical-dashboard/favourites-hub.tsx`:\n- Line 50: Update the favourites data handling around useSavedRegistryFavourites\nso it retains the returned status alongside items. Render the shared\nfault/recovery state for registry errors or unauthorized responses, and only\nshow “No favourites match” when the registry status is ready and the item list\nis empty.\n\nIn `@src/components/clinical-dashboard/use-saved-registry-favourites.ts`:\n- Around line 27-34: Update useSavedRegistryFavourites to expose and return the\nrefetch callback from useRegistryRecords, then thread it through the consuming\npage into its fault UI so the error state provides the existing retry action.\n\nIn `@src/components/therapy-compass/screens/search-screen.tsx`:\n- Around line 31-33: Update the search screen body rendering to suppress the\nresults and “No therapies match those filters” empty state whenever b.error is\nset, while preserving the existing loading and successful-results behavior. Use\nthe existing b.error state and the body rendering logic near the fault panel,\nmatching the services page’s faulted-state handling.\n\nIn `@tests/search-results-band-adoption.test.ts`:\n- Around line 151-157: Update the forced-colors validation in the search-results\nadoption test to fail when the final `.search-band` selector is absent instead\nof returning early. Determine the final forced-colors block’s closing boundary\nand assert that the selector’s position falls between its opener and closer,\nrather than relying only on opener offset ordering.\n\nIn `@tests/ui-accessibility.spec.ts`:\n- Around line 620-626: Update the rail style extraction in the band evaluation\nto return borderLeftColor instead of borderLeftWidth, then compare rail.color\nagainst that left-border color in the accent-rail assertion. Keep the existing\nwidth and neutral-color checks unchanged.\n\n---\n\nNitpick comments:\nIn `@src/components/clinical-dashboard/medication-prescribing-workspace.tsx`:\n- Around line 471-472: Expose a refetch or retry callback from\nuseMedicationCatalog, then pass it as onRetry or faultAction to the\nSearchResultsHeaderBand rendering the medication catalogue fault panel. Preserve\nthe existing catalog error, loading, and ready status behavior while enabling\nusers to retry failed requests.\n\nIn `@src/components/clinical-dashboard/search-results-header-band.tsx`:\n- Around line 410-435: Update the retry handler on the AsyncButton in the retry\naction block to catch rejected onRetry calls, preserving the existing\nsetRetrying(false) cleanup while surfacing the failure through the component’s\nestablished error or feedback mechanism instead of allowing an unhandled\nrejection.\n```\n\n

\n\n

🪄 Autofix (Beta)

\n\n❌ Autofix failed (check again to retry)\n\nFix all unresolved CodeRabbit comments on this PR:\n\n- [ ] Push a commit to this branch (recommended)\n- [ ] Create a new PR with the fixes\n\n

\n\n---\n\n

ℹ️ Review info

\n\n

⚙️ Run configuration

\n\n**Configuration used**: Path: .coderabbit.yaml\n\n**Review profile**: CHILL\n\n**Plan**: Pro\n\n**Run ID**: `c2e69a80-fdf3-4abf-ae6f-bbef25acf1da`\n\n

\n\n

📥 Commits

\n\nReviewing files that changed from the base of the PR and between 7f91d9946f07362cd49c0d9fe992d38d4ad06e64 and bbad2a9a9fb212d27d827f4c9b03fc41b4ceaf48.\n\n

\n\n

📒 Files selected for processing (24)

\n\n* `docs/search-chrome-behaviour.md`\n* `src/app/globals.css`\n* `src/components/clinical-dashboard/differentials-home.tsx`\n* `src/components/clinical-dashboard/document-search-results.tsx`\n* `src/components/clinical-dashboard/favourites-command-library-page.tsx`\n* `src/components/clinical-dashboard/favourites-hub.tsx`\n* `src/components/clinical-dashboard/medication-prescribing-workspace.tsx`\n* `src/components/clinical-dashboard/search-results-header-band.tsx`\n* `src/components/clinical-dashboard/universal-search-command-surface.tsx`\n* `src/components/clinical-dashboard/use-saved-registry-favourites.ts`\n* `src/components/forms/forms-search-results-page.tsx`\n* `src/components/formulation/formulation-home-page.tsx`\n* `src/components/services/services-navigator-page.tsx`\n* `src/components/therapy-compass/screens/search-screen.tsx`\n* `src/lib/app-modes.ts`\n* `tests/favourites-auth-gate.dom.test.tsx`\n* `tests/favourites-hub-unavailable-controls.dom.test.tsx`\n* `tests/mode-menu-prefetch.dom.test.tsx`\n* `tests/search-results-band-adoption.test.ts`\n* `tests/search-results-header-band.dom.test.tsx`\n* `tests/ui-accessibility.spec.ts`\n* `tests/ui-chrome-scroll.spec.ts`\n* `tests/ui-phone-scroll.spec.ts`\n* `tests/ui-smoke.spec.ts`\n\n

\n\n

\n\n","_🎯 Functional Correctness_ | _🟠 Major_ | _⚡ Quick win_\n\n**Compare the rail color with a color, not a width.**\n\n`rail.color` is a color string while `rail.leftWidth` is a length, so this assertion is effectively always true. Return `borderLeftColor` and compare the two colors; otherwise the regression test can pass while the rail remains neutral.\n\n

Proposed fix

\n\n```diff\n- return { width: style.borderTopWidth, color: style.borderTopColor, leftWidth: style.borderLeftWidth };\n+ return {\n+ width: style.borderTopWidth,\n+ color: style.borderTopColor,\n+ leftColor: style.borderLeftColor,\n+ };\n...\n- expect(rail.color, \"accent rail must use the clinical accent, not --border\").not.toBe(rail.leftWidth);\n+ expect(rail.color, \"accent rail must use the clinical accent, not --border\").not.toBe(rail.leftColor);\n```\n\n

\n\n\n\n

📝 Committable suggestion

\n\n> ‼️ **IMPORTANT**\n> Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.\n\n```suggestion\n const rail = await band.evaluate((node) => {\n const style = getComputedStyle(node);\n return {\n width: style.borderTopWidth,\n color: style.borderTopColor,\n leftColor: style.borderLeftColor,\n };\n });\n expect(rail.width, \"accent rail must be 2px, not the neutral 1px border\").toBe(\"2px\");\n expect(rail.color, \"accent rail must use the clinical accent, not --border\").not.toBe(rail.leftColor);\n expect(rail.color).not.toBe(\"rgb(229, 231, 235)\");\n```\n\n

\n\n\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@tests/ui-accessibility.spec.ts` around lines 620 - 626, Update the rail style\nextraction in the band evaluation to return borderLeftColor instead of\nborderLeftWidth, then compare rail.color against that left-border color in the\naccent-rail assertion. Keep the existing width and neutral-color checks\nunchanged.\n```\n\n

\n\n\n\n\n\n\n\n\n\n✅ Addressed in commit 65ba242","tests/ui-accessibility.spec.ts",626,"_🎯 Functional Correctness_ | _🟡 Minor_ | _⚡ Quick win_\n\n**Make the forced-colors assertion actually enforce the rule.**\n\nLine 152 silently passes if `.search-band` is removed. Also, comparing offsets does not prove the selector is *inside* the final forced-colors block; it could appear after that block closes. Fail when the selector is absent and inspect the matching block boundaries before asserting membership.\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@tests/search-results-band-adoption.test.ts` around lines 151 - 157, Update\nthe forced-colors validation in the search-results adoption test to fail when\nthe final `.search-band` selector is absent instead of returning early.\nDetermine the final forced-colors block’s closing boundary and assert that the\nselector’s position falls between its opener and closer, rather than relying\nonly on opener offset ordering.\n```\n\n

\n\n\n\n\n\n\n\n\n\n✅ Addressed in commits 65ba242 to 8ccd7f4","_🎯 Functional Correctness_ | _🟠 Major_ | _⚡ Quick win_\n\n**Error state still falls through to the \"No therapies match those filters\" empty state.**\n\nWith `b.error` set, `b.loading` is false and `results` is empty, so the body below (Lines 125-146) renders the empty state under the new fault panel — the page simultaneously says the search failed and that nothing matched. Suppress the body while faulted, as the services page does.\n\n

🐛 Proposed fix

\n\n```diff\n- {b.loading ? (\n+ {b.error ? null : b.loading ? (\n \n ) : (\n```\n

\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@src/components/therapy-compass/screens/search-screen.tsx` around lines 31 -\n33, Update the search screen body rendering to suppress the results and “No\ntherapies match those filters” empty state whenever b.error is set, while\npreserving the existing loading and successful-results behavior. Use the\nexisting b.error state and the body rendering logic near the fault panel,\nmatching the services page’s faulted-state handling.\n```\n\n

\n\n\n\n\n\n\n\n\n\n✅ Addressed in commit 65ba242","_🩺 Stability & Availability_ | _🟡 Minor_ | _⚡ Quick win_\n\n

🧩 Analysis chain

\n\n🏁 Script executed:\n\n```shell\n#!/bin/bash\nset -euo pipefail\n\nprintf 'Files:\\n'\ngit ls-files 'src/components/clinical-dashboard/*' | sed -n '1,120p'\n\nprintf '\\nOutline:\\n'\nast-grep outline src/components/clinical-dashboard/use-saved-registry-favourites.ts --view expanded || true\n\nprintf '\\nRelevant file excerpt:\\n'\nwc -l src/components/clinical-dashboard/use-saved-registry-favourites.ts\nsed -n '1,220p' src/components/clinical-dashboard/use-saved-registry-favourites.ts\n```\n\nRepository: BigSimmo/Database\n\nLength of output: 9454\n\n---\n\n🏁 Script executed:\n\n```shell\n#!/bin/bash\nset -euo pipefail\n\nprintf 'Search consumers and retry/refetch usage:\\n'\nrg -n \"useSavedRegistryFavourites|SavedRegistryFavouritesResult|refetch|retry\" src/components/clinical-dashboard src | sed -n '1,220p'\n```\n\nRepository: BigSimmo/Database\n\nLength of output: 23223\n\n---\n\n🏁 Script executed:\n\n```shell\n#!/bin/bash\nset -euo pipefail\n\necho '--- use-saved-registry-favourites.ts ---'\nnl -ba src/components/clinical-dashboard/use-saved-registry-favourites.ts | sed -n '1,220p'\n\necho\necho '--- consumers ---'\nrg -n \"useSavedRegistryFavourites|SavedRegistryFavouritesResult\" src | sed -n '1,200p'\n```\n\nRepository: BigSimmo/Database\n\nLength of output: 236\n\n---\n\n🏁 Script executed:\n\n```shell\n#!/bin/bash\nset -euo pipefail\n\nprintf '--- use-registry-records outline ---\\n'\nast-grep outline src/lib/use-registry-records.ts --view expanded || true\n\nprintf '\\n--- use-registry-records excerpt ---\\n'\npython3 - <<'PY'\nfrom pathlib import Path\np = Path('src/lib/use-registry-records.ts')\ntext = p.read_text().splitlines()\nfor i, line in enumerate(text, 1):\n if i <= 260:\n print(f\"{i:4}: {line}\")\nPY\n\nprintf '\\n--- favorites consumer slices ---\\n'\npython3 - <<'PY'\nfrom pathlib import Path\n\ntargets = [\n 'src/components/clinical-dashboard/favourites-hub.tsx',\n 'src/components/clinical-dashboard/favourites-command-library-page.tsx',\n 'src/components/clinical-dashboard/universal-search-command-surface.tsx',\n]\nfor t in targets:\n print(f\"\\n### {t}\")\n text = Path(t).read_text().splitlines()\n for i, line in enumerate(text, 1):\n if 'useSavedRegistryFavourites' in line or 'favouritesRegistryStatus' in line or 'savedRegistryFavourites' in line:\n start = max(1, i-15)\n end = min(len(text), i+35)\n for j in range(start, end+1):\n print(f\"{j:4}: {text[j-1]}\")\n print()\n break\nPY\n```\n\nRepository: BigSimmo/Database\n\nLength of output: 246\n\n---\n\n🏁 Script executed:\n\n```shell\n#!/bin/bash\nset -euo pipefail\n\npython3 - <<'PY'\nfrom pathlib import Path\n\ntargets = ['src/lib/use-registry-records.ts']\nfor t in targets:\n p = Path(t)\n print(f\"--- {t} ({sum(1 for _ in p.open())} lines) ---\")\n text = p.read_text().splitlines()\n for i, line in enumerate(text, 1):\n print(f\"{i:4}: {line}\")\nPY\n\npython3 - <<'PY'\nfrom pathlib import Path\n\ntargets = [\n 'src/components/clinical-dashboard/favourites-hub.tsx',\n 'src/components/clinical-dashboard/favourites-command-library-page.tsx',\n 'src/components/clinical-dashboard/universal-search-command-surface.tsx',\n]\nfor t in targets:\n p = Path(t)\n print(f\"\\n--- {t} ({sum(1 for _ in p.open())} lines) ---\")\n text = p.read_text().splitlines()\n needles = ['useSavedRegistryFavourites', 'favouritesRegistryStatus', 'savedRegistryFavourites']\n hits = [i for i, line in enumerate(text, 1) if any(n in line for n in needles)]\n if not hits:\n print(\"no hits\")\n continue\n for hit in hits[:5]:\n start = max(1, hit - 18)\n end = min(len(text), hit + 45)\n for i in range(start, end + 1):\n print(f\"{i:4}: {text[i-1]}\")\n print()\nPY\n```\n\nRepository: BigSimmo/Database\n\nLength of output: 50373\n\n---\n\n🏁 Script executed:\n\n```shell\n#!/bin/bash\nset -euo pipefail\n\npython3 - <<'PY'\nfrom pathlib import Path\n\np = Path('src/components/clinical-dashboard/search-results-header-band.tsx')\ntext = p.read_text().splitlines()\nneedles = ['status:', 'onRetry', 'retry', 'refetch', 'favouritesRegistryStatus']\nhits = [i for i, line in enumerate(text, 1) if any(n in line for n in needles)]\nprint(f\"--- {p} ({len(text)} lines) ---\")\nfor hit in hits[:12]:\n start = max(1, hit - 18)\n end = min(len(text), hit + 40)\n for i in range(start, end + 1):\n print(f\"{i:4}: {text[i-1]}\")\n print()\nPY\n\npython3 - <<'PY'\nfrom pathlib import Path\n\np = Path('src/components/clinical-dashboard/favourites-command-library-page.tsx')\ntext = p.read_text().splitlines()\nneedles = ['favouritesRegistryStatus', 'SearchResultsHeaderBand', 'refetch', 'Retry']\nhits = [i for i, line in enumerate(text, 1) if any(n in line for n in needles)]\nprint(f\"--- {p} ({len(text)} lines) ---\")\nfor hit in hits:\n start = max(1, hit - 18)\n end = min(len(text), hit + 45)\n for i in range(start, end + 1):\n print(f\"{i:4}: {text[i-1]}\")\n print()\nPY\n```\n\nRepository: BigSimmo/Database\n\nLength of output: 44524\n\n---\n\n\n\n

\n\n**Thread `refetch` through the favourites hook.** `useRegistryRecords` already exposes a retry callback, but `useSavedRegistryFavourites` drops it, so the favourites error state has no recovery action. Pass that callback through this hook and into the page’s fault UI.\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@src/components/clinical-dashboard/use-saved-registry-favourites.ts` around\nlines 27 - 34, Update useSavedRegistryFavourites to expose and return the\nrefetch callback from useRegistryRecords, then thread it through the consuming\npage into its fault UI so the error state provides the existing retry action.\n```\n\n

\n\n\n\n\n\n\n\n\n\n✅ Addressed in commits 65ba242 to 8ccd7f4","_🎯 Functional Correctness_ | _🟠 Major_ | _⚡ Quick win_\n\n**Preserve registry status in this list view.**\n\nThis component discards `status`, then treats an empty item list as “No favourites match.” A registry error or unauthorized response will therefore look like an empty library. Keep the status and render the shared fault/recovery state, or gate the empty state on `ready`.\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@src/components/clinical-dashboard/favourites-hub.tsx` at line 50, Update the\nfavourites data handling around useSavedRegistryFavourites so it retains the\nreturned status alongside items. Render the shared fault/recovery state for\nregistry errors or unauthorized responses, and only show “No favourites match”\nwhen the registry status is ready and the item list is empty.\n```\n\n

\n\n\n\n\n\n\n\n\n\n✅ Addressed in commit 65ba242","_🎯 Functional Correctness_ | _🟠 Major_ | _⚡ Quick win_\n\n**Gate the normal empty state on a successful registry read.**\n\nPassing `status` to the band does not prevent the later `scopedItems.length === 0` branch from rendering “No matches” during loading, error, or unauthorized states. Render `SearchResultsEmptyState` only when the registry status is `ready`, and use the recovery action for faulted states.\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@src/components/clinical-dashboard/favourites-command-library-page.tsx` around\nlines 1198 - 1201, Update the empty-results rendering in the favourites command\nlibrary page so SearchResultsEmptyState is shown only when\nfavouritesRegistryStatus is ready. For loading, error, or unauthorized statuses,\nsuppress the normal “No matches” branch and use the existing recovery action for\nfaulted states; retain the current empty-state behavior for successfully loaded\nresults.\n```\n\n

\n\n\n\n\n\n\n\n\n\n✅ Addressed in commit 65ba242","_🎯 Functional Correctness_ | _🟠 Major_ | _⚡ Quick win_\n\n**In record mode an `unavailableMessage` is now reported nowhere.**\n\nWhen `showRecordMatches` is true the band's status comes from `recordStatus` and `faultBody` is `undefined`, while the standalone alert is gated on `!showIdentityHeader` — so an auth-expired / API-unavailable / setup-warning message is silently dropped on the services and forms record paths that previously showed it. Keeping the ribbon's count valid (the stated intent) doesn't require discarding the notice.\n\n

🐛 Proposed fix: keep the notice when the ribbon isn't carrying it

\n\n```diff\n- {!showIdentityHeader && unavailableMessage ? (\n+ {/* The ribbon only carries this message on the documents path; in record\n+ mode its fault comes from the registry, so the notice still renders. */}\n+ {unavailableMessage && (showRecordMatches || !showIdentityHeader) ? (\n```\n

\n\n\n\n\n\nAlso applies to: 957-967\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@src/components/clinical-dashboard/document-search-results.tsx` around lines\n894 - 909, Update the record-mode handling in the status/faultBody logic and the\nrelated standalone alert near the corresponding duplicate block so an\nunavailableMessage remains surfaced when the ribbon does not carry it. Preserve\nrecordStatus-based status and valid ribbon counts, but avoid setting faultBody\nand alert visibility in a way that drops auth, API, or setup notices on services\nand forms record paths.\n```\n\n

\n\n\n\n\n\n\n\n\n\n✅ Addressed in commit 65ba242",912,"_🩺 Stability & Availability_ | _🔴 Critical_ | _⚡ Quick win_\n\n**Faulted catalogue with no results now falls into the results branch and dereferences a null `best`.**\n\nThe gate is `!best && !catalogFailed`, so when `catalogFailed` is true and `results` is empty (`best === null`) neither the loading nor the empty branch is taken, and control reaches the results grid, which reads `best.id` (Line 1081) and `best.kind` (Line 1141) unconditionally. That throws on every failed/unauthorized catalogue search — the exact state this PR is meant to render truthfully.\n\nSuppress the whole body while faulted (the band's fault panel already owns the reporting):\n\n

🐛 Proposed fix

\n\n```diff\n ) : /* A failed catalogue search is reported by the band's fault panel, which\n also owns the retry copy; this section is now the empty state only. */\n- !best && !catalogFailed ? (\n+ catalogFailed ? null : !best ? (\n \n\nNote the trailing `{best ? : null}` already guards itself, so only this branch needs the change.\n\n\n\n

📝 Committable suggestion

\n {`No catalogue matches for “${query}”`}\n

\n {hasSourceEvidence\n ? `No imported differential matched this search, but ${reviewedSourceCount.toLocaleString()} indexed source ${\n reviewedSourceCount === 1 ? \"match is\" : \"matches are\"\n } available in the library.`\n : \"Try a symptom, presentation, or diagnosis name — or browse the catalogue directly.\"}\n```\n\n

\n\n\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@src/components/clinical-dashboard/differentials-home.tsx` around lines 992 -\n1007, Update the main catalogue-content branching around the `!best &&\n!catalogFailed` empty-state condition so `catalogFailed` suppresses the entire\nresults body, preventing the results grid from rendering when `best` is null.\nPreserve the existing fault panel as the sole failure UI and leave the guarded\n`DifferentialsMobileCompareBar` behavior unchanged.\n```\n\n

\n\n\n\n\n\n\n\n\n\n✅ Addressed in commit 8be18c3",1021,"_🎯 Functional Correctness_ | _🟡 Minor_ | _⚡ Quick win_\n\n**Unauthorized fault copy promises a retry that is never offered.**\n\n`faultBody` is unconditional (\"Retry the search shortly, or browse the catalogue pages directly.\") while `onRetry` is deliberately omitted for `unauthorized`, so the panel tells the reader to retry and only renders browse links. Branch the body the way the title already does.\n\n

✏️ Proposed fix

\n\n```diff\n- faultBody=\"Retry the search shortly, or browse the catalogue pages directly.\"\n+ faultBody={\n+ catalog.status === \"unauthorized\"\n+ ? \"Sign in again to search, or browse the catalogue pages directly.\"\n+ : \"Retry the search shortly, or browse the catalogue pages directly.\"\n+ }\n```\n

\n\n\n\n

📝 Committable suggestion

\n\n> ‼️ **IMPORTANT**\n> Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.\n\n```suggestion\n faultTitle={\n catalog.status === \"unauthorized\"\n ? \"Sign in again to search the differentials catalogue\"\n : \"The differentials catalogue could not be searched\"\n }\n faultBody={\n catalog.status === \"unauthorized\"\n ? \"Sign in again to search, or browse the catalogue pages directly.\"\n : \"Retry the search shortly, or browse the catalogue pages directly.\"\n }\n // The fault copy promises two recoveries, so both have to exist: rerun the\n // search, and the catalogue links that the removed error section used to\n // carry. Without these the failed view tells the reader to act and gives\n // them nothing to act with.\n onRetry={catalog.status === \"unauthorized\" ? undefined : rerunSearch}\n```\n\n

\n\n\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@src/components/clinical-dashboard/differentials-home.tsx` around lines 919 -\n929, Update the faultBody expression in the catalog fault panel to branch on\ncatalog.status like faultTitle, omitting retry guidance for unauthorized\nresponses while retaining the retry and browse guidance for other failures; keep\nonRetry behavior unchanged.\n```\n\n

\n\n\n\n\n\n\n\n\n\n✅ Addressed in commits 65ba242 to 8ccd7f4","_📐 Maintainability & Code Quality_ | _🟡 Minor_ | _⚡ Quick win_\n\n**Make the typography rule internally consistent.**\n\nThe text says the query is the only heading-weight element, but then gives the figure weight `580` as well. Clarify whether both are emphasized so implementers do not receive conflicting guidance.\n\n

🤖 Prompt for AI Agents

\n\n```\nVerify each finding against current code. Fix only still-valid issues, skip the\nrest with a brief reason, keep changes minimal, and validate.\n\nIn `@docs/search-chrome-behaviour.md` around lines 53 - 60, Update the typography\nguidance in the section describing heading weights to consistently state whether\nthe query and figure both use weight 580, or only the query is emphasized; align\nthe opening claim with the explicit weight values and clarify the intended\nhierarchy for implementers.\n```\n\n

\n\n\n\n\n\n\n\n\n\n✅ Addressed in commits 65ba242 to 8ccd7f4","Confirmed, not fixed — **leaving this open for a human decision.** No disposition marker, deliberately.\n\nYou're right about the mechanism. `savedDifferentials` is derived from account data and never touches either registry, and demo mode merges `prototypeFavouriteItems`, so a failed service/form registry can coexist with a valid nonzero set of displayed items. Folding that into one status marks the whole band failed and hides a count that is true for what `FavouritesTable` is actually rendering.\n\nI'm not fixing it here because this is the **third** instance of the same underlying gap, and the first two I patched per-site:\n\n- documents (P1, `bbad2a9a`) — derived from `recordStatus` instead of the document API\n- favourites (this one)\n- and the general case you named earlier: \"represent partial failure separately\"\n\nThe band models trust as binary — `faulted` or not — and partial failure has no representation in it. Patching a third call site would make the same wrong shape harder to see, not easier. Doing it properly means a fourth state that renders and announces \"some sources loaded, one didn't\", which changes the `role=\"status\"` / `role=\"alert\"` split I settled with you earlier in this PR and needs a product call on what the count then means: all displayed items, or an explicit \"incomplete\".\n\nTwo candidate resolutions, both a follow-up rather than this branch:\n\n1. **Partial status** — add `partial` to the union, assert the count with an explicit incompleteness marker, and keep the fault panel as a secondary notice rather than a replacement.\n2. **Per-source truth** — the band reports only the source it counts, and each unavailable source raises its own notice below, which is closer to how services and forms already behave.\n\nInterim risk is bounded and in the safe direction: favourites currently **over**-reports failure (hides a valid count) rather than under-reporting it, which is the conservative failure mode for a clinical surface. That is why I'd rather leave it visible and open than paper over it.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","\nFixed in `8be18c37`. You were right, and this was worse than the gap it replaced: a Retry that cannot retry promises recovery and silently does nothing.\n\n`useDifferentialSearch` now exposes a `refetch`, mirroring `useRegistryRecords`. A failed request is never written to the cache (only `writeDifferentialCache` on success), so bumping a retry counter in the effect deps is enough to re-issue — no eviction needed, and the query/auth-identity clearing guards are untouched.\n\nTest is the one you specified: first catalogue request fails, `refetch()` is called, and both `kind=diagnosis` and `kind=presentation` are requested again with the band returning to `ready`. One pre-existing assertion moved from `toEqual` to `toMatchObject`, since the return now carries the callback.\n\n`unauthorized` still omits Retry — re-requesting on an expired session just fails again — and the browse links remain in `faultAction` for that case.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_",942,"**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Keep partial favourites results truthful**\n\nWhen a user has locally generated differential favourites—or any demo prototype items—alongside a saved service/form whose registry request fails, `useSavedRegistryFavourites` still returns the unaffected items but folds the registry failure into `favouritesRegistryStatus`. Passing that status here marks the entire band failed and hides the valid nonzero count while `FavouritesTable` below continues rendering those items. Represent the registry failure as partial or derive the band status from all displayed sources, with a test combining a failed registry request and an unaffected favourite.\n\nUseful? React with 👍 / 👎.","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Retry the failed catalogue request**\n\nWhen either `/api/differentials` catalogue request fails transiently, this Retry invokes `rerunSearch`, which only calls the parent `onRunSearch` document-evidence search. `useDifferentialSearch(query)` has no retry counter and the query remains unchanged, so the catalogue request is not repeated and the band stays faulted. Fresh evidence at this head is that the newly added `onRetry` targets the wrong request; expose a catalogue refetch action and add a test where the first catalogue request fails, Retry is clicked, and both catalogue endpoints are called again.\n\nUseful? React with 👍 / 👎.","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `bbad2a9a9f`\n \n\n

ℹ️ About Codex in GitHub

","\nFixed in `bbad2a9a`. The copy I wrote said \"Retry the search shortly, or browse the catalogue pages directly\" while I had removed both affordances along with the old error section — telling the reader to act and giving them nothing to act with.\n\n`onRetry` is now `rerunSearch`, and `faultAction` carries the two catalogue links (`Browse presentations` / `Browse diagnoses`) that the removed section held, rebuilt from the same `differentialRouteWithQuery` helper so the query is preserved.\n\nOne deliberate asymmetry: `onRetry` is omitted when the status is `unauthorized`. Re-running a search that failed on an expired session just fails again; the browse links still render, since those routes remain reachable.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","\nFixed in `bbad2a9a`. You're right, and this was the inverse of the defect the branch exists to fix: instead of claiming a result set that doesn't exist, it claimed failure over one that does — while `SearchRecordResults` rendered those very matches below.\n\nThe ribbon's status now derives from whichever source it is actually counting. When `showRecordMatches` is true it maps `recordStatus`; otherwise it falls back to the document-API path as before. `faultBody` follows the same split, so the document-API message no longer attaches itself to a registry-backed ribbon.\n\nI took the `recordStatus` branch rather than modelling partial failure. Partial state is the more faithful representation, but it needs its own design — a ribbon that says \"records loaded, documents didn't\" is a fourth thing to render and to announce, and inventing it inside this change would be worse than deferring it.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","\nFixed in `bbad2a9a`. This was the most important of the four — it defeated the entire point of the branch. Suppressing the count in the spine achieves nothing while a filter chip renders \"Forms 0\" two lines below it; the reader still sees a zero asserted about a search that never ran.\n\nFixed at the band rather than per call site, since you correctly noted medication and differentials leak the same way: a faulted band now drops `filterControls` and `mobileControls` entirely. A filter over a result set that failed to load has nothing to filter, so there is no state worth preserving there.\n\nAlso took the second half of your suggestion — the DOM test now asserts no digit anywhere in the ribbon region, not just in the status node, so a future count-bearing control cannot reintroduce this quietly.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","\nFixed in `bbad2a9a`. Confirmed by computed style before assuming either way — the rail was genuinely inert:\n\n```\nbefore: borderTopWidth 1px borderTopColor rgb(229, 231, 235) ← --border\nafter: borderTopWidth 2px borderTopColor rgb(11, 111, 134) ← --clinical-accent\n```\n\nThe block is now unlayered, with a comment recording why, since AGENTS.md already states component classes here are deliberately unlayered and this one was not.\n\nThe second half of your point was the more useful one: my own test asserted `toHaveClass(\"search-band\")`, so it passed against a rail that was not rendering. `tests/ui-accessibility.spec.ts` now asserts computed `borderTopWidth`/`borderTopColor`, plus the forced-colors thickening to `3px` — that last one matters because `--clinical-accent` resolves to `LinkText` in HCM and would otherwise be indistinguishable from the other three borders.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_","src/app/globals.css","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `ade17ac880`\n \n\n

ℹ️ About Codex in GitHub

","**_{_{![P1 Badge](https://img.shields.io/badge/P1-orange?style=flat)}} Derive the band failure from the displayed result source**\n\nIn the dashboard’s services/forms path, registry matches have their own `recordStatus` and can be ready while the unrelated dashboard/document request has set global `apiUnavailable`; this condition nevertheless marks the whole ribbon as failed, hides the valid `recordMatchCount`, and announces “Couldn’t search” while `SearchRecordResults` still renders those matches below. When `showRecordMatches` is true, derive failure/loading from `recordStatus` (or represent partial failure separately) instead of letting the document API state invalidate successful registry results.\n\nAGENTS.md reference: [AGENTS.md:L233-L237](https://github.com/BigSimmo/Database/blob/ade17ac880d0317e9b12a786bcc94e860d21e6b7/AGENTS.md#L233-L237)\n\nUseful? React with 👍 / 👎.","**_{_{![P1 Badge](https://img.shields.io/badge/P1-orange?style=flat)}} Hide result-count filters when the registry fails**\n\nWhen the forms registry is `error` or `unauthorized`, this always-mounted `filterControls` prop still renders `ResultTabs` with `displayedMatches.length`, so the faulted ribbon visibly contains “Forms 0” even though its status spine correctly refuses to assert a count. Medication and differential call sites similarly expose zero-valued filter counts during failures; suppress count-bearing controls unless the result source is ready, and add a faulted-page DOM assertion covering the entire ribbon rather than only the status node.\n\nAGENTS.md reference: [AGENTS.md:L233-L237](https://github.com/BigSimmo/Database/blob/ade17ac880d0317e9b12a786bcc94e860d21e6b7/AGENTS.md#L233-L237)\n\nUseful? React with 👍 / 👎.","\n\n\n\n\n\n\n\n@codex resolve actionable Codex review findings for this pull request and current head using the repository instructions. This is the pull request's single automatic repair pass: do not perform a fresh review, create new standalone findings, or request another review. Work only the existing unresolved Codex threads on the current head. Always fix P0 and P1 findings. For P2 and lower findings, fix only clear, scoped, low-risk issues; otherwise disposition them with a concise reason. After fixing or dispositioning a thread, reply in that thread with as the first line, followed by a concise summary; that marker authorizes the workflow to close that exact thread. If human input or new authorization is required, do not use the marker and leave the thread open with the blocker. Finish only after every actionable thread is fixed or dispositioned and closed, or explicitly left open for a human decision. Do not update the branch from main, address unrelated reviews, broaden scope, or create more than one scoped fix commit. Do not use external APIs, paid services, credentials, dependency changes, or broad refactors unless explicitly authorized. Add targeted tests where behavior changes and run the narrowest relevant validation.","\n### 💡 Codex Review\n\nHere are some automated review suggestions for this pull request.\n\n**Reviewed commit:** `389250d3f6`\n \n\n

ℹ️ About Codex in GitHub

","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Preserve recovery actions for a failed differential catalogue**\n\nWhen `useDifferentialSearch` returns `error` or `unauthorized`, this condition now suppresses the entire prior error section, including both catalogue browse links and the only contextual search action. The replacement band receives neither `onRetry` nor `faultAction`, even though its copy tells the user to retry or browse directly, leaving the failed result view with no way to perform either action. Keep the browse actions in the fault state or pass equivalent recovery controls into the band.\n\nUseful? React with 👍 / 👎.","**_{_{![P2 Badge](https://img.shields.io/badge/P2-yellow?style=flat)}} Move result-band overrides out of the component layer**\n\nOn every results-band render, the root also carries Tailwind’s `border` and `border-[color:var(--border)]` utilities. Tailwind’s utilities layer outranks `@layer components` regardless of selector specificity, so those utilities reset this declaration to a neutral 1px top border instead of the intended 2px accent; the same layering also prevents the zero-count color override from beating `text-[color:var(--text-heading)]`. Keep these overrides unlayered or remove the conflicting utilities, and verify computed styles rather than DOM class presence.\n\nAGENTS.md reference: [AGENTS.md:L245-L246](https://github.com/BigSimmo/Database/blob/389250d3f67c767a688a136a5a67572c932f08d0/AGENTS.md#L245-L246)\n\nUseful? React with 👍 / 👎.","\n### CI triage\nCI failed on this PR. Automated classification of the 2 failed job(s):\n\n- `Production UI` — **needs investigation**: inspect the failing step and uploaded diagnostics; rerun only after classifying the cause.\n- `PR required` — **needs investigation**: inspect the failing step and uploaded diagnostics; rerun only after classifying the cause.\n\nCompared with main CI run [#6289](https://github.com/BigSimmo/Database/actions/runs/30344474178) (success).\n\n_Classification is evidence routing, not permission to ignore a failure. Exact quarantined Playwright identities remain governed by the flake ledger._","\n\n\n[![Review Change Stack](https://storage.googleapis.com/coderabbit_public_assets/review-stack-in-coderabbit-ui.svg)](https://app.coderabbit.ai/change-stack/BigSimmo/Database/pull/1316?utm_source=github_walkthrough&utm_medium=github&utm_campaign=change_stack)\n\n\n\n\n> [!NOTE]\n> ## Reviews paused\n> \n> It looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the `reviews.auto_review.auto_pause_after_reviewed_commits` setting.\n> \n> Use the following commands to manage reviews:\n> - `@coderabbitai resume` to resume automatic reviews.\n> - `@coderabbitai review` to trigger a single review.\n> \n> Use the checkboxes below for quick actions:\n> - [ ] ▶️ Resume reviews\n> - [ ] 🔍 Trigger review\n\n\n\n\n

📝 Walkthrough

\n\n## Walkthrough\n\nThe shared `SearchResultsHeaderBand` now uses explicit status and fault props, suppresses untrusted counts, centralizes recovery messaging, and supports forced-colors styling. Search pages and favourites consumers were migrated, with app-mode adoption rules and expanded DOM, accessibility, retry, and browser test coverage.\n\n### Changes\n\n**Search results band**\n\n|Layer / File(s)|Summary|\n|---|---|\n|**Status-aware band contract and presentation**
`src/components/clinical-dashboard/search-results-header-band.tsx`, `src/app/globals.css`, `docs/search-chrome-behaviour.md`, `tests/search-results-header-band.dom.test.tsx`, `tests/ui-accessibility.spec.ts`|The band supports ready, loading, refetching, error, and unauthorized states, suppresses untrusted counts and controls, renders fault recovery, and applies shared typography and forced-colors styling.|\n|**Mode, document, account, and favourites status contracts**
`src/lib/app-modes.ts`, `src/components/clinical-dashboard/document-search-unavailable-status.ts`, `src/components/account-data-provider.tsx`, `src/components/clinical-dashboard/*favourites*`, `src/components/clinical-dashboard/universal-search-command-surface.tsx`|Search modes declare their results surface; document availability, account loading, and saved favourites expose structured statuses, reload behavior, and trusted-count handling.|\n|**Search page fault and loading integration**
`src/components/clinical-dashboard/{differentials-home,document-search-results,medication-prescribing-workspace}.tsx`, `src/components/forms/forms-search-results-page.tsx`, `src/components/services/services-navigator-page.tsx`, `src/components/formulation/formulation-home-page.tsx`, `src/components/therapy-compass/screens/search-screen.tsx`|Search surfaces pass explicit status, fault copy, retry handlers, and unauthorized actions to the shared band while removing duplicate notices and fault content.|\n|**Catalogue, favourites, and Therapy Compass recovery**
`src/components/clinical-dashboard/use-differential-catalog.ts`, `src/components/clinical-dashboard/use-saved-registry-favourites.ts`, `src/components/therapy-compass/{bindings.tsx,data/use-therapy-data.ts}`, `tests/*retry*`, `tests/use-differential-search.dom.test.tsx`|Catalogue, favourites, account, and Therapy Compass retries rerun failed requests and expose settled recovery state.|\n|**Adoption and browser validation**
`tests/search-results-band-adoption.test.ts`, `tests/search-results-header-band.dom.test.tsx`, `tests/ui-*.spec.ts`, `tests/*favourites*.tsx`, `tests/document-search-unavailable-status.test.ts`, `tests/saved-registry-favourites-status.test.ts`|Tests enforce band reachability, status invariants, forced-colors ordering, retry behavior, structured hook mocks, and visible search-control targeting.|\n|**Review and issue documentation**
`docs/branch-review-ledger.md`, `docs/outstanding-issues.md`|Review closeout rows and outstanding issue ledger entries are updated.|\n\n**Estimated code review effort:** 4 (Complex) | ~45 minutes\n\n**Possibly related PRs**\n\n- [BigSimmo/Database#1166](https://github.com/BigSimmo/Database/pull/1166): Overlaps with shared query-ribbon status behavior.\n- [BigSimmo/Database#758](https://github.com/BigSimmo/Database/pull/758): Related to Therapy Compass retry and recovery wiring.\n- [BigSimmo/Database#658](https://github.com/BigSimmo/Database/pull/658): Related to services-page integration with the shared results band.\n\n**Suggested labels:** `codex`\n\n**Suggested reviewers:** `claude`\n\n### Sequence Diagram(s)\n\n```mermaid\nsequenceDiagram\n participant SearchRoute\n participant RegistryOrCatalogue\n participant SearchResultsHeaderBand\n participant User\n SearchRoute->>RegistryOrCatalogue: read loading, error, or unauthorized state\n SearchRoute->>SearchResultsHeaderBand: pass resolved status and fault content\n SearchResultsHeaderBand->>SearchResultsHeaderBand: suppress count and controls when faulted\n SearchResultsHeaderBand->>User: render status region and fault alert\n User->>SearchResultsHeaderBand: activate retry\n SearchResultsHeaderBand->>RegistryOrCatalogue: invoke retry handler\n```\n\n

\n\n\n\n\n

🚥 Pre-merge checks | ✅ 4 | ❌ 1

\n\n### ❌ Failed checks (1 warning)\n\n| Check name | Status | Explanation | Resolution |\n| :----------------: | :--------- | :------------------------------------------------------------------------------------ | :--------------------------------------------------------------------------------- |\n| Docstring Coverage | ⚠️ Warning | Docstring coverage is 21.74% which is insufficient. The required threshold is 80.00%. | Write docstrings for the functions missing them to satisfy the coverage threshold. |\n\n

✅ Passed checks (4 passed)

\n\n

\n\n\n\n\n---\n\n\n\n\n_{Comment `@coderabbitai help` to get the list of available commands.}\n\n","[supa]:sjrfecxgysukkwxsowpy\nThis 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 ↗︎](https://supabase.com/dashboard/project/sjrfecxgysukkwxsowpy/settings/integrations).\n\n

\n\nPreview Branches by Supabase.\nLearn more about [Supabase Branching ↗︎](https://supabase.com/docs/guides/platform/branching).","base_ref","head_ref","pull_request_info",[902],{"_610":49,"_49":903},"## Correction: the therapy retry finding is marked addressed but is not fixed\n\n`use-therapy-data.ts` — *\"Scope retry settlement to its request key and attempt\"* now carries **\"✅ Addressed in commits b92c272 to 5911c2e\"**. I checked the code rather than the claim, and it is unchanged:\n\n- `retryWaitersRef` is still `useRef void>>([])` — no request-key or attempt token\n- `settleRetryWaiters` still drains the whole array unconditionally\n- nothing in that range touches the file\n\n`5911c2e` is a **merge of `main`**, and the two commits it brought in (`18bed6da`, `3a8edb93`) are the CSS/motion audit PRs #1296 and #1297. The range is what CodeRabbit re-scanned, not a fix.\n\nThe finding itself stands and I verified it independently earlier: the old effect's `!active` guard stops a *superseded* effect settling a *newer* reload, but nothing stops a newer effect settling an older waiter, because the ref is hook-scoped. The in-code comment asserts the stronger property — *\"Only the active attempt may settle Retry\"* — which the code does not provide.\n\n**Practical consequence:** it is now invisible to the `@Codex fix` sweep, so unlike the four findings just routed, nobody will pick it up. Severity is genuinely low — Retry's busy state ends early on a route change, and the hook already returns `loading: true` through the `state.requestKey !== requestKey` guard at that moment — but \"low severity\" and \"silently closed\" are different things.\n\n## Separately: `#093` is now confirmed pre-existing on `main`\n\nEarlier I said I had **not** shown the duplicate-root condition predates this branch, only that it was order-dependent. That gap is now closed, and not by me — `main` fixed it independently in `3a8edb93`, and its reasoning matches the CI evidence exactly:\n\n> checking count then immediately calling waitFor races that flicker into a strict-mode violation\n\nDecisive proof it predates this PR: at `631d90d2`, the commit *before* this branch's first commit, `tests/ui-overlap.spec.ts` already carried the comments *\"two `header#search` nodes, which trips Playwright strict mode\"* and *\"a second transient `header#search` can exist briefly.\"*\n\nSo the sole `Production UI` failure this PR ever produced was a known `main` condition, and the hardened helper is now merged in.\n\n**Verification on the merged tip:** `prettier --check .` clean; `npm run typecheck` and `npm run lint` exit 0; `npm run test` — **417 files, 4237 passed, 4 skipped**.\n\n---\n_Generated by [Claude Code](https://claude.ai/code)_",[934,935,936],[],{"_311":933},{"_308":932},{"_308":931},{"_276":277,"_267":278,"_279":280,"_281":62,"_282":18,"_283":18,"_284":285,"_286":287,"_288":238,"_289":130,"_290":291,"_292":132,"_293":294,"_295":-5,"_296":-5,"_297":930,"_299":300,"_301":302,"_303":304,"_305":306},{"_45":126,"_164":165,"_166":167,"_168":18,"_169":170,"_171":172,"_173":919,"_175":-5,"_176":-5,"_177":18,"_178":920,"_180":181,"_182":181,"_183":921,"_185":922,"_187":923,"_189":-5,"_190":191,"_192":193,"_194":195,"_196":924,"_58":-5,"_122":198,"_199":925,"_201":926,"_203":204,"_205":18,"_206":18},{"_136":912},[913,914,915],{"_148":918,"_150":161},{"_148":917,"_150":157},{"_148":916,"_150":151},{"_152":153,"_96":154,"_122":155},{"_152":158,"_96":154,"_122":159},{"_152":162,"_96":154,"_122":163},{"_269":270,"_271":72,"_272":-5,"_273":-5,"_274":-5,"_275":-5},[228],{"_228":927},{"_210":211,"_212":213,"_214":215,"_216":217,"_218":219,"_220":221,"_222":223,"_224":225,"_226":227},{},[],{"_209":62},{"_207":62,"_208":181},{"_45":230,"_152":231,"_232":928,"_234":167,"_235":236,"_237":238,"_239":240,"_241":242,"_243":244,"_245":67,"_246":18,"_247":-5,"_248":-5,"_249":-5,"_250":-5,"_251":-5,"_252":-5,"_253":254,"_255":-5,"_58":929},{"_261":262,"_45":263,"_96":264,"_265":266,"_267":268},{"_257":18,"_258":18,"_259":18,"_260":18},{"_261":262,"_152":262,"_307":-5,"_265":266,"_45":263},[],[],[],{"_96":601,"_100":97,"_359":949},{"_96":343,"_344":940},{"_96":316,"_317":318,"_319":320,"_321":937,"_323":-5},{"_96":321,"_324":325,"_131":132,"_326":-5,"_296":327,"_328":938,"_330":331,"_332":333,"_334":335,"_336":337,"_338":939},{"_340":341,"_342":-5},[],[941,942],{"_348":485,"_350":-5,"_351":946,"_353":-5},{"_348":349,"_350":-5,"_351":943,"_353":-5},[944],{"_355":356,"_357":358,"_359":945,"_361":62},[362,363,181,364,365,366,181,367,181,368,369,370,371,372,181,373,374,375,376,377,378,379,380,181,381,382,383,384,385,386,387,181,388,389,390,391,392,181,393,394,395,396,397,398,399,400,401,402,403,404,405,398,406,181,407,408,409,410,411,181,412,413,414,415,181,416,181,417,418,413,414,419,181,420,421,399,400,401,422,404,405,398,406,181,407,408,409,410,411,181,423,414,424,181,425,415,181,426,427,181,428,429,430,431,419,181,432,421,433,434,435,398,399,400,401,436,437,403,404,405,398,406,181,407,408,409,410,411,181,425,426,438,439,181,440,441,415,181,442,181,428,430,419,181,443,444,445,446,447,398,448,449,450,451,452,453,454,455,456,457,458,459,460,411,181,461,462,463,464,411,465,181,466,467,468,469,470,471,398,472,181,473,474,475,181,476,477,478,419,181,479,480,399,400,481,405,398,406,181,407,408,482,410,411,181,483,484,419,392],[947],{"_355":356,"_357":488,"_359":948,"_361":62},[490,181,491,181,492,181,493,494,495,496,181,497,498,499,500,380,181,501,502,503,181,504,505,506,387,181,507,508,509,510,387,181,511,512,513,514,515,516,517,518,519,387,181,520,521,522,523,524,525,380,181,526,527,528,529,530,531,532,533,534,535,536,537,538,539,540,541,542,543,544,181,545,546,547,548,549,411,550,551,552,553,554,555,556,539,557,558,181,559,560,561,562,563,564,565,566,567,568,569,398,570,571,572,181,573,574,575,576,577,578,579,580,581,582,583,584,585,586,579,580,587,588,589,590,591,592,471,593,594,595,596,181,573,593,597,595,598,181,599,600,387],[950,951,952,953,954,955,956],{"_610":49,"_49":621},{"_610":612,"_348":485,"_355":619,"_357":620},{"_610":49,"_49":618},{"_610":612,"_348":485,"_355":616,"_357":617},{"_610":49,"_49":615},{"_610":612,"_348":349,"_355":613,"_357":614},{"_610":49,"_49":611},[987,988,989],[],{"_311":986},{"_308":985},{"_308":984},{"_276":277,"_267":278,"_279":280,"_281":62,"_282":18,"_283":18,"_284":285,"_286":287,"_288":238,"_289":130,"_290":291,"_292":132,"_293":294,"_295":-5,"_296":-5,"_297":983,"_299":300,"_301":302,"_303":304,"_305":306},{"_45":126,"_164":165,"_166":167,"_168":18,"_169":170,"_171":172,"_173":972,"_175":-5,"_176":-5,"_177":18,"_178":973,"_180":181,"_182":181,"_183":974,"_185":975,"_187":976,"_189":-5,"_190":191,"_192":193,"_194":195,"_196":977,"_58":-5,"_122":198,"_199":978,"_201":979,"_203":204,"_205":18,"_206":18},{"_136":965},[966,967,968],{"_148":971,"_150":161},{"_148":970,"_150":157},{"_148":969,"_150":151},{"_152":153,"_96":154,"_122":155},{"_152":158,"_96":154,"_122":159},{"_152":162,"_96":154,"_122":163},{"_269":270,"_271":72,"_272":-5,"_273":-5,"_274":-5,"_275":-5},[228],{"_228":980},{"_210":211,"_212":213,"_214":215,"_216":217,"_218":219,"_220":221,"_222":223,"_224":225,"_226":227},{},[],{"_209":62},{"_207":62,"_208":181},{"_45":230,"_152":231,"_232":981,"_234":167,"_235":236,"_237":238,"_239":240,"_241":242,"_243":244,"_245":67,"_246":18,"_247":-5,"_248":-5,"_249":-5,"_250":-5,"_251":-5,"_252":-5,"_253":254,"_255":-5,"_58":982},{"_261":262,"_45":263,"_96":264,"_265":266,"_267":268},{"_257":18,"_258":18,"_259":18,"_260":18},{"_261":262,"_152":262,"_307":-5,"_265":266,"_45":263},[],[],[],{"_96":601,"_100":97,"_359":1002},{"_96":343,"_344":993},{"_96":316,"_317":318,"_319":320,"_321":990,"_323":-5},{"_96":321,"_324":325,"_131":132,"_326":-5,"_296":327,"_328":991,"_330":331,"_332":333,"_334":335,"_336":337,"_338":992},{"_340":341,"_342":-5},[],[994,995],{"_348":485,"_350":-5,"_351":999,"_353":-5},{"_348":349,"_350":-5,"_351":996,"_353":-5},[997],{"_355":356,"_357":358,"_359":998,"_361":62},[362,363,181,364,365,366,181,367,181,368,369,370,371,372,181,373,374,375,376,377,378,379,380,181,381,382,383,384,385,386,387,181,388,389,390,391,392,181,393,394,395,396,397,398,399,400,401,402,403,404,405,398,406,181,407,408,409,410,411,181,412,413,414,415,181,416,181,417,418,413,414,419,181,420,421,399,400,401,422,404,405,398,406,181,407,408,409,410,411,181,423,414,424,181,425,415,181,426,427,181,428,429,430,431,419,181,432,421,433,434,435,398,399,400,401,436,437,403,404,405,398,406,181,407,408,409,410,411,181,425,426,438,439,181,440,441,415,181,442,181,428,430,419,181,443,444,445,446,447,398,448,449,450,451,452,453,454,455,456,457,458,459,460,411,181,461,462,463,464,411,465,181,466,467,468,469,470,471,398,472,181,473,474,475,181,476,477,478,419,181,479,480,399,400,481,405,398,406,181,407,408,482,410,411,181,483,484,419,392],[1000],{"_355":356,"_357":488,"_359":1001,"_361":62},[490,181,491,181,492,181,493,494,495,496,181,497,498,499,500,380,181,501,502,503,181,504,505,506,387,181,507,508,509,510,387,181,511,512,513,514,515,516,517,518,519,387,181,520,521,522,523,524,525,380,181,526,527,528,529,530,531,532,533,534,535,536,537,538,539,540,541,542,543,544,181,545,546,547,548,549,411,550,551,552,553,554,555,556,539,557,558,181,559,560,561,562,563,564,565,566,567,568,569,398,570,571,572,181,573,574,575,576,577,578,579,580,581,582,583,584,585,586,579,580,587,588,589,590,591,592,471,593,594,595,596,181,573,593,597,595,598,181,599,600,387],[1003,1004,1005,1006,1007,1008,1009],{"_610":49,"_49":621},{"_610":612,"_348":485,"_355":619,"_357":620},{"_610":49,"_49":618},{"_610":612,"_348":485,"_355":616,"_357":617},{"_610":49,"_49":615},{"_610":612,"_348":349,"_355":613,"_357":614},{"_610":49,"_49":611},[1012,1013,1014,1015,1016,1017,1018,1019,1020,1021,1022,1023,1024,1025,1026,1027,1028,1029,1030,1031,1032,1033,1034,1035,1036,1037,1038,1039,1040,1041,1042,1043,1044,1045,1046,1047,1048,1049,1050,1051,1052,1053,1054,1055,1056,1057,1058,1059,1060,1061,1062,1063,1064,1065,1066,1067,1068,1069,1070,1071,1072,1073,1074,1075,1076,1077,1078,1079,1080,1081,1082,1083,1084,1085,1086,1087,1088,1089,1090,1091,1092,1093,1094,1095,1096,1097,1098,1099,1100,1101,1102,1103,1104,1105,1106,1107,1108,1109,1110,1111,1112,1113,1114,1115,1116,1117,1118,1119,1120,1121,1122,1123],{"_631":632,"_234":167,"_633":277,"_245":-5,"_324":228,"_634":-5,"_635":-5,"_636":637,"_638":639,"_640":-5,"_641":-5,"_642":643,"_644":18,"_645":62,"_646":57,"_647":18,"_648":-5},{"_96":601,"_100":297,"_359":1124},{"_284":285,"_286":287,"_898":291,"_899":132,"_293":-5,"_96":900},{"_96":761,"_286":897,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":896,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":895,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":894,"_348":887,"_764":765,"_766":-5},{"_96":761,"_286":893,"_348":842,"_764":765,"_766":-5},{"_96":761,"_286":892,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":891,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":890,"_348":848,"_764":765,"_766":849},{"_96":761,"_286":889,"_348":763,"_764":765,"_766":-5},{"_96":761,"_286":888,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":886,"_348":887,"_764":765,"_766":-5},{"_96":761,"_286":885,"_348":848,"_764":765,"_766":849},{"_96":761,"_286":884,"_348":763,"_764":765,"_766":-5},{"_96":761,"_286":883,"_348":842,"_764":765,"_766":-5},{"_96":761,"_286":882,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":881,"_348":842,"_764":765,"_766":879},{"_96":761,"_286":880,"_348":840,"_764":765,"_766":841},{"_96":761,"_286":878,"_348":842,"_764":765,"_766":879},{"_96":761,"_286":877,"_348":840,"_764":765,"_766":841},{"_96":761,"_286":876,"_348":843,"_764":765,"_766":-5},{"_96":761,"_286":875,"_348":842,"_764":765,"_766":-5},{"_96":761,"_286":873,"_348":842,"_764":765,"_766":874},{"_96":761,"_286":871,"_348":763,"_764":765,"_766":872},{"_96":761,"_286":870,"_348":840,"_764":765,"_766":841},{"_96":761,"_286":869,"_348":794,"_764":765,"_766":-5},{"_96":761,"_286":868,"_348":822,"_764":765,"_766":844},{"_96":761,"_286":867,"_348":797,"_764":765,"_766":798},{"_96":761,"_286":866,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":863,"_348":864,"_764":765,"_766":865},{"_96":761,"_286":862,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":861,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":860,"_348":848,"_764":765,"_766":849},{"_96":761,"_286":859,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":858,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":807,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":857,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":856,"_348":822,"_764":765,"_766":838},{"_96":761,"_286":855,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":854,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":853,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":852,"_348":842,"_764":765,"_766":-5},{"_96":761,"_286":851,"_348":843,"_764":765,"_766":-5},{"_96":761,"_286":850,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":847,"_348":848,"_764":765,"_766":849},{"_96":761,"_286":846,"_348":822,"_764":765,"_766":844},{"_96":761,"_286":845,"_348":840,"_764":765,"_766":841},{"_96":761,"_286":791,"_348":822,"_764":765,"_766":844},{"_96":761,"_286":791,"_348":843,"_764":765,"_766":-5},{"_96":761,"_286":791,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":791,"_348":842,"_764":765,"_766":-5},{"_96":761,"_286":839,"_348":840,"_764":765,"_766":841},{"_96":761,"_286":837,"_348":822,"_764":765,"_766":838},{"_96":761,"_286":836,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":835,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":834,"_348":822,"_764":765,"_766":-5},{"_96":761,"_286":833,"_348":820,"_764":765,"_766":-5},{"_96":761,"_286":832,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":831,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":830,"_348":763,"_764":765,"_766":824},{"_96":761,"_286":829,"_348":771,"_764":765,"_766":826},{"_96":761,"_286":828,"_348":771,"_764":765,"_766":826},{"_96":761,"_286":827,"_348":763,"_764":765,"_766":824},{"_96":761,"_286":825,"_348":771,"_764":765,"_766":826},{"_96":761,"_286":823,"_348":763,"_764":765,"_766":824},{"_96":761,"_286":821,"_348":822,"_764":765,"_766":-5},{"_96":761,"_286":819,"_348":820,"_764":765,"_766":-5},{"_96":761,"_286":818,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":817,"_348":792,"_764":765,"_766":-5},{"_96":761,"_286":816,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":815,"_348":794,"_764":765,"_766":795},{"_96":761,"_286":814,"_348":797,"_764":765,"_766":798},{"_96":761,"_286":813,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":812,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":811,"_348":794,"_764":765,"_766":795},{"_96":761,"_286":810,"_348":797,"_764":765,"_766":798},{"_96":761,"_286":809,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":808,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":807,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":806,"_348":794,"_764":765,"_766":795},{"_96":761,"_286":805,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":804,"_348":797,"_764":765,"_766":798},{"_96":761,"_286":803,"_348":788,"_764":765,"_766":-5},{"_96":761,"_286":802,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":801,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":800,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":799,"_348":792,"_764":765,"_766":-5},{"_96":761,"_286":796,"_348":797,"_764":765,"_766":798},{"_96":761,"_286":793,"_348":794,"_764":765,"_766":795},{"_96":761,"_286":791,"_348":792,"_764":765,"_766":-5},{"_96":761,"_286":790,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":789,"_348":771,"_764":765,"_766":-5},{"_96":761,"_286":787,"_348":788,"_764":765,"_766":-5},{"_96":761,"_286":786,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":784,"_348":485,"_764":765,"_766":785},{"_96":761,"_286":783,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":782,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":781,"_348":771,"_764":765,"_766":772},{"_96":761,"_286":780,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":779,"_348":763,"_764":765,"_766":767},{"_96":761,"_286":778,"_348":763,"_764":765,"_766":767},{"_96":761,"_286":777,"_348":773,"_764":765,"_766":774},{"_96":761,"_286":776,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":775,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":770,"_348":773,"_764":765,"_766":774},{"_96":761,"_286":770,"_348":763,"_764":765,"_766":767},{"_96":761,"_286":770,"_348":763,"_764":765,"_766":767},{"_96":761,"_286":770,"_348":771,"_764":765,"_766":772},{"_96":761,"_286":769,"_348":-5,"_764":-5,"_766":-5},{"_96":761,"_286":768,"_348":763,"_764":765,"_766":767},{"_96":761,"_286":762,"_348":763,"_764":765,"_766":767},[1125],{"_610":49,"_49":903},"task_e_6a689eaff6608322aac731fab88bca3f",1785241263.962635,"GitHub Mention: Search results band: truthful failure state, approved design, and adoption gate","has_generated_title","has_unread_turn","denormalized_metadata",{"_1157":35,"_1158":1159,"_1160":-5,"_1161":-5,"_1151":1162,"_1163":-5,"_1164":-5,"_1165":-5,"_1166":-5},"task_status_display",{"_1145":1146,"_1147":167,"_133":130,"_1148":-5,"_1149":316},"creator_workspace","external_pull_requests",[1138],{"_45":1139,"_1140":87,"_1141":1142,"_1143":132},"github-1243497866-1316","assistant_turn_id","pull_request",{"_276":277,"_267":278,"_279":280,"_281":62,"_282":18,"_283":18,"_284":285,"_286":287,"_288":238,"_289":130,"_290":291,"_292":132,"_293":294,"_295":-5,"_296":-5,"_297":1144,"_299":300,"_301":302,"_303":304,"_305":306},"codex_updated_sha",{"_261":262,"_152":262,"_307":-5,"_265":266,"_45":263},"latest_turn_status_display",{"_1150":87,"_115":116,"_1151":1152,"_98":1153,"_1154":-5,"_1155":-5,"_1156":316,"_117":-5},"environment_label","fix_task_id","initial_intent","turn_id","diff_stats",{"_330":331,"_332":333,"_334":335},[],"image_asset_pointer_citations","issue_counts","intent","num_follow_ups","repo_name_map",{},"last_user_turn_created_at","last_turn_status",{"_330":-5,"_332":-5,"_334":-5},"last_pushed_turn_id","start_branch_name","pushed_branch_name","cached_pull_request_data","dd",{"_1181":1182,"_1183":1184},"disablePrefetch","shouldPrefetchAccount","shouldPrefetchUser","shouldPrefetchSystemHints","manageRootRouteStylesheet","mobileWebHandoffEnabled","promoteCss","disableSSR","stripModulepreloadImports","paidCampaignAppRedirect","statsigGateEvaluationsPromise",["P",1180],"traceId","7413774938906322386","traceTime",1787761946976]