Standardize live search Playwright setup lane

docs-archived/implplan/SPRINT_20260308_009_FE_standard_live_search_playwright_setup.md

@@ -0,0 +1,82 @@
+# Sprint 20260308-009 - FE Standard Live Search Playwright Setup
+
+## Topic & Scope
+- Turn the ad hoc live-search ingestion/browser flow into a standard one-command Playwright acceptance lane for search.
+- Make the lane self-preparing where feasible: start the dedicated AdvisoryAI knowledge test database, start the local AdvisoryAI service, prepare sources, rebuild indexes, and gate the browser suite on a grounded smoke query.
+- Keep the implementation centered on `src/Web/StellaOps.Web` while allowing the minimum supporting doc updates in AdvisoryAI test/setup guidance.
+- Working directory: `src/Web/StellaOps.Web`.
+- Expected evidence: setup script/config, live setup artifact, Playwright live browser results, and updated operator docs.
+
+## Dependencies & Concurrency
+- Depends on archived search rollout/correction sprints plus archived `SPRINT_20260308_008_FE_live_search_ingestion_browser_validation.md`.
+- Safe parallelism: do not edit unrelated Router, platform-setup, or shell-cutover files while executing this lane.
+- Cross-module allowance: documentation updates are allowed in `docs/modules/advisory-ai/**` and `src/AdvisoryAI/__Tests/INFRASTRUCTURE.md` because they define the live-search setup contract consumed by this FE lane.
+
+## Documentation Prerequisites
+- `docs/qa/feature-checks/FLOW.md`
+- `docs/code-of-conduct/TESTING_PRACTICES.md`
+- `src/Web/StellaOps.Web/AGENTS.md`
+- `src/AdvisoryAI/__Tests/INFRASTRUCTURE.md`
+- `docs/modules/advisory-ai/knowledge-search.md`
+
+## Delivery Tracker
+
+### FE-LIVESETUP-001 - Standardize the Playwright live-search setup lane
+Status: DONE
+Dependency: none
+Owners: QA, Test Automation, Developer (FE)
+Task description:
+- Add a standard Playwright entrypoint for live search acceptance that boots the dedicated knowledge-test database, starts the source-run AdvisoryAI service, and runs a setup project before browser tests execute.
+- The setup project must prepare the ingestion sources, rebuild both search indexes, and fail fast when the live service is unhealthy or the grounded smoke query does not work.
+
+Completion criteria:
+- [x] A documented one-command live-search Playwright lane exists under `src/Web/StellaOps.Web`.
+- [x] The lane does not assume `stella` is already installed; it compiles or publishes the CLI when needed.
+- [x] Live setup artifacts record rebuild/query evidence for operator diagnosis.
+
+### FE-LIVESETUP-002 - Expand live browser verification on the standardized lane
+Status: DONE
+Dependency: FE-LIVESETUP-001
+Owners: QA, Test Automation
+Task description:
+- Run the live browser suite through the new standard lane and add at least one operator-relevant assertion beyond bare suggestion execution.
+- The added assertion should cover a real UX contract from the shipped search experience, not just service availability.
+
+Completion criteria:
+- [x] The standardized live lane executes the existing live search browser suite successfully.
+- [x] The live suite verifies at least one user-facing persistence/continuity behavior on real ingested data.
+- [x] Test evidence is captured with exact command and result counts.
+
+### FE-LIVESETUP-003 - Update docs and close the lane
+Status: DONE
+Dependency: FE-LIVESETUP-002
+Owners: Documentation, Project Manager
+Task description:
+- Update Web and AdvisoryAI setup docs to point at the new standard live-search acceptance lane.
+- Archive this sprint when the setup lane, verification, and docs are complete.
+
+Completion criteria:
+- [x] Web README documents the standard live-search Playwright command.
+- [x] AdvisoryAI infrastructure/search docs document the setup contract and prerequisites.
+- [x] Sprint execution log captures implementation and validation evidence before archival.
+
+## Execution Log
+| Date (UTC) | Update | Owner |
+| --- | --- | --- |
+| 2026-03-08 | Sprint created to replace manual live-search ingestion/browser steps with a standard Playwright setup lane. | Developer |
+| 2026-03-08 | Added the standard lane under `src/Web/StellaOps.Web` with `scripts/run-live-search-e2e.mjs`, `playwright.live-search.config.ts`, and a `live-search-setup` Playwright project that compiles the CLI if needed, runs `sources prepare`, rebuilds both indexes, and writes `output/playwright/live-search-setup.json`. | Developer |
+| 2026-03-08 | Extended the live browser suite so the standardized lane also verifies user-facing continuity: a grounded live search persists in Recent and can be removed through the clear-history icon. | QA |
+| 2026-03-08 | Validation: `npm run test:e2e:search:live` passed with `10 passed`, `3 skipped` in ~2.6 minutes; skips were the explicit corpus-unready branches for routes that are now ready enough to bypass those tests. | QA |
+
+## Decisions & Risks
+- Decision: standardize the search live lane as a dedicated acceptance path instead of forcing ingestion into every default FE test run.
+- Decision: use a wrapper script plus a dedicated Playwright config so the live lane can bring up the dedicated knowledge-test database before Playwright starts the source-run AdvisoryAI service and browser projects.
+- Decision: the setup project writes a JSON artifact under `output/playwright/live-search-setup.json` so failures can be diagnosed from corpus preparation and rebuild evidence, not only browser traces.
+- Risk: the setup lane depends on Docker, dotnet, Node/npm, and the dedicated AdvisoryAI knowledge test database compose file.
+- Mitigation: fail early with explicit setup diagnostics and keep the lane isolated from the fast mocked/offline FE loop.
+- Risk: CLI `--json` output can be prefixed by infrastructure logs on some hosts.
+- Mitigation: the setup parser now tolerates log-prefixed JSON and still fails hard when no valid payload is present.
+
+## Next Checkpoints
+- 2026-03-08: setup script/config implemented and locally executable.
+- 2026-03-08: live browser suite rerun through the standard lane with recorded evidence.

docs/modules/advisory-ai/knowledge-search.md

@@ -413,11 +413,27 @@ If the CLI is not built yet, the equivalent HTTP endpoints are:
 - `POST /v1/advisory-ai/index/rebuild` for the docs/OpenAPI/Doctor corpus
 - `POST /v1/search/index/rebuild` for unified overlay domains
 
+Standard browser acceptance lane for live search:
+
+```bash
+cd src/Web/StellaOps.Web
+npm run test:e2e:search:live
+```
+
+This command is now the standard local search-browser setup because it:
+- starts or reuses the dedicated AdvisoryAI knowledge-test Postgres
+- starts or reuses the source-run AdvisoryAI WebService on `http://127.0.0.1:10451`
+- publishes the local CLI when `.artifacts/stella-cli/` is missing
+- runs `advisoryai sources prepare --json`
+- rebuilds the knowledge and unified search indexes in the required order
+- proves a grounded smoke query before the browser tests run
+
 Current live verification coverage:
 - Rebuild order exercised against a running local service: `POST /v1/advisory-ai/index/rebuild` then `POST /v1/search/index/rebuild`, both with explicit `X-StellaOps-Scopes`, `X-StellaOps-Tenant`, and `X-StellaOps-Actor` headers
 - Verified live query: `database connectivity`
 - Verified live outcome: response includes `contextAnswer.status = grounded`, citations, and entity cards over ingested data
-- Verified live suggestion lane: `src/Web/StellaOps.Web/tests/e2e/unified-search-contextual-suggestions.live.e2e.spec.ts` now preflights corpus readiness, validates suggestion viability, executes every surfaced Doctor suggestion, asserts grounded answer states for surfaced live suggestions, verifies follow-up chips after result open, and verifies Ask-AdvisoryAI inherits the live query context
+- Verified live suggestion lane: `src/Web/StellaOps.Web/tests/e2e/unified-search-contextual-suggestions.live.e2e.spec.ts` now runs through the standard Playwright setup lane, preflights corpus readiness, validates suggestion viability, executes every surfaced Doctor suggestion, asserts grounded answer states for surfaced live suggestions, verifies follow-up chips after result open, verifies grounded-search history/clear behavior, and verifies Ask-AdvisoryAI inherits the live query context
+- Standard Playwright setup/config entrypoints: `src/Web/StellaOps.Web/scripts/run-live-search-e2e.mjs`, `src/Web/StellaOps.Web/playwright.live-search.config.ts`, and `src/Web/StellaOps.Web/tests/e2e/live-search.setup.ts`
 - Verified combined browser gate on 2026-03-08: `24/24` executed tests passed with `3` explicit route-unready skips across deterministic UX, telemetry-off search flows, self-serve answer panel, and the supported-route live suggestion lane against the ingested local corpus
 - Verified local corpus baseline on 2026-03-07 after `advisoryai sources prepare`: `documentCount = 470`, `chunkCount = 9050`, `apiOperationCount = 2190`, `doctorProjectionCount = 8`
 - Other routes still rely on deterministic mock-backed Playwright coverage until their ingestion parity is explicitly verified

src/AdvisoryAI/__Tests/INFRASTRUCTURE.md

@@ -334,6 +334,29 @@ E2E config: `src/Web/StellaOps.Web/playwright.e2e.config.ts`
 - Timeout: 60s per test
 - Workers: 1 (sequential)
 
+### Standard live-search browser lane
+
+For search verification against a freshly ingested local AdvisoryAI corpus, use the dedicated Web acceptance lane instead of manually sequencing database startup, CLI prepare, and rebuild calls:
+
+```bash
+cd src/Web/StellaOps.Web
+npm run test:e2e:search:live
+```
+
+This lane:
+- starts or reuses `devops/compose/docker-compose.advisoryai-knowledge-test.yml`
+- starts or reuses the source-run AdvisoryAI WebService on `http://127.0.0.1:10451`
+- publishes the local CLI to `.artifacts/stella-cli/` when needed
+- runs `advisoryai sources prepare --json`
+- rebuilds `POST /v1/advisory-ai/index/rebuild` then `POST /v1/search/index/rebuild`
+- proves the grounded smoke query `database connectivity`
+- runs the live browser suite through `src/Web/StellaOps.Web/playwright.live-search.config.ts`
+
+Useful overrides:
+- `PLAYWRIGHT_LIVE_SEARCH_SKIP_DOCKER=1`
+- `LIVE_ADVISORYAI_SEARCH_BASE_URL=http://127.0.0.1:10451`
+- `AdvisoryAI__KnowledgeSearch__ConnectionString=...`
+
 ---
 
 ## Configuration reference

src/Web/StellaOps.Web/README.md

@@ -71,6 +71,33 @@ npm run test:e2e
 
 The Playwright config auto-starts `npm run serve:test` and intercepts Authority redirects, so no live IdP is required. For CI/offline nodes, pre-install the required browsers via `npx playwright install --with-deps` and cache the results alongside your npm cache.
 
+### Standard live-search acceptance lane
+
+Use this lane when you need browser verification against a freshly ingested local AdvisoryAI corpus instead of mocked search responses:
+
+```bash
+npm run test:e2e:search:live
+```
+
+What it does:
+- starts or reuses the dedicated AdvisoryAI knowledge-test Postgres via `devops/compose/docker-compose.advisoryai-knowledge-test.yml`
+- starts or reuses the local AdvisoryAI WebService on `http://127.0.0.1:10451`
+- compiles the local CLI if `.artifacts/stella-cli/` is missing
+- runs `advisoryai sources prepare --json`
+- rebuilds `POST /v1/advisory-ai/index/rebuild` and `POST /v1/search/index/rebuild`
+- proves a grounded smoke query before the browser suite runs
+- executes the live search Playwright spec through `playwright.live-search.config.ts`
+
+Prerequisites:
+- `docker compose`
+- `dotnet`
+- Playwright browsers installed locally
+
+Optional environment overrides:
+- `PLAYWRIGHT_LIVE_SEARCH_SKIP_DOCKER=1` if the dedicated AdvisoryAI test database is already running
+- `LIVE_ADVISORYAI_SEARCH_BASE_URL` to target a non-default local AdvisoryAI host
+- `AdvisoryAI__KnowledgeSearch__ConnectionString` to override the dedicated test database connection string
+
 ## Running end-to-end tests
 
 Run `ng e2e` to execute the end-to-end tests via a platform of your choice. To use this command, you need to first add a package that implements end-to-end testing capabilities.

src/Web/StellaOps.Web/package.json

@@ -13,6 +13,7 @@
     "test:watch": "ng test",
     "test:ci": "npm run test",
     "test:e2e": "playwright test",
+    "test:e2e:search:live": "node ./scripts/run-live-search-e2e.mjs",
     "serve:test": "ng serve --configuration development --port 4400 --host 127.0.0.1 --ssl",
     "verify:chromium": "node ./scripts/verify-chromium.js",
     "ci:install": "npm ci --prefer-offline --no-audit --no-fund",

src/Web/StellaOps.Web/playwright.live-search.config.ts

@@ -0,0 +1,81 @@
+import path from 'node:path';
+
+import { defineConfig, devices } from '@playwright/test';
+
+// eslint-disable-next-line @typescript-eslint/no-var-requires
+const { resolveChromeBinary } = require('./scripts/chrome-path');
+
+const webRoot = __dirname;
+const repoRoot = path.resolve(__dirname, '..', '..', '..');
+const appPort = process.env.PLAYWRIGHT_PORT
+  ? Number.parseInt(process.env.PLAYWRIGHT_PORT, 10)
+  : 4400;
+const appBaseUrl = process.env.PLAYWRIGHT_BASE_URL ?? `https://127.0.0.1:${appPort}`;
+const liveSearchBaseUrl = process.env.LIVE_ADVISORYAI_SEARCH_BASE_URL?.trim()
+  || 'http://127.0.0.1:10451';
+const advisoryConnectionString = process.env.AdvisoryAI__KnowledgeSearch__ConnectionString?.trim()
+  || 'Host=localhost;Port=55432;Database=advisoryai_knowledge_test;Username=stellaops_knowledge;Password=stellaops_knowledge';
+
+process.env.LIVE_ADVISORYAI_SEARCH_BASE_URL = liveSearchBaseUrl;
+process.env.LIVE_ADVISORYAI_TENANT = process.env.LIVE_ADVISORYAI_TENANT?.trim() || 'test-tenant';
+process.env.LIVE_ADVISORYAI_SCOPES =
+  process.env.LIVE_ADVISORYAI_SCOPES?.trim()
+  || 'advisory-ai:view advisory-ai:operate advisory-ai:admin';
+
+const chromiumExecutable = resolveChromeBinary(__dirname) as string | null;
+
+export default defineConfig({
+  testDir: 'tests/e2e',
+  timeout: 120_000,
+  expect: { timeout: 10_000 },
+  fullyParallel: false,
+  retries: process.env.CI ? 1 : 0,
+  workers: 1,
+  reporter: [['list'], ['html', { open: 'never' }]],
+  use: {
+    baseURL: appBaseUrl,
+    ignoreHTTPSErrors: true,
+    trace: 'retain-on-failure',
+    screenshot: 'only-on-failure',
+    video: 'retain-on-failure',
+    ...(chromiumExecutable ? { launchOptions: { executablePath: chromiumExecutable } } : {}),
+  },
+  webServer: [
+    {
+      command: 'npm run serve:test',
+      cwd: webRoot,
+      reuseExistingServer: !process.env.CI,
+      url: appBaseUrl,
+      ignoreHTTPSErrors: true,
+      stdout: 'ignore',
+      stderr: 'ignore',
+      timeout: 120_000,
+    },
+    {
+      command: 'dotnet run --project src/AdvisoryAI/StellaOps.AdvisoryAI.WebService/StellaOps.AdvisoryAI.WebService.csproj --no-launch-profile',
+      cwd: repoRoot,
+      reuseExistingServer: !process.env.CI,
+      url: `${liveSearchBaseUrl}/health`,
+      stdout: 'ignore',
+      stderr: 'ignore',
+      timeout: 180_000,
+      env: {
+        ...process.env,
+        AdvisoryAI__KnowledgeSearch__ConnectionString: advisoryConnectionString,
+        ASPNETCORE_URLS: liveSearchBaseUrl,
+      },
+    },
+  ],
+  projects: [
+    {
+      name: 'live-search-setup',
+      testMatch: /live-search\.setup\.ts/,
+    },
+    {
+      name: 'live-search-chromium',
+      testMatch: /unified-search-contextual-suggestions\.live\.e2e\.spec\.ts/,
+      use: { ...devices['Desktop Chrome'] },
+      dependencies: ['live-search-setup'],
+    },
+  ],
+});

src/Web/StellaOps.Web/scripts/run-live-search-e2e.mjs

@@ -0,0 +1,108 @@
+#!/usr/bin/env node
+
+import net from 'node:net';
+import path from 'node:path';
+import { fileURLToPath } from 'node:url';
+import { spawnSync } from 'node:child_process';
+
+const __filename = fileURLToPath(import.meta.url);
+const __dirname = path.dirname(__filename);
+const webRoot = path.resolve(__dirname, '..');
+const repoRoot = path.resolve(webRoot, '..', '..', '..');
+const composeFile = path.join(
+  repoRoot,
+  'devops',
+  'compose',
+  'docker-compose.advisoryai-knowledge-test.yml',
+);
+
+async function main() {
+  const forwardedArgs = process.argv.slice(2);
+  const env = {
+    ...process.env,
+    LIVE_ADVISORYAI_SEARCH_BASE_URL:
+      process.env.LIVE_ADVISORYAI_SEARCH_BASE_URL?.trim() || 'http://127.0.0.1:10451',
+    PLAYWRIGHT_LIVE_SEARCH_SKIP_REBUILD: '1',
+  };
+
+  if (process.env.PLAYWRIGHT_LIVE_SEARCH_SKIP_DOCKER !== '1') {
+    runCommand('docker', ['compose', '-f', composeFile, 'up', '-d'], repoRoot);
+    await waitForTcpPort('127.0.0.1', 55432, 60_000);
+  }
+
+  const playwrightArgs = ['playwright', 'test', '--config', 'playwright.live-search.config.ts', ...forwardedArgs];
+  const result = process.platform === 'win32'
+    ? spawnSync('cmd.exe', ['/d', '/s', '/c', 'npx', ...playwrightArgs], {
+        cwd: webRoot,
+        env,
+        stdio: 'inherit',
+      })
+    : spawnSync('npx', playwrightArgs, {
+        cwd: webRoot,
+        env,
+        stdio: 'inherit',
+      });
+
+  if (result.error) {
+    throw result.error;
+  }
+
+  process.exit(result.status ?? 1);
+}
+
+function runCommand(command, args, cwd) {
+  const result = spawnSync(command, args, {
+    cwd,
+    stdio: 'inherit',
+    env: process.env,
+  });
+
+  if (result.error) {
+    throw result.error;
+  }
+
+  if ((result.status ?? 1) !== 0) {
+    throw new Error(`${command} ${args.join(' ')} exited with status ${result.status ?? 1}.`);
+  }
+}
+
+function waitForTcpPort(host, port, timeoutMs) {
+  const startedAt = Date.now();
+
+  return new Promise((resolve, reject) => {
+    const attempt = () => {
+      const socket = new net.Socket();
+
+      socket.setTimeout(2_000);
+      socket.once('connect', () => {
+        socket.destroy();
+        resolve();
+      });
+      socket.once('timeout', () => {
+        socket.destroy();
+        retry(new Error(`Timed out while connecting to ${host}:${port}.`));
+      });
+      socket.once('error', (error) => {
+        socket.destroy();
+        retry(error);
+      });
+      socket.connect(port, host);
+    };
+
+    const retry = (lastError) => {
+      if (Date.now() - startedAt >= timeoutMs) {
+        reject(new Error(`Port ${host}:${port} did not become ready within ${timeoutMs}ms: ${lastError.message}`));
+        return;
+      }
+
+      setTimeout(attempt, 1_000);
+    };
+
+    attempt();
+  });
+}
+
+main().catch((error) => {
+  console.error(`[live-search-e2e] ${error instanceof Error ? error.message : String(error)}`);
+  process.exit(1);
+});

src/Web/StellaOps.Web/tests/e2e/live-search.setup.ts

@@ -0,0 +1,40 @@
+import { expect, test } from '@playwright/test';
+
+import {
+  ensureLiveServiceHealthy,
+  rebuildLiveIndexes,
+  resolveLiveSearchRuntimeConfig,
+  runGroundedSmokeQuery,
+  runSourcesPrepare,
+  writeLiveSearchSetupArtifact,
+} from './support/live-search-support';
+
+test.describe('Live search setup', () => {
+  test('prepares the ingested corpus and proves grounded retrieval before browser execution', async ({}, testInfo) => {
+    testInfo.setTimeout(300_000);
+
+    const config = resolveLiveSearchRuntimeConfig();
+    await ensureLiveServiceHealthy(config, 180_000);
+
+    const sourcesPrepare = runSourcesPrepare(config);
+    const { knowledge, unified } = await rebuildLiveIndexes(config);
+    const smokeQuery = await runGroundedSmokeQuery(config);
+    const artifactPath = writeLiveSearchSetupArtifact(config, {
+      preparedAtUtc: new Date().toISOString(),
+      baseUrl: config.baseUrl,
+      tenant: config.tenant,
+      sourcesPrepare,
+      knowledgeRebuild: knowledge,
+      unifiedRebuild: unified,
+      smokeQuery,
+    });
+
+    expect((smokeQuery['contextAnswer'] as Record<string, unknown> | undefined)?.['status']).toBe('grounded');
+    expect(Array.isArray(smokeQuery['cards']) ? smokeQuery['cards'].length : 0).toBeGreaterThan(0);
+
+    await testInfo.attach('live-search-setup.json', {
+      path: artifactPath,
+      contentType: 'application/json',
+    });
+  });
+});

src/Web/StellaOps.Web/tests/e2e/support/live-search-support.ts

@@ -0,0 +1,267 @@
+import { execFileSync } from 'node:child_process';
+import { existsSync, mkdirSync, writeFileSync } from 'node:fs';
+import path from 'node:path';
+
+export type LiveSearchRuntimeConfig = {
+  baseUrl: string;
+  tenant: string;
+  scopes: string;
+  actor: string;
+  advisoryConnectionString: string;
+  repoRoot: string;
+  webRoot: string;
+};
+
+export type LiveSearchSetupReport = {
+  preparedAtUtc: string;
+  baseUrl: string;
+  tenant: string;
+  sourcesPrepare: Record<string, unknown>;
+  knowledgeRebuild: Record<string, unknown>;
+  unifiedRebuild: Record<string, unknown>;
+  smokeQuery: Record<string, unknown>;
+};
+
+const DEFAULT_LIVE_SEARCH_BASE_URL = 'http://127.0.0.1:10451';
+const DEFAULT_LIVE_TENANT = 'test-tenant';
+const DEFAULT_LIVE_SCOPES = 'advisory-ai:view advisory-ai:operate advisory-ai:admin';
+const DEFAULT_LIVE_ACTOR = 'playwright-live-setup';
+const DEFAULT_ADVISORY_CONNSTRING =
+  'Host=localhost;Port=55432;Database=advisoryai_knowledge_test;Username=stellaops_knowledge;Password=stellaops_knowledge';
+
+export function resolveLiveSearchRuntimeConfig(): LiveSearchRuntimeConfig {
+  const repoRoot = findRepoRoot(process.cwd());
+  const webRoot = path.join(repoRoot, 'src', 'Web', 'StellaOps.Web');
+
+  return {
+    baseUrl: process.env['LIVE_ADVISORYAI_SEARCH_BASE_URL']?.trim() || DEFAULT_LIVE_SEARCH_BASE_URL,
+    tenant: process.env['LIVE_ADVISORYAI_TENANT']?.trim() || DEFAULT_LIVE_TENANT,
+    scopes: process.env['LIVE_ADVISORYAI_SCOPES']?.trim() || DEFAULT_LIVE_SCOPES,
+    actor: process.env['LIVE_ADVISORYAI_ACTOR']?.trim() || DEFAULT_LIVE_ACTOR,
+    advisoryConnectionString:
+      process.env['AdvisoryAI__KnowledgeSearch__ConnectionString']?.trim() || DEFAULT_ADVISORY_CONNSTRING,
+    repoRoot,
+    webRoot,
+  };
+}
+
+export async function ensureLiveServiceHealthy(
+  config: LiveSearchRuntimeConfig,
+  timeoutMs = 120_000,
+): Promise<void> {
+  const start = Date.now();
+  let lastError = 'service did not respond';
+
+  while (Date.now() - start < timeoutMs) {
+    try {
+      const response = await fetch(`${config.baseUrl}/health`);
+      if (response.ok) {
+        return;
+      }
+
+      lastError = `status ${response.status}`;
+    } catch (error) {
+      lastError = error instanceof Error ? error.message : String(error);
+    }
+
+    await delay(1_000);
+  }
+
+  throw new Error(`Live AdvisoryAI health check failed at ${config.baseUrl}/health: ${lastError}`);
+}
+
+export function ensureLocalCliPublished(config: LiveSearchRuntimeConfig): string {
+  const publishDirectory = path.join(config.repoRoot, '.artifacts', 'stella-cli');
+  const cliBinaryName = process.platform === 'win32' ? 'StellaOps.Cli.exe' : 'StellaOps.Cli';
+  const cliBinaryPath = path.join(publishDirectory, cliBinaryName);
+
+  if (existsSync(cliBinaryPath)) {
+    return cliBinaryPath;
+  }
+
+  execFileSync(
+    'dotnet',
+    [
+      'publish',
+      path.join(config.repoRoot, 'src', 'Cli', 'StellaOps.Cli', 'StellaOps.Cli.csproj'),
+      '-c',
+      'Release',
+      '-o',
+      publishDirectory,
+    ],
+    {
+      cwd: config.repoRoot,
+      env: process.env,
+      stdio: 'pipe',
+    },
+  );
+
+  if (!existsSync(cliBinaryPath)) {
+    throw new Error(`CLI publish completed without producing ${cliBinaryPath}.`);
+  }
+
+  return cliBinaryPath;
+}
+
+export function runSourcesPrepare(
+  config: LiveSearchRuntimeConfig,
+  cliBinaryPath = ensureLocalCliPublished(config),
+): Record<string, unknown> {
+  const stdout = execFileSync(
+    cliBinaryPath,
+    ['advisoryai', 'sources', 'prepare', '--json'],
+    {
+      cwd: config.repoRoot,
+      env: {
+        ...process.env,
+        STELLAOPS_BACKEND_URL: config.baseUrl,
+      },
+      stdio: 'pipe',
+      encoding: 'utf8',
+    },
+  );
+
+  return parseJsonObject(stdout, 'sources prepare');
+}
+
+export async function rebuildLiveIndexes(
+  config: LiveSearchRuntimeConfig,
+): Promise<{ knowledge: Record<string, unknown>; unified: Record<string, unknown> }> {
+  const adminHeaders = buildLiveHeaders(config, 'advisory-ai:admin');
+
+  const knowledge = await postJson(`${config.baseUrl}/v1/advisory-ai/index/rebuild`, adminHeaders);
+  const unified = await postJson(`${config.baseUrl}/v1/search/index/rebuild`, adminHeaders);
+
+  return { knowledge, unified };
+}
+
+export async function runGroundedSmokeQuery(
+  config: LiveSearchRuntimeConfig,
+): Promise<Record<string, unknown>> {
+  return postJson(
+    `${config.baseUrl}/v1/search/query`,
+    buildLiveHeaders(config),
+    {
+      q: 'database connectivity',
+      k: 5,
+      includeSynthesis: false,
+      ambient: {
+        currentRoute: '/ops/operations/doctor',
+      },
+    },
+  );
+}
+
+export function writeLiveSearchSetupArtifact(
+  config: LiveSearchRuntimeConfig,
+  report: LiveSearchSetupReport,
+): string {
+  const outputDirectory = path.join(config.webRoot, 'output', 'playwright');
+  const outputPath = path.join(outputDirectory, 'live-search-setup.json');
+  mkdirSync(outputDirectory, { recursive: true });
+  writeFileSync(outputPath, `${JSON.stringify(report, null, 2)}\n`, 'utf8');
+  return outputPath;
+}
+
+export function buildLiveHeaders(
+  config: LiveSearchRuntimeConfig,
+  scopes = config.scopes,
+): Record<string, string> {
+  return {
+    'content-type': 'application/json',
+    'x-stellaops-scopes': scopes,
+    'x-stellaops-tenant': config.tenant,
+    'x-stellaops-actor': config.actor,
+  };
+}
+
+async function postJson(
+  url: string,
+  headers: Record<string, string>,
+  body?: Record<string, unknown>,
+): Promise<Record<string, unknown>> {
+  const response = await fetch(url, {
+    method: 'POST',
+    headers,
+    body: body ? JSON.stringify(body) : undefined,
+  });
+
+  const raw = await response.text();
+  if (!response.ok) {
+    throw new Error(`Request to ${url} failed with status ${response.status}: ${raw}`);
+  }
+
+  return parseJsonObject(raw, url);
+}
+
+function parseJsonObject(raw: string, source: string): Record<string, unknown> {
+  const trimmed = raw.trim();
+  if (!trimmed) {
+    return {};
+  }
+
+  try {
+    const parsed = JSON.parse(trimmed) as Record<string, unknown>;
+    return parsed && typeof parsed === 'object' ? parsed : {};
+  } catch {
+    const extracted = tryExtractEmbeddedJson(trimmed);
+    if (extracted) {
+      return extracted;
+    }
+
+    throw new Error(`Expected JSON output from ${source}, received: ${trimmed.slice(0, 400)}`);
+  }
+}
+
+function tryExtractEmbeddedJson(raw: string): Record<string, unknown> | null {
+  const firstBrace = raw.indexOf('{');
+  const lastBrace = raw.lastIndexOf('}');
+  if (firstBrace >= 0 && lastBrace > firstBrace) {
+    const candidate = raw.slice(firstBrace, lastBrace + 1);
+    try {
+      const parsed = JSON.parse(candidate) as Record<string, unknown>;
+      return parsed && typeof parsed === 'object' ? parsed : {};
+    } catch {
+      // Fall through to narrower candidates.
+    }
+  }
+
+  for (let index = raw.indexOf('{'); index >= 0; index = raw.indexOf('{', index + 1)) {
+    try {
+      const parsed = JSON.parse(raw.slice(index)) as Record<string, unknown>;
+      return parsed && typeof parsed === 'object' ? parsed : {};
+    } catch {
+      // Keep scanning for the start of the JSON payload.
+    }
+  }
+
+  return null;
+}
+
+function findRepoRoot(startDirectory: string): string {
+  let current = path.resolve(startDirectory);
+
+  while (true) {
+    if (isRepoRoot(current)) {
+      return current;
+    }
+
+    const parent = path.dirname(current);
+    if (parent === current) {
+      throw new Error(`Unable to locate the Stella Ops repository root from ${startDirectory}.`);
+    }
+
+    current = parent;
+  }
+}
+
+function isRepoRoot(candidate: string): boolean {
+  return existsSync(path.join(candidate, 'global.json'))
+    && existsSync(path.join(candidate, 'src'))
+    && existsSync(path.join(candidate, 'docs'))
+    && existsSync(path.join(candidate, 'devops'));
+}
+
+function delay(ms: number): Promise<void> {
+  return new Promise((resolve) => setTimeout(resolve, ms));
+}

src/Web/StellaOps.Web/tests/e2e/unified-search-contextual-suggestions.live.e2e.spec.ts

@@ -1,12 +1,18 @@
 import { expect, test, type Page, type Route } from '@playwright/test';
 
 import { policyAuthorSession } from '../../src/app/testing';
+import {
+  buildLiveHeaders,
+  ensureLiveServiceHealthy,
+  rebuildLiveIndexes,
+  resolveLiveSearchRuntimeConfig,
+} from './support/live-search-support';
 import { waitForEntityCards, waitForResults } from './unified-search-fixtures';
 
-const liveSearchBaseUrl = process.env['LIVE_ADVISORYAI_SEARCH_BASE_URL']?.trim() ?? '';
+const liveRuntime = resolveLiveSearchRuntimeConfig();
-const liveTenant = process.env['LIVE_ADVISORYAI_TENANT']?.trim() || 'test-tenant';
+const liveSearchBaseUrl = liveRuntime.baseUrl;
-const liveScopes = process.env['LIVE_ADVISORYAI_SCOPES']?.trim()
+const liveTenant = liveRuntime.tenant;
-  || 'advisory-ai:view advisory-ai:operate advisory-ai:admin';
+const liveScopes = liveRuntime.scopes;
 
 const mockConfig = {
   authority: {
@@ -162,8 +168,10 @@ test.describe('Unified Search - Live contextual suggestions', () => {
 
   test.beforeAll(async ({}, testInfo) => {
     testInfo.setTimeout(120_000);
-    await ensureLiveServiceHealthy(liveSearchBaseUrl);
+    await ensureLiveServiceHealthy(liveRuntime);
-    await rebuildLiveIndexes(liveSearchBaseUrl);
+    if (process.env['PLAYWRIGHT_LIVE_SEARCH_SKIP_REBUILD'] !== '1') {
+      await rebuildLiveIndexes(liveRuntime);
+    }
     liveRouteStates.clear();
 
     for (const routeConfig of liveRouteConfigs) {
@@ -319,6 +327,34 @@ test.describe('Unified Search - Live contextual suggestions', () => {
     expect(String(capturedTurnBodies.at(-1)?.['content'] ?? '')).toMatch(/grounded answer|best next step/i);
   });
 
+  test('persists a grounded live search in recent history and clears it with the icon action', async ({ page }) => {
+    await routeLiveUnifiedSearch(page);
+    await openDoctor(page);
+
+    const searchInput = page.locator('app-global-search input[type="text"]');
+    await searchInput.focus();
+    await waitForResults(page);
+
+    await page.locator('.search__suggestions .search__chip', {
+      hasText: /database connectivity/i,
+    }).first().click();
+
+    await expect(searchInput).toHaveValue('database connectivity');
+    await waitForResults(page);
+    await expect(page.locator('[data-answer-status="grounded"]')).toBeVisible();
+
+    await searchInput.fill('');
+    await waitForResults(page);
+
+    const recentGroup = page.locator('.search__group').filter({ hasText: 'Recent' });
+    await expect(recentGroup).toContainText('database connectivity');
+
+    const clearButton = page.locator('.search__clear-history');
+    await expect(clearButton).toBeVisible();
+    await clearButton.click();
+    await expect(recentGroup).toHaveCount(0);
+  });
+
   for (const routeConfig of liveRouteConfigs.filter((route) => route.key !== 'doctor')) {
     test(`${routeConfig.label} suppresses surfaced starter chips when the live route corpus is unready`, async ({ page }) => {
       const state = requireLiveRouteState(routeConfig.key);
@@ -534,6 +570,8 @@ async function routeLiveUnifiedSearch(
   page: Page,
   capturedRequests?: Array<Record<string, unknown>>,
 ): Promise<void> {
+  const liveHeaders = buildLiveHeaders(liveRuntime);
+
   await page.route('**/api/v1/search/query', async (route) => {
     const rawBody = route.request().postData() ?? '{}';
     const parsedBody = safeParseRequest(rawBody);
@@ -543,12 +581,7 @@ async function routeLiveUnifiedSearch(
 
     const response = await fetch(`${liveSearchBaseUrl}/v1/search/query`, {
       method: 'POST',
-      headers: {
+      headers: liveHeaders,
-        'content-type': 'application/json',
-        'x-stellaops-scopes': liveScopes,
-        'x-stellaops-tenant': liveTenant,
-        'x-stellaops-actor': 'playwright-live',
-      },
       body: rawBody,
     });
 
@@ -571,38 +604,6 @@ async function routeLiveUnifiedSearch(
   });
 }
 
-async function ensureLiveServiceHealthy(baseUrl: string): Promise<void> {
-  const response = await fetch(`${baseUrl}/health`);
-  if (!response.ok) {
-    throw new Error(`Live AdvisoryAI health check failed with status ${response.status}.`);
-  }
-}
-
-async function rebuildLiveIndexes(baseUrl: string): Promise<void> {
-  const headers = {
-    'content-type': 'application/json',
-    'x-stellaops-scopes': 'advisory-ai:admin',
-    'x-stellaops-tenant': liveTenant,
-    'x-stellaops-actor': 'playwright-live',
-  };
-
-  const knowledgeResponse = await fetch(`${baseUrl}/v1/advisory-ai/index/rebuild`, {
-    method: 'POST',
-    headers,
-  });
-  if (!knowledgeResponse.ok) {
-    throw new Error(`Knowledge rebuild failed with status ${knowledgeResponse.status}.`);
-  }
-
-  const unifiedResponse = await fetch(`${baseUrl}/v1/search/index/rebuild`, {
-    method: 'POST',
-    headers,
-  });
-  if (!unifiedResponse.ok) {
-    throw new Error(`Unified rebuild failed with status ${unifiedResponse.status}.`);
-  }
-}
-
 function safeParseRequest(rawBody: string): Record<string, unknown> {
   try {
     const parsed = JSON.parse(rawBody) as Record<string, unknown>;
@@ -613,12 +614,7 @@ function safeParseRequest(rawBody: string): Record<string, unknown> {
 }
 
 async function fetchLiveSuggestionViability(rawBody: string): Promise<Record<string, unknown>> {
-  const headers = {
+  const headers = buildLiveHeaders(liveRuntime);
-    'content-type': 'application/json',
-    'x-stellaops-scopes': liveScopes,
-    'x-stellaops-tenant': liveTenant,
-    'x-stellaops-actor': 'playwright-live',
-  };
 
   const directResponse = await fetch(`${liveSearchBaseUrl}/v1/search/suggestions/evaluate`, {
     method: 'POST',