test(e2e): add chat acceptance bundle

AGENTS.md

@@ -216,6 +216,8 @@ cd ../web; corepack.cmd pnpm typecheck; corepack.cmd pnpm exec vite build
 
 UI 工作流变更必须有行为断言，不只截图：共享 unit/contract + Desktop/Web Playwright + Visual QA，证据等级按 `real-e2e-acceptance` 标注。Vite renderer 不等于 packaged Desktop；stub/fixture/readiness-only 必须写 `real_tested=false`。
 
+跨平台前端 acceptance、Playwright/Visual QA 聚合和 manifest 生成优先用 Node/TypeScript runner。PowerShell 只用于 Windows、Tauri、installer、sidecar、release、签名等 OS-specific gate；不要为普通前端验收继续新增大型 `.ps1` 核心脚本。
+
 禁止无保护力测试：
 
 | 反模式 | 禁止原因 |

app/desktop/package.json

@@ -16,6 +16,7 @@
     "test:e2e:chat-flow": "playwright test --config playwright.config.ts --project=chromium chat-flow-ui.spec.ts",
     "test:e2e:data-boundary": "playwright test --config playwright.config.ts --project=chromium desktop-data-boundary.spec.ts",
     "test:visual:chat-flow": "node scripts/manual-chat-flow-check.mjs",
+    "test:acceptance:chat-flow": "pnpm test:e2e:chat-flow && pnpm test:visual:chat-flow",
     "test:e2e:smoke": "playwright test --config playwright.config.ts --project=chromium smoke.spec.ts",
     "test:e2e:ui": "playwright test --ui",
     "analyze": "pnpm build && npx vite-bundle-visualizer",

app/package.json

@@ -21,6 +21,7 @@
     "test:e2e:web": "pnpm --filter agenthub-web test:e2e",
     "test:e2e:desktop": "pnpm --filter agenthub-desktop test:e2e",
     "test:smoke:matrix": "powershell -NoProfile -ExecutionPolicy Bypass -File ../scripts/smoke/verify-e2e-smoke-matrix.ps1 -RepoRoot ..",
+    "test:acceptance:chat-flow": "node ../scripts/verify/chat-acceptance.mjs --repo-root ..",
     "lint:css": "stylelint \"**/*.css\" --ignore-path .stylelintignore"
   },
   "pnpm": {

app/web/package.json

@@ -15,6 +15,7 @@
     "test:e2e:stubbed-hub": "playwright test --config playwright.config.ts --project=chromium chat-flow-contract.spec.ts web-stubbed-hub-replay-smoke.spec.ts task-contract.spec.ts",
     "test:e2e:approved-real-stub": "pnpm test:e2e:stubbed-hub",
     "test:visual:chat-flow": "node scripts/manual-chat-flow-check.mjs",
+    "test:acceptance:chat-flow": "pnpm test:e2e:chat-flow && pnpm test:visual:chat-flow",
     "lint": "eslint src",
     "typecheck": "tsc -p tsconfig.json --noEmit"
   },

docs/progress/MASTER.md

@@ -114,3 +114,4 @@ Per-task telemetry is stored in GitHub issue comments before task closure. Adapt
 | 2026-06-29 | Phase 3 sync | Updated Phase 3 live state after #408: milestone #19 is 2/3 complete, adaptive drift_score is 1, #388 has a drift warning, and T3.3 (#388) is the active next task. |
 | 2026-06-29 | T3.3 implementation | Hardened observed/approved-real manifest boundaries, kept packaged-release claims separate, aligned smoke-matrix contract checks with current stubbed-Hub replay names, and verified shared/contract gates with `real_tested=false`; PR pending. |
 | 2026-06-29 | Phase 3 complete | #388 merged via #410 and closed manually because non-default base did not auto-close it; milestone #19 is closed at 3/3 with adaptive drift_score 2, so Phase 4 requires a lightweight checkpoint before T4.1. |
+| 2026-06-29 | T4.1 implementation | Added a Node-based focused chat acceptance bundle for shared unit, Desktop/Web Playwright, and Desktop/Web Visual QA; package entry passed with `real_tested=false`; PR pending. |

scripts/verify/chat-acceptance.mjs

@@ -0,0 +1,316 @@
+#!/usr/bin/env node
+/*
+ * Focused chat workflow acceptance bundle.
+ *
+ * Runs the merge-useful Desktop/Web chat gates and writes a machine-readable
+ * manifest. This is Vite/browser/fixture evidence only unless a future task
+ * adds a separate approved-real or packaged-release row.
+ */
+
+import { spawn } from 'node:child_process';
+import fs from 'node:fs';
+import path from 'node:path';
+import process from 'node:process';
+
+const args = parseArgs(process.argv.slice(2));
+const repoRoot = path.resolve(args.repoRoot ?? '.');
+const artifactRoot = path.resolve(repoRoot, args.artifactRoot ?? path.join('.tmp', 'chat-acceptance', `run-${process.pid}`));
+const outputPath = path.resolve(repoRoot, args.outputPath ?? path.join(path.relative(repoRoot, artifactRoot), 'chat-acceptance-manifest.json'));
+const timeoutMs = Number(args.commandTimeoutSec ?? 300) * 1000;
+
+fs.mkdirSync(artifactRoot, { recursive: true });
+
+const startedAt = new Date();
+const corepack = process.platform === 'win32' ? 'corepack.cmd' : 'corepack';
+
+const rows = [];
+
+const sharedUnitTests = [
+  'src/chatview/adapter.test.ts',
+  'src/chatview/components/AgentGroup.rendering.test.tsx',
+  'src/chatview/components/UserMessage.rendering.test.tsx',
+  'src/chatview/components/Transcript.autoscroll.test.tsx',
+  'src/chatview/components/Transcript.css.test.ts',
+  'src/transcript/runtimeDiagnostics.test.ts',
+  'src/workbench/AgentHubWorkbench.test.tsx',
+];
+
+await runGate({
+  name: 'shared-chat-unit',
+  surface: 'shared',
+  evidenceLevel: 'fixture-unit',
+  claim: 'Shared transcript ordering, optimistic send, markdown/table, diagnostic filtering, and card grouping unit coverage',
+  cwd: path.join(repoRoot, 'app', 'shared'),
+  command: corepack,
+  args: ['pnpm', '--dir', path.join(repoRoot, 'app', 'shared'), 'exec', 'vitest', 'run', ...sharedUnitTests],
+  skip: args.skipSharedUnit,
+  skipReason: 'skipped by --skip-shared-unit',
+});
+
+await runGate({
+  name: 'desktop-chat-playwright',
+  surface: 'desktop',
+  evidenceLevel: 'playwright-ui',
+  claim: 'Desktop Vite renderer chat flow Playwright coverage; not packaged Desktop',
+  cwd: path.join(repoRoot, 'app', 'desktop'),
+  command: corepack,
+  args: ['pnpm', '--dir', path.join(repoRoot, 'app', 'desktop'), 'run', 'test:e2e:chat-flow'],
+  skip: args.skipDesktopPlaywright,
+  skipReason: 'skipped by --skip-desktop-playwright',
+});
+
+await runGate({
+  name: 'web-chat-playwright',
+  surface: 'web',
+  evidenceLevel: 'playwright-ui',
+  claim: 'Web Vite renderer chat flow Playwright coverage; Hub-shaped stubs are not real login or model execution',
+  cwd: path.join(repoRoot, 'app', 'web'),
+  command: corepack,
+  args: ['pnpm', '--dir', path.join(repoRoot, 'app', 'web'), 'run', 'test:e2e:chat-flow'],
+  skip: args.skipWebPlaywright,
+  skipReason: 'skipped by --skip-web-playwright',
+});
+
+await runGate({
+  name: 'desktop-chat-visual-qa',
+  surface: 'desktop',
+  evidenceLevel: 'visual-qa',
+  claim: 'Desktop chat Visual QA screenshot and geometry checks at 1440x810',
+  cwd: path.join(repoRoot, 'app', 'desktop'),
+  command: corepack,
+  args: ['pnpm', '--dir', path.join(repoRoot, 'app', 'desktop'), 'run', 'test:visual:chat-flow'],
+  artifacts: ['app/desktop/.tmp/manual-chat-flow-uiux/desktop-1440x810-chat-flow.png'],
+  skip: args.skipDesktopVisualQa,
+  skipReason: 'skipped by --skip-desktop-visual-qa',
+});
+
+await runGate({
+  name: 'web-chat-visual-qa',
+  surface: 'web',
+  evidenceLevel: 'visual-qa',
+  claim: 'Web chat Visual QA screenshot and geometry checks at 1440x810',
+  cwd: path.join(repoRoot, 'app', 'web'),
+  command: corepack,
+  args: ['pnpm', '--dir', path.join(repoRoot, 'app', 'web'), 'run', 'test:visual:chat-flow'],
+  artifacts: ['app/web/.tmp/manual-chat-flow-uiux/web-1440x810-chat-flow.png'],
+  skip: args.skipWebVisualQa,
+  skipReason: 'skipped by --skip-web-visual-qa',
+});
+
+const endedAt = new Date();
+const counts = {
+  passed: rows.filter((row) => row.status === 'passed').length,
+  failed: rows.filter((row) => row.status === 'failed').length,
+  skipped: rows.filter((row) => row.status === 'skipped').length,
+  total: rows.length,
+};
+const status = counts.failed > 0 ? 'failed' : counts.passed === 0 ? 'skipped' : counts.skipped > 0 ? 'passed_with_skips' : 'passed';
+const executedRows = rows.filter((row) => row.status !== 'skipped');
+const manifest = {
+  schema: 'agenthub.chat_acceptance_bundle.v1',
+  status,
+  real_tested: false,
+  evidence_levels: unique(executedRows.map((row) => row.evidence_level)),
+  planned_evidence_levels: unique(rows.map((row) => row.evidence_level)),
+  generated_at: endedAt.toISOString(),
+  started_at: startedAt.toISOString(),
+  ended_at: endedAt.toISOString(),
+  duration_ms: endedAt.getTime() - startedAt.getTime(),
+  artifact_root: artifactRoot,
+  counts,
+  boundaries: {
+    real_tokendance_id_login: false,
+    real_cli_or_model_api: false,
+    packaged_desktop: false,
+    signing: false,
+    release_upload: false,
+    production_deploy: false,
+  },
+  exclusions: [
+    'real TokenDance ID login',
+    'real CLI/model/API execution',
+    'packaged Tauri/Desktop installer',
+    'sidecar packaging proof',
+    'signing',
+    'release upload',
+    'production deployment',
+  ],
+  rows,
+};
+
+fs.mkdirSync(path.dirname(outputPath), { recursive: true });
+fs.writeFileSync(outputPath, `${JSON.stringify(manifest, null, 2)}\n`);
+
+console.log(`\nChat acceptance status: ${status}`);
+console.log(`Manifest: ${outputPath}`);
+console.log('Evidence boundary: real_tested=false; no real login, model/API execution, packaged Desktop, signing, release upload, or production deploy.');
+
+if (counts.failed > 0) {
+  process.exitCode = 1;
+}
+
+async function runGate(gate) {
+  const commandText = [gate.command, ...gate.args].map(quoteArg).join(' ');
+  const artifacts = (gate.artifacts ?? []).filter(Boolean);
+
+  if (gate.skip) {
+    rows.push({
+      name: gate.name,
+      surface: gate.surface,
+      evidence_level: gate.evidenceLevel,
+      real_tested: false,
+      status: 'skipped',
+      exit_code: null,
+      duration_ms: 0,
+      command: commandText,
+      working_directory: gate.cwd,
+      claim: gate.claim,
+      evidence: gate.skipReason,
+      artifacts,
+    });
+    console.log(`SKIP  ${gate.name} - ${gate.skipReason}`);
+    return;
+  }
+
+  console.log(`RUN   ${gate.name}`);
+  const started = Date.now();
+  const result = await runCommand(gate.command, gate.args, gate.cwd);
+  const status = result.exitCode === 0 ? 'passed' : 'failed';
+  rows.push({
+    name: gate.name,
+    surface: gate.surface,
+    evidence_level: gate.evidenceLevel,
+    real_tested: false,
+    status,
+    exit_code: result.exitCode,
+    duration_ms: Date.now() - started,
+    command: commandText,
+    working_directory: gate.cwd,
+    claim: gate.claim,
+    evidence: shortenText(`${result.stdout}\n${result.stderr}`),
+    artifacts,
+  });
+  console.log(`${status === 'passed' ? 'PASS' : 'FAIL'}  ${gate.name}${status === 'failed' ? ` - exit ${result.exitCode}` : ''}`);
+}
+
+function runCommand(command, commandArgs, cwd) {
+  return new Promise((resolve) => {
+    const spawnCommand = process.platform === 'win32' ? process.env.ComSpec ?? 'cmd.exe' : command;
+    const spawnArgs = process.platform === 'win32'
+      ? ['/d', '/s', '/c', [command, ...commandArgs].map(quoteArg).join(' ')]
+      : commandArgs;
+    let child;
+    try {
+      child = spawn(spawnCommand, spawnArgs, {
+      cwd,
+      shell: false,
+      windowsHide: true,
+      env: {
+        ...process.env,
+        AGENTHUB_EDGE_AUTH_TOKEN: '',
+        AGENTHUB_CHAT_ACCEPTANCE_ARTIFACT_ROOT: artifactRoot,
+      },
+      });
+    } catch (error) {
+      resolve({ exitCode: 127, stdout: '', stderr: error instanceof Error ? error.message : String(error) });
+      return;
+    }
+    let stdout = '';
+    let stderr = '';
+    let timedOut = false;
+    const timer = setTimeout(() => {
+      timedOut = true;
+      child.kill('SIGTERM');
+      setTimeout(() => child.kill('SIGKILL'), 1000).unref();
+    }, timeoutMs);
+
+    child.stdout.on('data', (chunk) => { stdout += chunk.toString(); });
+    child.stderr.on('data', (chunk) => { stderr += chunk.toString(); });
+    child.on('error', (error) => {
+      clearTimeout(timer);
+      resolve({ exitCode: 127, stdout, stderr: `${stderr}\n${error.message}` });
+    });
+    child.on('close', (code, signal) => {
+      clearTimeout(timer);
+      if (timedOut) {
+        resolve({ exitCode: 124, stdout, stderr: `${stderr}\nTimed out after ${timeoutMs / 1000} seconds; signal=${signal ?? 'none'}` });
+        return;
+      }
+      resolve({ exitCode: code ?? 1, stdout, stderr });
+    });
+  });
+}
+
+function redactSecretLike(value) {
+  return value
+    .replace(/(Authorization:\s*Bearer\s+)[^"'\s,}]+/gi, '$1<redacted-token>')
+    .replace(/\b(bearer\s+)[a-z0-9._-]{12,}/gi, '$1<redacted-token>')
+    .replace(/\b(sk|ghp|gho|ghu|ghs|glpat|xox[baprs])-[-_A-Za-z0-9]{8,}/gi, '<redacted-token>')
+    .replace(/((?:access[_-]?token|refresh[_-]?token|id[_-]?token|client_secret|password)\s*[=:]\s*)[^"'\s,}]+/gi, '$1<redacted-secret>')
+    .replace(/("?(?:access[_-]?token|refresh[_-]?token|id[_-]?token|client_secret|password)"?\s*:\s*")[^"]+/gi, '$1<redacted-secret>');
+}
+
+function shortenText(value, max = 4000) {
+  const safe = redactSecretLike(value);
+  return safe.length <= max ? safe : `${safe.slice(0, max)}\n...<truncated>...`;
+}
+
+function unique(values) {
+  return [...new Set(values)].sort();
+}
+
+function quoteArg(value) {
+  if (!value) return '""';
+  return /[\s"]/.test(value) ? `"${value.replaceAll('"', '\\"')}"` : value;
+}
+
+function parseArgs(rawArgs) {
+  const parsed = {};
+  for (let i = 0; i < rawArgs.length; i += 1) {
+    const arg = rawArgs[i];
+    if (arg === '--') {
+      continue;
+    }
+    switch (arg) {
+      case '--repo-root':
+      case '-RepoRoot':
+        parsed.repoRoot = rawArgs[++i];
+        break;
+      case '--artifact-root':
+      case '-ArtifactRoot':
+        parsed.artifactRoot = rawArgs[++i];
+        break;
+      case '--output-path':
+      case '-OutputPath':
+        parsed.outputPath = rawArgs[++i];
+        break;
+      case '--command-timeout-sec':
+      case '-CommandTimeoutSec':
+        parsed.commandTimeoutSec = rawArgs[++i];
+        break;
+      case '--skip-shared-unit':
+      case '-SkipSharedUnit':
+        parsed.skipSharedUnit = true;
+        break;
+      case '--skip-desktop-playwright':
+      case '-SkipDesktopPlaywright':
+        parsed.skipDesktopPlaywright = true;
+        break;
+      case '--skip-web-playwright':
+      case '-SkipWebPlaywright':
+        parsed.skipWebPlaywright = true;
+        break;
+      case '--skip-desktop-visual-qa':
+      case '-SkipDesktopVisualQa':
+        parsed.skipDesktopVisualQa = true;
+        break;
+      case '--skip-web-visual-qa':
+      case '-SkipWebVisualQa':
+        parsed.skipWebVisualQa = true;
+        break;
+      default:
+        throw new Error(`Unknown argument: ${arg}`);
+    }
+  }
+  return parsed;
+}