feat(e2e): add test daemon CLI with ANSI dashboard and Agent SDK bridge

- index.ts: CLI entry point (--full, --spec, --watch, --agent flags)
- runner.ts: programmatic WDIO launcher with result streaming
- dashboard.ts: ANSI terminal UI (pass/fail/skip/running icons, summary)
- agent-bridge.ts: NDJSON stdin/stdout for Agent SDK queries
  (status, rerun, failures, reset-cache)
- Standalone package at tests/e2e/daemon/

tests/e2e/daemon/README.md

@@ -0,0 +1,122 @@
+# E2E Test Daemon
+
+Terminal dashboard for running and monitoring Agent Orchestrator E2E tests. Supports smart caching, file watching, and agent SDK integration.
+
+## Prerequisites
+
+- Built Tauri debug binary (`npm run tauri build -- --debug --no-bundle`)
+- `tauri-driver` installed (`cargo install tauri-driver`)
+- Node.js 20+
+
+## Install
+
+```bash
+cd tests/e2e/daemon
+npm install
+```
+
+The daemon reuses WDIO deps from the root `node_modules`. Its own `package.json` lists `@wdio/cli` and `@wdio/local-runner` for clarity, but they resolve from the workspace root.
+
+## Usage
+
+```bash
+# Run all specs (with smart cache — skips recently-passed specs)
+npm start
+
+# Full run — reset cache, run everything
+npm run start:full
+
+# Filter by spec name pattern
+npx tsx index.ts --spec phase-a
+
+# Watch mode — re-run on spec file changes
+npm run start:watch
+
+# Agent mode — accept NDJSON queries on stdin, respond on stderr
+npm run start:agent
+
+# Combine flags
+npx tsx index.ts --full --watch --agent
+```
+
+## Dashboard
+
+The terminal UI shows:
+
+```
+  Agent Orchestrator — E2E Test Daemon  RUNNING  12.3s
+  ────────────────────────────────────────────
+  ✓ smoke                        1.2s
+  ✓ workspace                    0.8s
+  ⟳ settings
+  · phase-a-structure
+  · phase-a-agent
+  ⏭ phase-b-grid
+
+  ────────────────────────────────────────────
+  2 passed  │  0 failed  │  1 skipped  │  1 running  │  2 pending  │  2.0s
+```
+
+Status icons:
+- `✓` green — passed
+- `✗` red — failed (error message shown below)
+- `⏭` gray — skipped (cached)
+- `⟳` yellow — running
+- `·` white — pending
+
+## Smart Cache
+
+The daemon reads from the shared `test-results/results.json` (managed by `results-db.ts`). Specs that passed in any of the last 5 runs are skipped unless `--full` is used.
+
+## Agent Bridge Protocol
+
+When started with `--agent`, the daemon accepts NDJSON queries on **stdin** and responds on **stderr** (stdout is used by the dashboard).
+
+### Queries
+
+**Status** — get current test state:
+```json
+{"type": "status"}
+```
+Response:
+```json
+{"type": "status", "running": false, "passed": 15, "failed": 2, "skipped": 1, "pending": 0, "total": 18, "failures": [{"name": "phase-b-grid", "error": "WDIO exited with code 1"}]}
+```
+
+**Rerun** — trigger a new test run:
+```json
+{"type": "rerun", "pattern": "phase-a"}
+```
+Response:
+```json
+{"type": "rerun", "specsQueued": 1}
+```
+
+**Failures** — get detailed failure list:
+```json
+{"type": "failures"}
+```
+Response:
+```json
+{"type": "failures", "failures": [{"name": "phase-b-grid", "specFile": "phase-b-grid.test.ts", "error": "WDIO exited with code 1"}]}
+```
+
+**Reset cache** — clear smart cache:
+```json
+{"type": "reset-cache"}
+```
+Response:
+```json
+{"type": "reset-cache", "ok": true}
+```
+
+## Architecture
+
+```
+index.ts          CLI entry point, arg parsing, main loop
+runner.ts         WDIO Launcher wrapper, spec discovery, smart cache
+dashboard.ts      ANSI terminal UI (no external deps)
+agent-bridge.ts   NDJSON stdio interface for agent integration
+```
+
+The daemon reuses the project's existing `wdio.conf.js` and `infra/results-db.ts`.