- api/agents: tool-schemas (semantic allowlist), agent-loop (multi-step + budget), circuit-breaker, llm-client/tool-executor interfaces - api/agents: WpRestExecutor (real WP REST), OpenRouterLlmClient - e2e/agent: multi-step prompts + run-agent-spike runner (scores rendered changes) - ADR 0017; 101 unit tests green
- e2e: Playwright chat-flow (sign up -> chat) with api/web webServer harness - api: sandbox services subset/media-proxy/manifest/gc + DockerClient contract - api: SandboxManager orchestrator (fake-Docker tested); /health route - 53 unit tests green; Playwright e2e green
Stand up the monorepo skeleton and land the first Phase 0 artifacts: pairing threat model, plugin catalog, builder detect, 20-prompt golden harness, and synthetic 2GB mirror timing.