HyperAI User Control System

The HyperAI User Control System is the orchestrator for the autonomy dashboards. The active runtime surface is autonomy-and-symphony first: the browser shell promotes the autonomy/symphony command plane backed by backend/server.js on port 5000, while every other lane remains quarantined, degraded, local-only, or optional until the Autonomous Queue deliberately promotes it.

Runtime truth

• Product surface: C:\Users\pc\HyperAI_Phoenix_Master\hyperai-user-control-system
• Frontend entry: src/main.tsx
• App root: src/App.tsx
• CI/runtime backend path: backend/server.js (listens on port 5000)
• Proven live shell endpoints: GET /api/health, GET /api/symphony/status, POST /api/symphony/start, POST /api/symphony/stop
• Runtime authority probe: GET /api/runtime/capabilities
• Runtime classification probe: GET /api/runtime/state reports Dormant/Recoverable/Operational/Autonomous plus heartbeat/probe proofs so the dashboard can display the boundary’s current status.
• CI source of truth: .github/workflows/ci.yml
• Hard verification gates: npm run ci:build, npm run ci:browser-smoke

Runtime rules

1. Always probe port 5000 before touching backend/server.js. If another process already owns the port, document its PID and start time and treat it as the live process until it is intentionally recycled.
2. Non-core browser lanes (backend/server.ts, NotebookLM, empathy, Vietnamese, user CRUD, websocket routes) remain quarantined behind the VITE_ENABLE_* flags; do not assume they are authoritative without new Autonomous Queue evidence.
3. Use /api/runtime/capabilities and the Runtime Lane Status cockpit to distinguish a fresh backend listener from a stale process that merely still responds.
4. Follow the Autonomous Queue cycle before pulling new runtime conclusions:
   • python tools/hyperai_autonomous_cycle.py
   • memory/master_autonomous_todo.md
   • memory/runtime_execution_todo.md Process-first and delta-first means that if nothing changed since the last cycle, do not restart the backend nor run expensive verification.

Quarantine flags

• VITE_ENABLE_VIETNAMESE_RUNTIME
• VITE_ENABLE_EMPATHY_RUNTIME
• VITE_ENABLE_USER_RUNTIME
• VITE_ENABLE_NOTEBOOKLM_RUNTIME
• VITE_ENABLE_WEBSOCKET_RUNTIME

When these flags are true, the corresponding lane is released from fallback mode. When they are unset or false, clients return static/local-only responses and the lane is treated as non-live in documentation and tooling.

Memory & autonomy playbooks

• Record every runtime decision in the memory capsules: memory/agent1_*.md through memory/agent6_*.md, and keep memory/agent6_synthesis_capsule.md in sync after major changes.
• Use the Autonomous Queue before changing runtime or toolchain knowledge. The queue files track the outstanding work for Russell (tooling drift), Erdos (backend contracts), Kuhn (frontend composition), and Sartre (synthesis/backlog refresh).
• The README and scripts should cite those artifacts when describing required commands or verification steps.

Tooling & verification narrative

• Required: npm run ci:build, npm run ci:browser-smoke. The CI workflows (.github/workflows/ci.yml → build-artifact + browser-agent-smoke) rely on these gates to prove the symphony path.
• Advisory: npm test -- --runInBand, npm run lint. These commands now execute via scripts/tooling/run-legacy-jest.mjs and scripts/tooling/run-legacy-eslint.mjs, enforcing jest.config.js and .eslintrc.cjs. They exit early with a clear message when the symphony runtime lane is not enabled because the broader lanes are intentionally quarantined.
• When the advisory commands fail, note the failure in the Autonomous Queue instead of assuming the live backend is broken. The queue already tracks the tooling drift work for Russell.

Command guidance

• scripts/build.sh: installs dependencies, probes port 5000, warns that backend/server.ts is informational only, and runs npm run ci:build.
• scripts/test.sh: runs npm run ci:build, then the advisory legacy tests and lint (npm test -- --runInBand || true, npm run lint || true), while restating that they are non-critical because the symphony lane is locked.
• scripts/deploy.sh: enforces the local-first rule—if port 5000 already has a listener, it reports the owning PID/start time, refuses to boot a second backend/server.js, and instructs the operator to resolve the stale process first.
• scripts/setup-dev.sh: provides the same runtime story while bootstrapping dependencies; it recommends running the Autonomous Queue, probing port 5000, and using npm run ci:build as the primary verification gate.

Running locally

1. npm install
2. python tools/hyperai_autonomous_cycle.py
3. Probe port 5000 before restarting the backend; capture the owning PID and timestamps.
4. npm run ci:build
5. Run npm run ci:browser-smoke only when dashboard routing or the active symphony endpoints change.

Contribution & reporting

Before piecework, check memory/master_autonomous_todo.md to understand which agent owns the scope and how it still needs to evolve. Update the queue, re-run the cycle, and refresh the relevant capsule when you finalize a decision so the automation queue and memory/project_state.json stay aligned.