Harden lead capture queue
Builder + Sentinel
Next: Verify production observability gates before launch.
Mission Control is the coordination layer for LightGrid operators, missions, governance, memory, and supervised AI recommendations.
Control Pattern
Review submitted DUCK proof, triage contributor follow-up, and leave an audit trail.
Forming
watchCore stewards and builders are being mapped into accountable lanes.
4 reviews
watchActions are staged for review, simulation, or approval only.
71%
readyPriority missions have owners, but execution gates remain supervised.
Seeded
stableStatic operating memory is available for Phase 1 planning.
Locked
stableNo wallet, transfer, or token execution is connected.
Watching
watchSignals are manually modeled until adapters are approved.
Guarded
stableAI recommendations are filtered through policy and human approval.
Draft
blockedReplication checklist is not yet approved for field execution.
Operational work is staged as assigned, queued, blocked, or under review.
Builder + Sentinel
Next: Verify production observability gates before launch.
Scribe + Steward
Next: Draft first operator readiness checklist.
Architect + Steward
Next: Separate advisory proposals from executable authority.
Scribe
Next: Summarize RSVPs, ticket CTAs, and partner follow-up.
Architect
Next: Wait for policy and land/legal review criteria.
Treasury Keeper
Next: Keep all actions simulated until approval ledger exists.
Governance work is modeled as review and simulation, not live execution.
Reviewer: Steward Circle
Role authority cannot grant execution rights yet.
Reviewer: Architect + Sentinel
Simulation only; no governance execution is live.
Reviewer: Treasury Keeper
No wallet integration exists in Phase 1.
Reviewer: Scribe + Steward
Metrics inform decisions; they do not automate decisions.
Roles define responsibility, not unchecked authority.
System design, dependency planning, and readiness decomposition.
Threat modeling, policy risk detection, and control validation.
Implementation planning, test coverage, and safe patch proposals.
Governance interpretation, approvals, and escalation routing.
Decision records, summaries, onboarding, and operator memory.
Authenticated operator cockpit for missions, recommendations, approval requests, and decision history.
Hydration
Locked
Operator
No authenticated operator
Authority
Read-only fallback is disabled when persistence is configured.
Mission Control requires an authenticated operator session.
Admin
Can review approvals and manage operator-grade gates.
Operator
Can create missions, recommendations, and action requests.
Viewer
Read-only operator context; mutation controls stay gated.
Empty state
Panels render clear empty messages when no records are available.
Loading state
Server hydration keeps loading off the client until authenticated data is ready.
Error state
Hydration failures fall back safely without exposing backend internals.
Durable mission records when Supabase is available; fallback records otherwise.
Operator-facing form shells for protected API workflows. Browser forms do not carry server secrets.
Form submission requires authenticated operator authority through protected API routes. No execution action is available from this console.
Requests can be proposed or reviewed. Approval never executes the request.
AI signals are classified for human review and optional action-request creation.
Append-only decision records for action requests.
Append-only review records for AI-generated signals.
Append-only audit trail for mutation attempts, denials, failures, and accepted changes.
Trace related API responses and audit events by request ID. Request IDs are diagnostic handles, not credentials or approval proof.
Signals are modeled locally in Phase 1. No external telemetry adapters are connected.
Execution disabled until policy ledger is connected.
Decision records need source citations before approval use.
External adapters are intentionally disconnected in Phase 1.
Treasury actions remain simulated and awaiting human approval.
Recommendations are advisory. Policy filters and humans approve before controlled workflow.
Requires operator review; no autonomous deployment.
Policy review required before governance routing.
Human approval required for field execution.
Execution readiness is gated by identity, policy, approval, and audit design.
Operator archetypes and roles exist as seed model.
Mission types are modeled but not yet governance-approved.
Decision pattern is defined; durable policy ledger is next.
Ledger concept exists; Mission Control does not persist yet.
AI Recommendation -> Policy Review -> Human Approval -> Controlled Execution -> Audit Ledger.
Phase 1 uses typed seed data only.
No production automation or external API connection is present.
All actions remain proposed, queued, simulated, or awaiting approval.
Queued for steward approval; no action executes from this page.
Simulation only; no vote execution is connected.
Awaiting approval; no wallet, transfer, or token action exists.
Proposed update requires review before becoming operating memory.
Phase 2 adds server-only mission persistence with strict validation. The page still falls back to typed seed data when Supabase is unavailable.
Mission records are modeled as proposed, active, blocked, completed, or archived. Operational work remains queued for operator review.
Policy outcomes are append-only by design and are not exposed through update or delete repository methods.
Every sensitive recommendation should become a policy decision record before approval or denial.
Authority is represented as a boundary, not as a completed RBAC system.
API writes are gated server-side. Authenticated operator roles are intentionally deferred until the identity layer is finalized.
Governance actions can be stored as draft, review, approved, rejected, or deferred.
Governance records capture required authority and deadlines without implying live governance execution.
Readiness gates persist the conditions required before later phases can advance.
Gates are tracked as not started, in progress, satisfied, or blocked so operators can see what authority and evidence are missing.
Identity -> persistence -> policy ledger -> approval flow -> AI recommendation queue.
No autonomous agents are introduced in this phase. The system is preparing durable mission records and audit-ready policy decisions first.
Phase 3 introduces an authenticated operator model before any AI recommendation queue or execution workflow.
Operator identity is resolved server-side. Temporary development headers are placeholders only and are not a production authentication pattern.
Mutations require explicit Mission Control authorities such as mission:create or policy_decision:append.
Operator authority is now the required gate for Mission Control mutations. Admins receive all authorities, operators receive scoped mutation rights, and viewers remain read-only.
Mission, readiness, governance, and policy ledger writes are checked before validation side effects.
Unauthorized requests are denied. Insufficient-authority requests are rejected. Public errors remain generic.
Action request types are modeled for future approval workflow, but no execution routes exist.
Action requests can be proposed, reviewed, approved, denied, or expired in later phases. Autonomous execution remains disabled.
Operator context is a protected server concern, not a client-controlled role claim.
AI recommendations are still deferred until the approval and policy layers mature.
Phase 4 stores approval requests and decision history. Approval is not execution.
Requires mission:update. Requester and approver must be separate operators. Approved requests authorize only future controlled workflow.
Requires governance_action:update. Requester and approver must be separate operators. Approved requests authorize only future controlled workflow.
Approve, reject, and expire decisions append immutable history records.
Status transitions stop at approved, denied, or expired. Approved does not mean executed, and no execution adapter is connected.
Phase 5 accepts AI-generated signals for policy classification and human review only.
Create an operator-reviewed action request for lead queue observability verification.
AI may recommend. Operators may request. Admins may approve. Systems still do not execute.
Keep node replication as a reviewed planning item until policy gates are satisfied.
AI may recommend. Operators may request. Admins may approve. Systems still do not execute.
Recommendation intake -> policy classification -> human review -> optional action-request creation -> no execution.
Accepted recommendations become reviewed signals. Action-worthy recommendations must be converted into Phase 4 action requests and still require approval before any future controlled workflow path.
This is the first static operating surface for Mission Control. It is not a completed automation platform.
These are planned implementation steps, not live integrations.