Four domains. One governance architecture. Each jurisdiction implements the same constitutional pattern — deterministic laws, audit-chained evidence, human authority at every boundary — in a different operational context.
The constitutional governance layer for autonomous agents with shell and browser access. JSON in, JSON out. Three decision states. Full violation reporting with precedence traces and session-level composition tracking.
Deterministic safety boundaries between AI proposals and flight execution. Geofencing, fuel reserves, and crew authority compiled to pure-function rules at every decision point.
Human authorship protected as a formal governance constraint. AI may draft prose; it cannot decide fate. Character mortality and plot decisions are human-only state transitions with full audit trails.
CI/CD tells you what deployed. Vigil tells you what executed. Light Station tells you what the Principal needs to decide — a persistent operating picture of agent proposals, policy verdicts, and human escalations.
The governance pattern doesn't change between jurisdictions — only the laws composed within it. Vigil enforces Boundary, Resource, and Authority Laws for desktop agents. FlightLaw enforces Spatial, Resource, and Authority Laws for autonomous aviation. The kernel, the evidence chain, and the verifier are the same underneath.
This is what "constitutional" means in practice: a single architectural contract that holds across every domain it's applied to. The framework is the proof. The jurisdictions are the evidence.
AgentVector framework →