
Execution governance for AI agents and autonomous systems. A server-enforced state machine that controls what agents can do and when, with human approval gates and a full audit trail. Agents discover allowed transitions from the API; critical states require human sign-off (hard 403) before bots may enter.
