agent-authority

Free AI agent authority, permissions and reachability resources for understanding what agents can access before they act.

Agentic Primitives — the open-source trust substrate for AI agents: identity that can sign (ERC-4337), authority as a revocable grant (ERC-7710), receipts the owner carries. Developer kit: packages, contracts, skills.

The authority layer for agentic AI, as a DeepSeek Harness plugin. Governs every tool call against your business rules via TapPass /v1/govern.

Schema-first governance prototype for host-level AI agents: typed capabilities, authorization, provenance, and policy decisions. Ring 0 and CI boundary checks are implemented; runtime enforcement is not.