29-pact

AI governance is enforced by mapping organizational structures with D/T/R addresses, applying knowledge clearance, and using verification gradients to maintain agent accountability. This framework provides explicit constraints and audit trails for autonomous operations.

D/T/R addressing is a containment and mapping method used in AI governance to define organizational structures. It coordinates agent accountability by scoping roles and tasks through envelopes, ensuring clearance and traceable audit trails.

To set up AI agent governance rules, create a GovernanceEngine from your YAML organizational definition. You can then run verify_action to check permissions for a specific role and action within the framework.

AI agent governance frameworks integrate directly with MCP tooling for enforcement and tools governance. This integration enables continuous auditing, compliance verification, and action traceability across your autonomous agent infrastructure.

Governance envelopes scope AI agent roles and tasks through monotonic tightening rules. This mechanism ensures that agent permissions are progressively restricted and verified, maintaining strict accountability and compliance boundaries.

Verification gradients apply audit trails and compliance checks for AI agents, but require accurate organizational mapping and MCP tooling integration. Inaccurate D/T/R address definitions or YAML configurations can limit enforcement effectiveness.