Agentic Covenants

Govern (GV) · Identity

Identity at the domain layer

·

Who is allowed to put this agent into the world, under what authority, accountable to what policy, with what retirement criteria?

What this cell does

Domain leadership signs charter authorizing a class of agents. Names roles permitted to create agents and the escalation path.

Cell notes

Charter, Identity / Domain

Structural question. Does each domain that operates agents have a signed domain charter naming who in the domain is permitted to create agents, what the escalation path is, and what audit trail location records charter signatures?

Owner. Domain authority (Platform Engineering Director, Security Director, Customer Engineering Director, etc.). Counter-signed by the AI Governance Council.

Template fragment

This cell is satisfied by §4 (Identity) and §1 (Scope) of ../../templates/domain-charter.md. Specifically:

  • - §1 names the agent classes the domain authorizes.
  • - §4 names the roles in this domain authorized to create agents and the escalation path.

Audit prompts

  • - Which domains operate agents? Is there a domain charter for each?
  • - Who in [domain] is allowed to create agents? Is that role-based or named-individual-based?
  • - Where is the audit trail of charter signatures for this domain?

Operational tie-in

The domain charter constrains who in the domain can sign new agent charters. When a [Domain] employee proposes a new agent, the named role here is who reviews and signs.

Citation

NIST CSF 2.0 GV.RR-02 (roles, responsibilities, and authorities); GV.OV-01 (oversight). NIST AI RMF GOVERN 2.1, GOVERN 5.1 (engaging stakeholders). ISO/IEC 42001 §A.6.2 (AI objectives). EU AI Act Art. 26 (deployer obligations).

Crosswalk

NIST CSF 2 0GV.RR-02, GV.OV-01
NIST AI RMFGOVERN 2.1, GOVERN 5.1
ISO IEC 42001§A.6.2
EU AI ACTArt. 26