Agentic Covenants

Govern (GV) · Supply chain

Supply chain at the server side layer

external · Outside the agent entirely

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

Org-wide allowlist of approved foundation models. Org-wide policy on MCP server approval, third-party dependency approval, vendor risk assessment integrated with procurement.

Cell notes

Charter, Supply chain / Organizational

Structural question. Does the org maintain an allowlist of approved foundation models, an MCP server approval policy, a third-party dependency approval policy, and integrate AI vendor risk into procurement?

Owner. AI Governance Council with Procurement and Security.

Template fragment

§6 (Approved foundation models) and §7 (Vendor risk assessment) of ../../templates/organizational-policy.md.

Audit prompts

  • - What foundation models are on the approved list? When were they last reviewed?
  • - Does procurement integrate AI vendor risk? Is SOC 2 Type II required for production agents?
  • - Are subprocessors of approved vendors reviewed annually?

Operational tie-in

The approved-models list is the upper bound on every agent charter's dependencies.foundation_model. PR review must reject any charter that names a model not on the org list.

Citation

NIST CSF 2.0 GV.SC-01 through GV.SC-10 (cybersecurity supply chain risk management). NIST AI RMF GOVERN 6.1, GOVERN 6.2 (third-party risk), MAP 4.1. ISO/IEC 42001 §A.10 (third party). EU AI Act Art. 25 (relationships along the AI value chain). NIST SP 800-161 Rev. 1. NIST SP 800-218A.

Crosswalk

NIST CSF 2 0GV.SC-01, GV.SC-02, GV.SC-03, GV.SC-04, GV.SC-05, GV.SC-06, GV.SC-07, GV.SC-08, GV.SC-09, GV.SC-10, ID.RA-09
NIST AI RMFGOVERN 6.1, GOVERN 6.2, MAP 4.1
ISO IEC 42001§A.10
EU AI ACTArt. 25
OTHERNIST SP 800-161 Rev. 1, NIST SP 800-218A