Skip to content
Michael Rishi Forrester

Agentic Covenants

Agentic Inventory Matrix

What agents exist, what they touch, what threats they face?

Filter

All 15Self-declaredOperator-declaredDiscoveredIdentityAuthorizationBlast radiusApproval gatingSupply chain

Layers, by what they can actually enforce

  • Self-declared — advisory

    The agent reports itself. It can be wrong, stale, or absent entirely.

  • Operator-declared — deterministic-when-invoked

    The operator registered it. Accurate only as far as the operator knows.

  • Discovered — deterministic-and-external

    Found independently by scanning the target. Trusts neither of the above.

Showing 5 of 15.

  • Identity at Discovered

    CloudTrail / GCP Audit Logs of SA and IAM principal usage. K8s controller watching SA+RoleBinding by naming pattern. Reverse-lookup from credential fingerprints in Sentinels.

  • Authorization at Discovered

    K8s RBAC API list, AWS IAM Access Analyzer effective permissions, Kyverno PolicyReports of policies actually applied.

  • Blast radius at Discovered

    Threat-modeling output (MAESTRO Layer 7, MITRE ATLAS, lateral-movement path analysis). Behavioral observation of what the agent has touched.

  • Approval gating at Discovered

    Registry of charter files in source control. Last-modified, last-PR-merged. Cross-reference with self-declared and operator-declared.

  • Supply chain at Discovered

    Image-registry pull events, package-manager logs, runtime introspection of loaded models and connected MCP servers, SBOM scanning. Drift between actual and authorized = alert.