Agent Governance
Permissions define what an agent is allowed to do. aiclavis reconstructs what the system can actually cause — across authority, evidence, memory, delegation, execution, revocation and recovery.
THE CAUSAL PATH WE RECONSTRUCT
aiclavis does not examine these as separate components. It reconstructs the causal chain they form together.
What we reconstruct
What the system formally permits.
What the system can actually change.
What information entered the decision.
How behavior changes over time.
Whether authority actually terminates.
Whether the accountable owner can really correct the consequence.
Why existing controls are not enough
Identity, permissions, logging, approval, versioning and runtime controls can each work correctly. The harder question is whether their composition still produces the intended governance outcome across the full causal path.
A system can pass every local check and still fail at the level of the chain.
aiclavis reconstructs how authority, evidence, memory and consequential effects propagate through agentic systems — from the initiating principal to the external outcome and, when needed, back through revocation and recovery.