Trust Center

Logging and audit trace

Logging is designed for accountable operations and reviewer-ready evidence. We capture request and control-plane metadata without exposing chain-of-thought by default.

Request context

Timestamp, request ID, actor type, and policy profile.

Agent identity

Agent instance ID, cryptographic signature verification state, and task key.

Tool access

Allowed tool scopes, requested invocation, approval outcome, and execution status.

Output governance

Output classification, reviewer disposition, and retention bucket.

Export model for client and reviewer audits

Audit data can be exported by customer environment and time window for procurement reviews, incident analysis, and governance attestations.

Back to trust centerAI governance