Run adversarial attack playbooks against a governed agent. Executes red-team playbooks that simulate prompt injection, privilege escalation,

Manually attest or vouch for an agent. AGENT_ID is the DID of the agent to attest (e.g. did:mesh:agent-alpha-001).

Create an attestation record for an agent's current governance posture. The attestation maps the AGT trust score to a risk label and packages

View audit logs.

Return recent injection audit log entries.

Append every governance event to the immutable audit log.

Append every governance event to the immutable audit log.

Append every governance event to the immutable audit log.

Append every governance event to the audit log.

Append every governance event to the immutable audit log.

Create a persistent audit log.

The underlying audit log.

Access the audit log for inspection.

Return a copy of the audit trail.

Return a copy of the audit trail.

Return a copy of the scan audit history.

Return the full audit log.

Return the audit log.

Get audit service statistics.

Get the escalation audit trail.

Get all governance decisions made.

Parameterized fixture returning an AdversarialAuditor for each pack.

Current token count after refilling for elapsed time.

Energy available for discharge (kWh)

Return list of active scanner names.

List which signal sources are configured.

Return average evaluation latency in seconds, or 0.0.

Return average evaluation latency in seconds, or 0.0.

Average cost per query execution.

Average cost per chain execution.

Average latency in milliseconds.

Average latency across all events.

Average query latency in milliseconds.

Average chain latency in milliseconds.

Agent that gives wrong answers.

Function without type hints - should fail.

Derive a badge from the overall score.

Compute current baseline statistics.

Agent call limits and timeout enforcement.

Tool allow/block, content scan, human approval.