audit-agent

Auditing agent conformance requires reading the standard template, enumerating target agents, and validating each against 15 criteria. The audit reports per-criterion statuses with explicit evidence and locations for any structural gaps found in the definitions.

Agent template conformance checking validates agent definitions against 15 official template criteria to ensure structural alignment. It flags gaps with concrete evidence and line references, supporting repository-wide governance and quality control.

Yes, batch auditing is supported by invoking the tool with the 'all' parameter to enumerate and validate every agent in the repository. This produces per-criterion status reports with evidence for all non-passing findings across the entire codebase.

Running an audit generates per-criterion reports that attach explicit evidence and locations for every non-passing finding. The output includes line references to both the standard template and the specific agent definitions flagged for structural gaps.

Agent template auditing is limited to validating structural conformance against 15 predefined template criteria. It focuses on identifying structural gaps with evidence, but does not evaluate the logical correctness, semantic accuracy, or runtime behavior of the agents.