The Agent Skills Directory

[SAFE]: The skill's behavior is consistent with its stated purpose of auditing codebases for lifecycle and state-machine bugs. It does not perform unexpected network requests, hardcode credentials, or attempt to persist on the system.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it is designed to ingest and process arbitrary codebase content.
Ingestion points: External code files located at the user-specified path or within the entire repository.
Boundary markers: The skill does not explicitly define boundary markers or delimiters for the code it audits within the instruction brief.
Capability inventory: The skill has the capability to perform autonomous investigations (via figure-out) and create project tickets (via ticket-up).
Sanitization: There is no mention of sanitizing or escaping the audited code content before it is processed by the agent.

audit-state-machines