The Agent Skills Directory

[SAFE]: The skill incorporates strong safety constraints within its instruction contract, explicitly gating credential access, network calls, and production side effects.
[SAFE]: The routing rules provide clear boundaries, instructing the agent to hand off tasks related to security audits or repository-wide infrastructure repairs to more appropriate specialized skills.
[PROMPT_INJECTION]: The skill processes untrusted repository data such as source code, logs, and error messages to perform its diagnosis, which creates a potential surface for indirect prompt injection. Ingestion points: the skill reads local files and captures tool outputs (logs, test results) during the diagnosis phase. Boundary markers: while the skill emphasizes a 'bug boundary', it does not define specific delimiters for separating data from instructions. Capability inventory: the skill is authorized to perform file edits and execute shell commands for validation purposes. Sanitization: there is no explicit mention of sanitizing or filtering ingested repository data before processing.

bug-fix