The Agent Skills Directory

[SAFE]: The skill focuses on analytical tasks and static design review. It includes explicit negative constraints that prohibit the agent from performing risky operations like fault injection, production system access, or credential reading.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external inputs such as retry policies and redacted design fixtures. While this creates a potential surface for indirect injection, the skill mitigates this by enforcing a mandatory 'Input Audit' to verify evidence and a strict requirement to trace every conclusion back to its source.

retry-testing