The Agent Skills Directory

[SAFE]: The skill is purely instructional and defines a framework for the agent to use when evaluating code quality and structural health in pull requests.
[DATA_EXPOSURE]: The skill explicitly states it only produces findings and does not mutate files or perform external network operations. It is designed to operate within the provided context of a code diff.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input (code diffs). However, it lacks dangerous capabilities such as shell execution or file-system writing, and its output is limited to informative review text for the user, minimizing the impact of potential injection within the reviewed code.

structural-review