The Agent Skills Directory

[SAFE]: The skill establishes a detailed and systematic process for code review, focusing on quality metrics such as mutation scores, data state verification, and repeatability. It does not contain instructions for data exfiltration or unauthorized system access.
[COMMAND_EXECUTION]: The skill utilizes local development tools including git diff, python scripts (e.g., code_graph), and standard mutation testing frameworks such as Stryker, PITest, and mutmut. These are used appropriately within the context of a code quality review and do not represent a security risk.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted content (changed production code and test files) from the repository. This is the primary function of the skill; while this creates an attack surface, the skill includes structured phases and specific quality gates that act as a framework for the agent's analysis, reducing the impact of potential injection within the reviewed code. The risk is classified as low and inherent to the tool's purpose.

integration-test-review