The Agent Skills Directory

[SAFE]: A thorough review of the skill instructions, reference files, and metadata found no evidence of malicious intent, credential harvesting, or unauthorized remote access.
[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves analyzing external source code and test files, which serves as a potential ingestion point for indirect prompt injection. However, the skill implements a strict logical framework (e.g., Value Bar, Retention Bar) that forces the agent to judge code based on assertions and behavior rather than prose, effectively mitigating the risk of following instructions embedded in code comments.
[COMMAND_EXECUTION]: The skill workflow includes the use of standard development tools, such as git for diffing and repository-specific test runners for validation. These operations are transparent, scoped to the auditing task, and follow best practices for verifying changes before landing code.

test-audit