The Agent Skills Directory

[SAFE]: The skill uses tools like Bash, Grep, and Read to explore the repository. All programmed commands are used for discovery (finding test files, counting source files, checking configs) and do not perform any destructive or unauthorized actions.
[SAFE]: The instructions establish strong operational boundaries, explicitly directing the agent to only report findings and never modify code or write new tests.
[INDIRECT_PROMPT_INJECTION]: The skill reads and processes the content of source code and test files, which is an inherent risk of its primary function. However, the potential for malicious instructions in the analyzed code to influence the agent is addressed by the analytical and reporting-only nature of the skill. -- Ingestion points: Repository files read via Read and Bash commands in SKILL.md. -- Boundary markers: None identified for delimiting code content during processing. -- Capability inventory: The skill has access to analysis tools but is prohibited from modifying the filesystem. -- Sanitization: There is no specific evidence of input sanitization or filtering of the file contents.

qa