The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to execute a wide variety of CLI utilities including npx, pytest, ruff, go, and cargo. It also executes a Python script to parse package.json for project detection. The broad access to the shell for running these tools represents a standard but significant capability.
[REMOTE_CODE_EXECUTION]: The skill relies on npx to run several Node.js tools (eslint, biome, madge, knip, vitest, jest). If these tools are not available in the local environment, npx will fetch them from the public npm registry, leading to the execution of external code.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from the local repository to drive its logic.
Ingestion points: Reads project configuration files (package.json, pyproject.toml) and source code files through various analysis tools (SKILL.md).
Boundary markers: The instructions do not define delimiters or specific safety instructions for handling the output of these tools.
Capability inventory: The skill has access to the Bash and Edit tools, allowing it to execute commands and modify files based on tool output.
Sanitization: There is no evidence of sanitization for data extracted from local files before it is processed by the agent. This creates a surface where malicious content in the repository (e.g., in test failures or linter errors) could influence the agent's subsequent actions.

verify