The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to audit and generate documentation by ingesting and processing untrusted user-provided source code (Python, TypeScript, JavaScript) and REST API definitions. This creates a vulnerability surface for indirect prompt injection where malicious instructions could be embedded within docstrings or code comments to influence the agent's behavior.
Ingestion points: User-supplied source code files, API schemas, and documentation blocks (found across SKILL.md and reference files).
Boundary markers: The skill does not define specific prompt delimiters or instructions to ignore embedded commands within the code being analyzed to isolate the agent's logic from the data.
Capability inventory: The skill involves the agent reading, analyzing, and writing to source files (implied by the documentation tasks).
Sanitization: The instructions do not include steps to sanitize, escape, or validate content extracted from untrusted code comments before processing or re-inserting it into documentation.
[EXTERNAL_DOWNLOADS]: The skill references standard documentation linting and coverage packages, including pydocstyle, interrogate, and eslint-plugin-jsdoc. These are well-known, reputable tools within the development ecosystem for ensuring documentation quality.
[COMMAND_EXECUTION]: The skill provides instructions and examples for executing commands associated with documentation linting and coverage reporting (e.g., pydocstyle --convention=google src/). These represent standard development workflows.

documentation