The Agent Skills Directory

[COMMAND_EXECUTION]: The skill's test suite, scripts/test_prose_lint.py, utilizes subprocess.run to execute the local linter script (scripts/prose_lint.py). The execution is performed using list-based arguments, which is a secure practice to prevent command injection, and targets only the skill's own local files.
[EXTERNAL_DOWNLOADS]: The documentation reference files (e.g., AGENTSMD-REFERENCE.md, CLARITY-REFERENCE.md) contain numerous links to external technical resources, research papers, and documentation standards from well-known services and organizations, including Google Developers, GitHub, and the Linux Foundation. These are provided for research and educational purposes.
[INDIRECT_PROMPT_INJECTION]: The skill processes arbitrary user-provided markdown files through its linter script.
Ingestion points: The scripts/prose_lint.py script reads file content from paths provided as command-line arguments.
Boundary markers: None present; the linter operates directly on the raw file content.
Capability inventory: The linter script is limited to reading files and outputting text-based findings. The test script can execute the linter.
Sanitization: The linter logic uses regex and masking functions to ignore markdown structural elements (like code blocks and links) and focus analysis on prose text, which serves as a basic form of content filtering.

write-docs