The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/investigate.py executes local git commands via subprocess.run using argument lists, which is a secure pattern that prevents shell injection.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface. Ingestion points: scripts/investigate.py reads data from git log outputs and ADR files in .agents/architecture/. Boundary markers: The generated report does not use specific delimiters to separate external content from instructions. Capability inventory: Uses subprocess.run for git commands and Path.read_text() for file access. Sanitization: Path traversal validation is present, but content from commit messages and files is not sanitized before inclusion in the final report.
[SAFE]: The implementation includes a validate_path_no_traversal function that explicitly protects against directory traversal attacks (CWE-22) by ensuring target paths resolve within the current working directory.

chestertons-fence