The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/check_doc_staleness.py executes local git commands using subprocess.run. This is used to analyze repository churn and track file changes to identify stale documentation. The implementation uses a list-based argument approach without a shell, mitigating command injection risks.
[INDIRECT_PROMPT_INJECTION]: The skill processes local Markdown files and codebase files to verify documentation accuracy. This represents a potential surface for indirect prompt injection if the files contain malicious instructions. However, the skill implements path filtering to prevent access to sensitive system files (e.g., blocking absolute paths and shell expansion characters) and limits analysis to the repository scope.
Ingestion points: scripts/check_doc_staleness.py reads Markdown files (.md) and cited code files within the repository.
Boundary markers: The skill encourages the use of specific path/to/file:line citations to verify claims.
Capability inventory: Uses subprocess.run for git history and Path.open for file content verification.
Sanitization: Implements path validation in parse_citations to exclude absolute paths, home directory references, and shell expansion characters.

update-docs