The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/changed_lines.py executes git commands using subprocess.run to retrieve diff data.\n
Evidence: The run_git and git_output functions wrap subprocess.run to call git subcommands.\n
Security Context: The implementation uses argument lists rather than shell strings, preventing shell injection vectors.\n- [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface when processing untrusted diff data.\n
Ingestion points: scripts/changed_lines.py ingests untrusted code content via git diff.\n
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore instructions embedded in hunks.\n
Capability inventory: The skill script executes subprocess.run for git-specific operations.\n
Sanitization: Code content from the diff is passed directly to the agent context without sanitization.

branch-risk-audit