The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/simplify_scope.py executes the git command using subprocess.run. Arguments are passed as a list, which prevents shell-based command injection. While user input can influence flags passed to git diff via the --ref parameter, the impact is confined to the functionality of the git binary itself.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because file paths from the repository are interpolated directly into the agent's instructions without sanitization.
Ingestion points: scripts/simplify_scope.py captures file paths from git diff output.
Boundary markers: Absent. File names are listed under a markdown header in the generated prompt without isolation or escaping.
Capability inventory: The agent is authorized to read and write files and execute project-specific tests (SKILL.md).
Sanitization: Absent. The script does not validate or escape file paths to prevent them from containing prompt injection payloads.

simplify