The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a Python script scripts/worktree_convention.py to execute git commands via subprocess.run. The implementation avoids using a shell and passes arguments as a list to mitigate injection risks.\n- [INDIRECT_PROMPT_INJECTION]: The skill represents a surface for indirect prompt injection as it processes branch names and paths that may originate from untrusted sources.\n
Ingestion points: Branch names and repository paths passed as arguments to scripts/worktree_convention.py (e.g., via --branch and --repo-root).\n
Boundary markers: The skill instructions require the agent to use the helper script for planning, and the script uses subprocess.run with list-based arguments.\n
Capability inventory: The skill can modify the .gitignore file, write metadata to the filesystem, and execute git worktree operations.\n
Sanitization: The script includes sanitize_slug to remove path separators from branch names and validate_explicit_root to prevent path traversal and repository escapes.

git-worktree