The Agent Skills Directory

[SAFE]: The skill implements a 'Safety First' policy requiring explicit user confirmation at two distinct gates before any destructive actions (branch or worktree deletion) are taken.- [COMMAND_EXECUTION]: The skill uses the Bash tool to execute git commands (e.g., git branch, git log, git fetch). These operations are restricted to local repository management and are intended for the skill's primary purpose.- [INDIRECT_PROMPT_INJECTION]: The skill processes external data in the form of git branch names and commit messages.
Ingestion points: Branch names and logs are read via git branch and git log commands in SKILL.md (Phase 1, 2, and 3).
Boundary markers: The instructions explicitly mention the need for user confirmation gates to review the analysis before proceeding.
Capability inventory: The skill has the ability to delete branches (git branch -d/-D) and remove worktrees (git worktree remove) via the Bash tool.
Sanitization: The skill includes a specific security instruction to always quote "$branch" in commands to prevent characters in branch names from breaking shell expansion or causing command injection.

git-cleanup