The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute a variety of shell commands for repository management and project setup. This includes Git operations (git worktree add, git rev-parse, git check-ignore), project-specific dependency installation (npm install, cargo build, pip install, poetry install, go mod download), and testing suites (npm test, cargo test, pytest, go test). These commands are standard for development workflows and are used within the context of the user's project.
[INDIRECT_PROMPT_INJECTION]: The skill includes instructions to read configuration preferences from CLAUDE.md and project manifests like package.json or Cargo.toml.
Ingestion points: Reads project-local files including CLAUDE.md, package.json, Cargo.toml, requirements.txt, pyproject.toml, and go.mod.
Boundary markers: The agent is instructed to use specific grep patterns (e.g., grep -i "worktree.*director") to filter content from CLAUDE.md.
Capability inventory: The skill has capabilities to write to .gitignore, commit changes to the repository, and execute arbitrary build/test commands based on the detected project type.
Sanitization: There is no explicit sanitization of the content extracted from CLAUDE.md before it is used to determine directory paths, which represents a standard but notable surface for influencing agent behavior via untrusted repository files.
[SAFE]: The skill explicitly incorporates safety best practices, such as verifying that new worktree directories are ignored by Git to prevent accidental disclosure or tracking of worktree-local files. It also enforces a 'fix broken things immediately' policy by ensuring the environment is correctly configured before proceeding.

using-git-worktrees