The Agent Skills Directory

[COMMAND_EXECUTION] (LOW): The skill automatically executes environment setup and test baseline commands (e.g., npm install, npm test, pip install, cargo build) based on files detected in the project root. This is a standard developer workflow, but it allows a repository to trigger arbitrary command execution via post-install hooks or test scripts. Severity is reduced to LOW as this is the primary purpose of the skill.\n- [EXTERNAL_DOWNLOADS] (LOW): Triggers external package downloads through official package managers. This behavior is expected for workspace setup but involves interacting with external registries based on repository-defined dependencies.\n- [PROMPT_INJECTION] (LOW): The skill is susceptible to indirect prompt injection via the CLAUDE.md file, which it uses to determine configuration preferences without validation.\n
Ingestion points: The CLAUDE.md file is read via grep to find directory preferences.\n
Boundary markers: None; information retrieved from the repository files is used directly in shell commands.\n
Capability inventory: The skill can execute git worktree, npm, cargo, pip, poetry, and go commands, and can modify the repository's .gitignore and commit changes.\n
Sanitization: No sanitization or escaping is performed on the data extracted from CLAUDE.md or branch names before they are interpolated into shell commands (e.g., path="$LOCATION/$BRANCH_NAME").

git-worktrees