The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple shell commands to manage the development environment and project state.
Uses git worktree add to create new workspaces and git check-ignore to verify directory exclusion.
Automatically triggers package managers based on detected manifest files: npm install, cargo build, pip install, poetry install, and go mod download.
Executes project test runners including npm test, cargo test, pytest, and go test ./... to verify the baseline code state.
Modifies and commits changes to the .gitignore file if a local worktree directory is not already ignored.
[INDIRECT_PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests and processes untrusted data from the repository.
Ingestion points: Reads environment preferences from CLAUDE.md (specifically searching for worktree.*director) and analyzes project manifests like package.json and Cargo.toml (SKILL.md).
Boundary markers: Absent; the skill does not use specific delimiters or instructions to ignore potential commands embedded in manifest files or CLAUDE.md.
Capability inventory: Possesses significant capabilities including file writes, directory creation, dependency installation, and arbitrary command execution via test runners (SKILL.md).
Sanitization: Lacks explicit sanitization or validation for strings extracted from CLAUDE.md or the results of project setup commands.

using-git-worktrees