The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes shell commands for repository management and environment verification.\n
Shell commands such as git worktree add, git rev-parse, and git check-ignore are utilized for Git operations as described in SKILL.md.\n
It executes package manager commands including npm install, cargo build, pip install, and go mod download to configure the workspace in SKILL.md.\n
It runs test suites using commands like npm test, cargo test, and pytest in SKILL.md.\n- [EXTERNAL_DOWNLOADS]: The skill downloads dependencies from official package registries based on project configuration.\n
Fetches Node.js packages via npm install from the official registry.\n
Fetches Rust dependencies via cargo build from the official crates registry.\n
Fetches Python packages via pip install or poetry install from the official PyPI registry.\n
Fetches Go modules via go mod download from the official Go proxy.\n- [INDIRECT_PROMPT_INJECTION]: The skill reads project configuration files which could contain untrusted data.\n
Ingestion points: The agent reads the CLAUDE.md file and manifest files like package.json to determine directory preferences and project types in SKILL.md.\n
Boundary markers: No delimiters or instructions are present to prevent the agent from acting on malicious instructions hidden within project files.\n
Capability inventory: The skill can execute shell commands, install packages, and run tests as defined in SKILL.md.\n
Sanitization: There is no evidence of sanitization for strings extracted from CLAUDE.md before they are used to influence the agent's actions.

using-git-worktrees