using-git-worktrees

Pass

Audited by Gen Agent Trust Hub on Sep 15, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted repository files, presenting an indirect prompt injection and execution surface if the target workspace is malicious.
  • Ingestion points: Evaluates local workspace configuration files including CLAUDE.md, package.json, Cargo.toml, requirements.txt, pyproject.toml, and go.mod in SKILL.md.
  • Boundary markers: Present. Explicitly includes an untrusted content boundary definition specifying the use of <EXTERNAL_DATA> delimiters in SKILL.md.
  • Capability inventory: Triggers execution of package management tools and test runners (npm install, cargo build, pip install, poetry install, go mod download, npm test, cargo test, pytest, go test) in SKILL.md.
  • Sanitization: Absent. Relies on the user or calling agent to ensure the workspace repository itself is trusted before invoking automated installation and testing routines.
  • [COMMAND_EXECUTION]: The skill executes multiple system commands to set up the development environment.
  • Evidence: Invokes commands such as git worktree add, git check-ignore, and various package manager setup/test suites in SKILL.md. These actions are well-aligned with the primary utility purpose of the skill.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 15, 2026, 09:01 AM
Security Audit — agent-trust-hub — using-git-worktrees