using-git-worktrees

Pass

Audited by Gen Agent Trust Hub on Oct 26, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill executes multiple shell commands to manage git state and worktrees (e.g., git rev-parse, git worktree add, git branch). It also automates project-specific commands like npm test, cargo test, and pytest. These commands are necessary for the skill's primary function of workspace isolation and verification.\n- [EXTERNAL_DOWNLOADS]: The skill uses standard dependency managers (npm install, poetry install, pip install, go mod download) to set up the development environment. This involves downloading third-party packages from official registries based on the project's configuration files.\n- [INDIRECT_PROMPT_INJECTION]: The skill operates on and executes code from the local repository (untrusted data). This creates a vulnerability surface where malicious instructions could be embedded in build files (package.json, Cargo.toml) or test scripts which the agent executes automatically.\n
  • Ingestion points: Repository configuration files (package.json, Cargo.toml, requirements.txt, pyproject.toml, go.mod) and test scripts.\n
  • Boundary markers: None present to distinguish between benign configuration and malicious instructions embedded in scripts.\n
  • Capability inventory: Command execution via shell (subprocess calls for git, npm, pip, cargo, etc.).\n
  • Sanitization: No sanitization or verification of the repository content is performed before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 26, 2026, 06:32 PM