using-git-worktrees

Warn

Audited by Gen Agent Trust Hub on Feb 22, 2026

Risk Level: MEDIUMEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • EXTERNAL_DOWNLOADS (MEDIUM): The skill automatically invokes package managers (npm install, pip install, poetry install, go mod download) based on the presence of project manifest files. This behavior can be exploited to perform remote code execution if a project contains a malicious manifest targeting public registries.
  • COMMAND_EXECUTION (MEDIUM): The skill executes test suites (npm test, cargo test, pytest, go test) to verify a clean workspace baseline. This involves executing arbitrary code located within the repository's test files, which is a significant risk when working on untrusted codebases or branches.
  • PROMPT_INJECTION (LOW): The skill demonstrates an indirect prompt injection surface by reading configuration preferences from CLAUDE.md and repository manifests without adequate sanitization or boundary enforcement.
  • Ingestion points: CLAUDE.md (read via grep), package.json, Cargo.toml, requirements.txt, pyproject.toml, go.mod.
  • Boundary markers: Absent; the skill does not utilize delimiters or specific warnings to ignore embedded instructions within these data sources.
  • Capability inventory: File system modification (git worktree add, git commit), network access (via package managers), and arbitrary shell execution (via test runners).
  • Sanitization: Absent; paths and configuration values derived from repository files are used directly in shell commands.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 22, 2026, 04:54 AM