worktrees
Pass
Audited by Gen Agent Trust Hub on Jul 26, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes
gitcommands for worktree creation, isolation detection, and repository configuration (e.g., modifying.gitignore). It also runs system package managers likenpm,pip,cargo, andgobased on detected lockfiles to prepare the workspace.\n- [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by reading and executing test commands directly from a repository file (docs/agents/project.md).\n - Ingestion points: The agent reads the
docs/agents/project.mdfile to determine the test execution command.\n - Boundary markers: No boundary markers or 'ignore embedded instructions' warnings are used when processing the documentation content.\n
- Capability inventory: The skill utilizes subprocess execution for
git, package management, and arbitrary test commands.\n - Sanitization: There is no evidence of command validation or sanitization before the test command is executed in the shell.
Audit Metadata