worktree
Pass
Audited by Gen Agent Trust Hub on Sep 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill facilitates git repository management by executing multiple shell commands, including git and a custom tool named doctrine. These commands perform operations like creating worktrees, checking repository status, and verifying build baselines.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests external configuration from the CLAUDE.md file and repository metadata (such as worktree structures and git status) to determine its operational logic. This represents a vulnerability surface where a malicious repository could influence the agent's behavior or workspace configuration.
- Ingestion points: CLAUDE.md file, git repository state, and project-provided verification commands.
- Boundary markers: None specified in the instructions.
- Capability inventory: Shell command execution, directory creation/removal, and modification of the .gitignore file.
- Sanitization: No explicit validation or sanitization of project-provided inputs or CLAUDE.md contents is documented.
Audit Metadata