setup
Pass
Audited by Gen Agent Trust Hub on Apr 4, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses Git command-line tools (
git worktree,git fetch,git status) to manage development environments. These are standard operations within the scope of a repository management skill.- [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by relying on a file within the project repository to define agent behavior. - Ingestion points: Reads
AGENTS.mdfrom the repository root. - Boundary markers: Absent. The skill instructions explicitly treat the contents of the file as a 'behavior contract' governing downstream logic.
- Capability inventory: Git commands, file system checks, and branch management.
- Sanitization: Absent. There is no logic to validate or sanitize instructions contained within the
AGENTS.mdfile before execution.
Audit Metadata