implement
Pass
Audited by Gen Agent Trust Hub on Nov 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input from repository files such as spec.md, AGENTS.md, and CLAUDE.md. These inputs drive high-privilege agent actions including code generation, test execution, and pull request management. This creates a vulnerability surface where malicious content in a specification file could influence the agent's behavior. * Ingestion points: docs/specs/SLUG/spec.md, AGENTS.md, CLAUDE.md. * Boundary markers: Absent. * Capability inventory: git commit, npx (skill installation), pr (pull request creation), babysit-pr (CI bot interaction), tool execution (tdd, llm-evals), and local server execution. * Sanitization: Absent.
- [EXTERNAL_DOWNLOADS]: The skill references and provides instructions for downloading external tools and skills using npx skills@latest add toy-crane/skills. These resources are hosted on NPM (a well-known service) and the author's GitHub repository.
- [COMMAND_EXECUTION]: The skill executes various system-level commands for version control (Git), package management (npx), and local server management to facilitate the product handoff.
Audit Metadata