project-checklist
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is instructed to read and follow project-specific rules defined in files like AGENTS.md and CLAUDE.md within the target repository. This creates a surface where instructions within the audited files could influence the agent's audit logic.
- Ingestion points: usecases/check.md (Step 2: Read target AGENTS.md and CLAUDE.md).
- Boundary markers: Not specified.
- Capability inventory: Execution of gh (GitHub CLI) for repository management and vp (Vite+ verify) for local project execution.
- Sanitization: The skill does not mention sanitizing or escaping the contents of these files before interpreting them as instructions.
- [COMMAND_EXECUTION]: The skill uses the gh tool to perform administrative actions on GitHub repositories, such as setting secrets, modifying workflow permissions, and updating branch protection rules. These commands are executed as part of the repository setup process.
- [EXTERNAL_DOWNLOADS]: The skill clones a template repository (masseater/typescript-template) from the author's GitHub account to facilitate project verification and setup. Evidence: usecases/check.md utilizes ghq get or gh repo clone for this purpose.
Audit Metadata