opensource-contributions
Pass
Audited by Gen Agent Trust Hub on Sep 2, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill's primary function is to read and process external repository data such as CONTRIBUTING.md and project templates. The instructions direct the agent to follow project-specific guidelines, which are untrusted sources. This represents a standard risk surface for agents interacting with external data, partially mitigated by the skill's explicit transparency and disclosure requirements.
- Ingestion points: references/phase-0-before-you-start.md and references/phase-2-pull-requests.md describe reading and adopting external project files into the agent's context.
- Boundary markers: The skill uses structured templates and a validation script, but lacks explicit prompt delimiters when interpolating external content.
- Capability inventory: The skill utilizes git, gh (GitHub CLI), and standard shell commands for repository management.
- Sanitization: The included script scripts/check-pr-template-compliance.py performs structural validation of markdown files but does not sanitize text content.
- [SAFE]: The included Python script scripts/check-pr-template-compliance.py uses standard library modules (re, sys) to perform safe text processing and structural validation without network or sensitive file access.
Audit Metadata