git-workflow

Pass

Audited by Gen Agent Trust Hub on Mar 8, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: A thorough audit of the 32 files in this skill found no evidence of malicious patterns, such as prompt injection, data exfiltration, or obfuscated code. The skill is purely instructional.
  • [COMMAND_EXECUTION]: The skill includes numerous examples of git, gh (GitHub CLI), and npm commands. These are standard tools used for their intended purpose of managing repository workflows and implementing development hooks (e.g., Husky).
  • [EXTERNAL_DOWNLOADS]: The skill references several external resources for documentation and tooling, including git-scm.com, conventionalcommits.org, and official GitHub documentation. All referenced domains are well-known technology sites or trusted organizations.
  • [PROMPT_INJECTION]: The skill provides guidelines for agents to analyze git history and pull requests (Category 8 surface). 1. Ingestion points: Untrusted data enters via git commit messages or PR text provided by the user for review. 2. Boundary markers: The skill defines structured output formats (e.g., [category] description) to separate analysis from content. 3. Capability inventory: The skill suggests command snippets to the user but does not contain internal mechanisms for arbitrary shell execution or network requests. 4. Sanitization: No explicit input sanitization is defined, but the instructions focus on objective formatting checks (e.g., conventional commit compliance) rather than executing embedded text.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 8, 2026, 03:05 PM