pr
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE]: The skill consists entirely of Markdown templates, YAML configuration, and metadata files (LICENSE, CREDITS, PROVENANCE). It contains no scripts, binaries, or active execution components.
- [SAFE]: The content is limited to providing structural guidance for PR descriptions, including examples of Mermaid diagrams, pseudocode, and diff formats. All referenced URLs in the metadata point to legitimate public repositories for attribution purposes (e.g., Humanlayer, Matt Pocock's skills repo).
- [INDIRECT_PROMPT_INJECTION]: While the skill processes user data (domain language from a hypothetical
GLOSSARY.mdand technical changes) to generate a PR body, it has no active capabilities (network, file writes, command execution) to exploit. The risk is negligible as it merely acts as a text formatter.
Audit Metadata