load-pr-comments
Pass
Audited by Gen Agent Trust Hub on Jun 20, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill employs standard GitHub CLI (
gh) commands and basic shell utilities such asmkdir,grep, andprintfto interact with the GitHub API and manage local task files. These operations are transparent and consistent with the skill's documented functionality for managing PR feedback. - [PROMPT_INJECTION]: The skill presents a surface for indirect prompt injection as it processes external GitHub PR comments at runtime.
- Ingestion points: External data enters the context through PR comment bodies retrieved from GitHub via GraphQL or MCP tools (SKILL.md).
- Boundary markers: The skill rewrites comments into markdown task files using standard headers and checklists but does not include explicit protective delimiters or 'ignore embedded instructions' warnings for the external content.
- Capability inventory: The skill's inherent capabilities are limited to file system management and API reads. However, the generated task files are designed for consumption by downstream agents which may have broader capabilities.
- Sanitization: Instructions specify preserving the substance of code suggestions and feedback verbatim to ensure accuracy, which means no sanitization or filtering of the original comment text is performed.
Audit Metadata