summarize-review
Pass
Audited by Gen Agent Trust Hub on Aug 2, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill employs dynamic command syntax to automatically detect the current repository and branch context. It also uses git and gh tools to perform its primary function of code analysis. These operations are transparent and consistent with the skill's documented purpose.
- [PROMPT_INJECTION]: The skill is designed to handle untrusted code from Pull Request authors. It addresses potential indirect prompt injection through several mechanisms:
- Ingestion points: External code is accessed via git show and PR data through gh api calls.
- Boundary markers: The skill requires the user to explicitly confirm the draft before any automated posting occurs.
- Capability inventory: The skill uses diffing and API reading tools, with the ability to post comments through the gh tool.
- Sanitization: It specifically directs the agent to use quoted heredocs for shell commands to prevent untrusted content from being interpreted as code.
Audit Metadata