pr-reviewer-summary
Pass
Audited by Gen Agent Trust Hub on Aug 26, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill facilitates the processing of untrusted external content, creating an indirect prompt injection surface.
- Ingestion points: The skill ingests data from Pull Request diffs, URLs, changed-file summaries, commit history, and ticket details as specified in SKILL.md and references/pr-body-workflow.md.
- Boundary markers: There are no instructions to use delimiters or explicit prompts to ignore instructions within the source text, increasing the likelihood that the agent may follow commands embedded in the review evidence.
- Capability inventory: The skill is capable of generating shell commands (e.g., test runners) and manual verification steps in the final output, which could be manipulated by an attacker to social-engineer a reviewer.
- Sanitization: The instructions do not define any validation or sanitization requirements for the processed text evidence.
Audit Metadata