pr-writing-review

Warn

Audited by Gen Agent Trust Hub on Feb 16, 2026

Risk Level: MEDIUMPROMPT_INJECTION
Full Analysis
  • PROMPT_INJECTION (MEDIUM): The skill is susceptible to Indirect Prompt Injection (Category 8) because it ingests and analyzes untrusted data.
  • Ingestion points: The script scripts/extract_pr_reviews.py fetches data (comments, code suggestions, and file evolution) from user-provided GitHub PR URLs.
  • Capability inventory: Extracted content is provided to the LLM for 'Pattern recognition', 'Paragraph comparison', and 'style lesson synthesis'.
  • Boundary markers: The skill documentation does not mention the use of delimiters or specific boundary markers to isolate untrusted PR content from the agent's system instructions.
  • Sanitization: There is no evidence of sanitization, filtering, or validation performed on the external content before it is processed by the AI.
  • Risk: A malicious PR could contain instructions (e.g., in a code suggestion or feedback comment) designed to trick the agent into ignoring its constraints or performing unauthorized actions while the 'analysis' is being conducted.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Feb 16, 2026, 12:46 PM