reviewgate-loop

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 0.75). Outsider free text can enter the LLM context via gh api --paginate .../issues/$PR_NUMBER/comments and .../pulls/$PR_NUMBER/comments, where the skill reads untrusted comment body text (including <!-- reviewgate-finding: and <!-- reviewgate-summary -->) from other authors into the agent’s runtime context.