relay

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

• Third-party content exposure detected (high risk: 0.85). The skill explicitly ingests user-generated GitHub issue content as a required Task AC source (see "gh issue view " in Step 1) and then uses that content to build dispatch prompts, rubrics, and drive implementation/review actions, so untrusted third-party instructions could materially influence agent behavior.