adr-review

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

• Third-party content exposure detected (high risk: 0.70). The skill's Phase 0 "Related Work Research" (references/debate-protocol.md) explicitly instructs agents to run gh issue/pr searches and include those open Issues/PRs findings in every agent prompt, meaning the agents fetch and interpret user-generated GitHub content (untrusted third-party material) that can materially influence decisions.