weekly-retro

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs the agent to fetch and read GitHub issues and repo data (see SKILL.md Phase 1 commands like "gh search issues --owner $YOUR_OWNER ..." and "gh issue list -R ...") and then uses those user-generated items to decide, create, and route new issues, so it ingests untrusted third-party content that can influence actions.