idea-validation

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md workflow explicitly tells the agent to "Search for Reddit threads, forum posts, review complaints" and to find competitors/1-star reviews from public sites (Problem Check / Competitor Check), meaning the agent must fetch and interpret untrusted, user-generated web content that can materially influence its validation verdict.