validating-startup-ideas

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md "Finding problems" section explicitly instructs mining complaints from public, user-generated sources (Reddit, G2, Capterra, Upwork job posts, App Store reviews), which the agent is expected to read and use to validate ideas, exposing it to untrusted third-party content that could carry indirect prompt injection.