ad-angle-miner

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

• Third-party content exposure detected (high risk: 0.95). Yes—this skill’s required workflow explicitly instructs mining outsider-authored free text from public sources (e.g., G2/Capterra/Trustpilot reviews, Reddit threads, Twitter/X posts, and competitor ads) via the upstream scrapers, and then uses that collected text as “Proof quotes” in the LLM’s Phase 2/3 extraction and ranking context.