market-research

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs Layer 1 agents (see agents/competitor-agent.md and agents/consumer-landscape-agent.md) and the SKILL.md "Research Tool Priority" to fetch/scrape public third‑party sources (pricing pages, G2/Capterra, Crunchbase, Reddit, forums, arbitrary web search results via WebSearch/Firecrawl/Defuddle), and those untrusted user-generated and public sources are read and synthesized to drive gap identification, opportunity ranking, and downstream actions—meeting the criteria for indirect prompt injection risk.