product-analysis

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly instructs gathering substitution evidence from public, user-generated sources (e.g., review sites like G2/Capterra, Reddit/forums, comparison blog posts and competitor websites in the "Substitution Evidence Gathering" and related sections) and even suggests spawning a "Competitor research" subagent to collect and interpret that material, so the agent is expected to read untrusted third‑party content as part of its workflow.