company-research

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

• Third-party content exposure detected (high risk: 1.00). The skill's Research Flow and individual steps (e.g., "Phase 1 — Verification", Steps 2–8, Source Priority Reference and Step 12) explicitly require web searches and ingestion of open/public third‑party content — MCA filings, LinkedIn, Tracxn/Crunchbase, Google Business/app store reviews and social media — which the agent is expected to read and use to drive KServe Fit, ICP scoring, and next actions, creating a clear surface for indirect prompt injection.