startup-validator

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

• Third-party content exposure detected (high risk: 0.85). The required workflow includes runtime web_fetch of full articles from third-party sites (e.g., Gartner/McKinsey/Statista/Crunchbase/industry reports) and then uses that fetched readable text in the agent’s analysis/reporting, which can include outsider-authored free text (indirect prompt injection risk).