idea-testing

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

• Third-party content exposure detected (high risk: 0.90). The skill's required workflow and the "Discovery Experiments" (references/discovery-experiments.md) explicitly direct using and reading open/public third‑party sources—e.g., Google Trends/Keyword Planner and public discussion forums (Facebook groups, Reddit, Jumia/Jiji reviews)—and instruct the agent/user to interpret those untrusted, user‑generated contents to drive experiment selection and Persevere/Pivot/Kill decisions, which could allow indirect prompt injection.