longitudinal-cross-sectional-research

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow (SKILL.md "2. Collect source material") and the prompt template in references/method.md explicitly instruct the agent to gather and cite public sources—including blogs, company pages, filings, and "community evidence for user sentiment"—which are untrusted, user-generated third-party web content the agent is expected to read and use to form its analysis.