dcf-model

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md explicitly instructs the agent to "Fetch data from MCP servers, user provided data, and the web" (see "Step 1: Data Retrieval and Validation" and "Available Data Sources" / "Web search/fetch"), so the skill ingests open/public third‑party web content that is untrusted and that directly feeds the DCF calculations and model decisions.