eda-research

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

• Third-party content exposure detected (high risk: 0.85). The skill explicitly searches and fetches open web vendor content (e.g., "Search JLC" via mcp__jlc__component_search and mcp__jlc__library_fetch) and downloads external datasheets (datasheets/*.pdf) which the agent is required to read and use to make selection/fix decisions, exposing it to untrusted third-party content.