blog-author

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

• Third-party content exposure detected (high risk: 0.80). The skill's required Stage 1 Research instructs the agent to "gather more context from multiple sources" when web/retrieval tools are available and explicitly allows using community posts or personal blogs as supporting context (SKILL.md, "Stage 1: Research" and source preference), meaning the agent will fetch and interpret open/public, user-generated content that can influence subsequent planning and writing.