ai-rag-pipeline

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

• Third-party content exposure detected (high risk: 0.90). SKILL.md explicitly shows fetching and ingesting open web content (e.g., running tavily/search-assistant, exa/search, and tavily/extract on arbitrary URLs in the "Basic RAG Pipeline", "Multi-Source Research", and "URL Content Analysis" examples) and then passes those untrusted results into LLM prompts for analysis and decision-making, so third-party content can materially influence agent behavior.