ai-rag-pipeline

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

• Third-party content exposure detected (high risk: 0.90). The SKILL.md explicitly shows fetching and ingesting untrusted public web content (e.g., using tavily/search-assistant, exa/search and tavily/extract in the "Basic RAG Pipeline", "Multi-Source Research" and "URL Content Analysis" examples, including extracting from arbitrary URLs like "https://example.com/research-paper"), which are then injected into LLM prompts and can materially influence agent behavior.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.90). The Quick Start includes a direct remote install-and-execute command (curl -fsSL https://cli.inference.sh | sh) which fetches and runs code at install/runtime and the skill repeatedly relies on the resulting infsh CLI, so https://cli.inference.sh is a runtime external dependency that executes remote code.