tavily

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

• Third-party content exposure detected (high risk: 0.90). The skill explicitly performs web searches and content extraction from public sites via the Tavily API (see SKILL.md and scripts/search.mjs & scripts/extract.mjs), returning/exposing AI-optimized snippets and full extracted page text that the agent is instructed to read and use for summarization/answers, so untrusted third-party content could inject actionable instructions.

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

• Potentially malicious external URL detected (high risk: 1.00). The skill's runtime scripts perform POST fetches to https://api.tavily.com/search and https://api.tavily.com/extract (requiring TAVILY_API_KEY) and consume the returned "answer" / "content" fields which are injected into the agent's output/context, meaning remote content directly controls prompts/response generation.