tavily-crawl

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

• Third-party content exposure detected (high risk: 0.95). The skill’s runtime calls Tavily’s POST https://api.tavily.com/crawl with a user-supplied url, and Tavily returns scraped page raw_content (outsider-authored public web text) that is then ingested into the agent’s LLM context.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill issues runtime POSTs to https://api.tavily.com/crawl (and https://api.tavily.com/map) to fetch external webpage content that is intended to be injected into LLM context for agentic workflows, so remotely-hosted content can directly influence prompts and agent behavior.