tavily

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

• Third-party content exposure detected (high risk: 0.90). The skill calls Tavily's live web search API (see "How to Use" and the POST to https://api.tavily.com/search in SKILL.md) and returns/searches public webpages (optionally including raw page content and an "answer" + results array) that the agent is instructed to read and use in downstream LLM/agent workflows, exposing it to untrusted third-party content that could carry indirect prompt injection.