ollama-web

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md "Web Search" and "Web Fetch" sections instruct the agent to call Ollama Cloud's web_search and web_fetch APIs with user queries or arbitrary URLs, which return content from public websites and links that the agent is expected to read and use, exposing it to untrusted third‑party web content.