jina

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

• Third-party content exposure detected (high risk: 1.00). The SKILL.md explicitly instructs the agent to fetch and convert arbitrary web pages and perform real-time web searches (e.g., "x jina reader https://example.com/article" and "x jina search ..."), meaning the agent will read untrusted public web content that can influence its decisions and tool use.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill fetches arbitrary web pages at runtime and converts them into Markdown for the model to read (e.g., "x jina reader https://example.com/article"), so remote content from external URLs can directly control prompts provided to the agent.