doc

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

• Third-party content exposure detected (high risk: 0.90). The skill instructs the agent to fetch arbitrary external documentation via the Jina.ai reader (https://r.jina.ai/) into $PROJECT/docs/external and then explicitly tells the agent to read/reference those cached docs (e.g., @docs/external/.../README.md), so untrusted public webpages can directly influence its decisions and actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly fetches documentation at runtime using the Jina.ai reader (https://r.jina.ai/), saves that content, and injects/references it into the agent context (e.g., @docs/external/...), so remote content can directly control prompts.