read-doc

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

• Third-party content exposure detected (high risk: 0.80). The skill uses WebFetch to retrieve and navigate public documentation pages (the user-confirmed documentation URL and its subpages) and then extracts and summarizes that third-party web content as part of its workflow, exposing the agent to untrusted public web content.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill fetches the user-confirmed documentation URL (the "[url]" from Step 1) at runtime and injects the retrieved documentation into the model context to produce code-generation prompts, so the external documentation URL is a required runtime dependency that directly controls agent behavior.