webextrator

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

• Third-party content exposure detected (high risk: 0.95). The required runtime workflow calls POST /webextrator/extract with a user-supplied url, and the service renders and ingests the fetched page’s readable content (including arbitrary third-party text) into its schema.org/LLM extraction pipeline—i.e., outsider-authored free text from public web pages.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill issues runtime POSTs to https://api.acedata.cloud/webextrator/extract (and related /render) with a user-supplied "url" (e.g., "https://example.com"), causing the service to fetch arbitrary external pages and return their text/markdown which can be injected into an LLM prompt and thus directly control agent behavior.