antv-x6-editor

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

• Third-party content exposure detected (medium risk: 0.30). Runtime path uses a context retrieval service (GET https://sive.antv.antgroup.com/api/v1/context/retrieve) where the user-provided query can cause the workflow to ingest retrieved Markdown content from search results, so outsider-authored free text can be returned and ingested based on submitted keywords.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill explicitly performs runtime GET requests to retrieve context from https://sive.antv.antgroup.com/api/v1/context/retrieve?query=flowchart+stencil+port&library=x6 (shown in the curl example), and that retrieved content is injected as context for the skill so it can directly control prompts.