vision-analysis

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

• Third-party content exposure detected (high risk: 0.75). The required workflow ingests an image from a user-provided file path or URL (public web content or downloaded/forwarded attachments authored by outsiders), and the agent then calls the MiniMax_understand_image MCP tool which reads that image content as LLM-readable text for analysis.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs the agent at runtime to fetch and follow remote setup instructions from https://platform.minimaxi.com/docs/token-plan/mcp-guide, so external content could directly control configuration commands or agent behavior.