ai-ecommerce-expert-product-background-replacement

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

• Third-party content exposure detected (medium risk: 0.30). 运行时在脚本 scripts/imiva_mcp.py 中，create_smart_refine_task 的工具调用参数（含 prompt 等用户/工作流输入的自由文本）会通过 MCP tools/call 直接发送给 @infimind/ecom-content-cli，因此会让外部自由文本进入所需工作流的可读上下文。

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

• Potentially malicious external URL detected (high risk: 0.90). The skill invokes npx to run the remote npm package "@infimind/ecom-content-cli@latest" at runtime (see scripts/imiva_mcp.py and config examples), which causes npm to fetch and execute remote code.