ai-ecommerce-expert-batch-sku-content-production

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

• Third-party content exposure detected (low risk: 0.10). 在脚本 scripts/imiva_mcp.py 运行时，LLM 通过 create_visual_migration_task 把用户在 referenceImage/productImages/subjectDescriptions 等字段中提供的自由文本（尤其是 subjectDescriptions）作为 MCP 工具入参发送给 @infimind/ecom-content-cli（https://imiva.ecpro.com），但该入参不是从外部由“第三方提交”的可监控队列/页面/票据等来源读取后再注入的运行时抓取文本。

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 fetch and run the remote npm package "@infimind/ecom-content-cli@latest" at runtime (e.g., via subprocess [npx, "-y", PACKAGE]), which executes remote code fetched from the npm registry.