cloud-functions

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

• Third-party content exposure detected (medium risk: 0.30). The skill’s runtime workflow can read function execution logs and log details (which may include outsider-supplied request bodies) via queryFunctions(action="listFunctionLogs") and queryFunctions(action="getFunctionLogDetail") / callCloudApi “GetFunctionLogDetail”, with no requirement to pre-select a trusted log source item.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill contains TCR image URIs such as ccr.ccs.tencentyun.com/your-ns/demo-app:demo-app-001 (and :demo-app-002) that are used at deploy/runtime for CustomImage HTTP Functions to pull and run container images, which fetch and execute remote code.