ai-model-nodejs

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The prompt shows and instructs embedding user-supplied secrets directly into code and API payloads (e.g., secretId/secretKey in tcb.init and Secret.ApiKey in CreateAIModel), which would require the agent to place secret values verbatim into generated requests or code.

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

• Third-party content exposure detected (high risk: 0.80). The SKILL.md's "Custom onboarding" workflow explicitly allows registering arbitrary external BaseUrl endpoints via CreateAIModel (e.g., an OpenAI-compatible URL) and then calling those custom models (ai.createModel + generateText/generateImage), which means the agent will ingest and act on untrusted third-party model responses—creating a clear avenue for indirect prompt injection.