qinghu-pod-pattern-variations

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.80). The skill instructs the agent to handle API keys by having the user obtain a token and pass it via command-line argument (qhkit config set --token <密钥>), which involves passing secret values through the agent context/output.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs the agent to download, build, and execute Node.js from nodejs.org and install packages from npmjs.com/npmmirror.com during runtime environment bootstrapping, which constitutes an unverifiable operational dependency.