feat(ai): set the default temperature to 0 and allow users to configure it

[yuyutaotao]

Motivation

• Expose per-intent sampling temperature so callers can control model sampling via environment or agent config.
• Preserve deterministic behavior for vlm-ui-tars by defaulting its temperature to 0.0 unless overridden.
• Improve observability by surfacing the resolved temperature in profile statistics logs.

Description

• Add environment keys and constants for MIDSCENE_MODEL_TEMPERATURE, MIDSCENE_INSIGHT_MODEL_TEMPERATURE, and MIDSCENE_PLANNING_MODEL_TEMPERATURE and wire them into MODEL_ENV_KEYS and related config structures.
• Extend environment types and IModelConfig with a temperature field and parse it in parseOpenaiSdkConfig to include temperature in the resolved model config.
• Use the resolved modelConfig.temperature in callAI (falling back to 0.0 for vlm-ui-tars) and include it in commonConfig sent to the SDK.
• Append temperature to the comma-separated debugProfileStats output in both streaming and non-streaming code paths, and update English/Chinese model-config docs to document the new env variables.

Testing

• No automated tests were run for this change.

---

Codex Task

[netlify]

✅ Deploy Preview for midscene ready!

Name	Link
🔨 Latest commit	b4a674d
🔍 Latest deploy log	https://app.netlify.com/projects/midscene/deploys/694bb8d6d4f82e0008bc412a
😎 Deploy Preview	https://deploy-preview-1654--midscene.netlify.app
📱 Preview on mobile	Toggle QR Code... Use your smartphone camera to open QR code link.

---

To edit notification comments on pull requests, go to your Netlify project configuration.