fix: honor openai.useChatCompletions in agent mode

[WinnCook]

Summary

• route daemon agent OpenAI models through the shared OpenAI resolver so sidepanel chat honors openai.useChatCompletions
• preserve custom OPENAI_BASE_URL behavior for OpenAI-compatible servers
• add daemon agent regression coverage for env, config, base URL, and streaming agent paths

Closes #155.

Validation

• pnpm -s check
• full Vitest suite passed during pnpm -s check (345 files passed, 26 skipped)

Notes

• kept the change scoped to the agent resolver and its tests only
• did not add new user-facing flags or settings
• no reporter follow-up comment posted in this pass

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: d46c63f000

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".