fix: default Terra reasoning to medium

[Honguan]

Summary

• change the default --terra-reasoning value from high to medium
• keep generated Terra agent expectations and tests aligned
• update the Chinese and English configuration examples

Validation

• npm test (109 passed)
• npm run check
• npm run test:package

The primary-model --set-default behavior remains unchanged.

[chatgpt-codex-connector]

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: 3d66340050

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".

[chatgpt-codex-connector]

Keep the default profile out of hierarchy warnings

With a fresh install and no reasoning flags, this changes Terra to medium while Luna remains xhigh; reasoningWarnings() treats that two-rank gap as a reversed plan/execute hierarchy, so every default installation—and the newly documented example—now emits warning: Luna is materially stronger than Terra. If this is the supported default profile, update the warning policy alongside it so users are not told that the package's own defaults are misconfigured.

Useful? React with 👍 / 👎.