#19311: feat: add github-copilot gpt-5.3-codex with xhigh support (AI-assisted)

## Summary - add `gpt-5.3-codex` to GitHub Copilot default model IDs - set `reasoning: true` for GitHub Copilot Codex models in `buildCopilotModelDefinition` - allow `/think xhigh` for `github-copilot/gpt-5.3-codex` - add unit tests for Copilot model defaults/reasoning and thinking-level coverage - update GitHub Copilot provider docs (EN + zh-CN) Closes #19301 ## Testing - Degree of testing: fully tested (targeted) - Ran: `corepack pnpm vitest run src/auto-reply/thinking.test.ts src/providers/github-copilot-models.test.ts` - Ran: `corepack pnpm vitest run --config vitest.e2e.config.ts src/auto-reply/reply.directive.directive-behavior.accepts-thinking-xhigh-codex-models.e2e.test.ts` - Result: all targeted tests passed (26 tests across 3 files) ## AI Assistance - This PR was AI-assisted via Codex. <!-- greptile_comment --> <h3>Greptile Summary</h3> This PR adds `gpt-5.3-codex` to the GitHub Copilot provider's default model list, enables `reasoning: true` dynamically for all codex models via a regex check, and adds `github-copilot/gpt-5.3-codex` to the xhigh thinking allowlist. The core logic changes are clean and consistent with existing patterns. - Added `gpt-5.3-codex` to `DEFAULT_MODEL_IDS` and `XHIGH_MODEL_REFS` - Replaced hardcoded `reasoning: false` with a dynamic regex check (`/(?:^|-)codex(?:$|-)/i`) so all codex models get `reasoning: true` automatically - Updated error message in e2e test to reflect the new model in the xhigh hint - Added new unit tests for `buildCopilotModelDefinition` and `getDefaultCopilotModelIds` - `docs/zh-CN/providers/github-copilot.md` was manually edited but per repo guidelines it's a generated file; the `source_hash` in frontmatter is now stale, which may cause the i18n pipeline to overwrite these changes on the next run <h3>Confidence Score: 4/5</h3> - This PR is safe to merge; the core logic changes are correct and well-tested, with only minor style concerns. - The changes are straightforward — adding a model ID to two arrays and introducing a regex-based reasoning flag. Tests cover the new behavior. The only concern is the manually edited zh-CN generated file with a stale source_hash, which could be overwritten by the i18n pipeline. - docs/zh-CN/providers/github-copilot.md — manually edited generated file with stale source_hash _{Last reviewed commit: 3a2b038} <!-- greptile_other_comments_section --> _{(4/5) You can add custom instructions or style guidelines for the agent [here](https://app.greptile.com/review/github)!} **Context used:** - Context from `dashboard` - AGENTS.md ([source](https://app.greptile.com/review/custom-context?memory=0d0c8278-ef8e-4d6c-ab21-f5527e322f13)) <!-- /greptile_comment -->