chore: add Groq Qwen3.6 preset

[c121914yu]

Summary

• Add Groq qwen/qwen3.6-27b LLM preset.
• Set the preset from Groq official model data: 131,072 context, 32,768 max output tokens, vision enabled, reasoning enabled, and tool choice enabled.

Evidence

• Groq model list: https://console.groq.com/docs/models
• Groq Qwen3.6 model card: https://console.groq.com/docs/model/qwen/qwen3.6-27b

Validation

• node .codex/skills/model-provider-updater/scripts/model_provider_presets.mjs inventory
• bun run test passed: 39 files / 246 tests. Existing warnings: missing optional .env.test / .env.test.local, and Vitest coverage package version mismatch.
• bunx tsx -e "...ProviderConfigSchema.parse(groq)..." passed for the Groq provider data.
• bun tsc --noEmit currently fails on existing sdk/factory/src/tool-factory.test.ts(285,9) stream mock typing; this is unrelated to the static Groq preset change.

[Copilot]

Pull request overview

Adds a new Groq LLM preset entry for qwen/qwen3.6-27b to the provider’s static model inventory so it can be selected/configured consistently alongside existing Groq presets.

Changes:

• Add Groq preset for qwen/qwen3.6-27b with maxContext: 131072 and maxTokens: 32768.
• Mark the model as vision: true, reasoning: true, reasoningEffort: true, and toolChoice: true, matching the Groq provider config shape used elsewhere in the file.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.