Add GroqCloud provider support

[00-Aakash-00]

Summary

Adds GroqCloud as a first-class LLM provider option in Dexter (OpenAI-compatible Chat Completions), including provider routing, model selection UI entries, and env/docs updates.

Changes

Provider registry + routing

• Added groq to the canonical provider registry (src/providers.ts) with:
  • modelPrefix: groq:
  • apiKeyEnvVar: GROQ_API_KEY
  • fastModel: groq:llama-3.1-8b-instant
• Implemented GroqCloud model factory in src/model/llm.ts using LangChain's ChatOpenAI against Groq's OpenAI-compatible base URL (https://api.groq.com/openai/v1).
  • Stored model IDs keep the groq: prefix for routing; the prefix is stripped before sending to Groq (groq:llama-... → llama-...).

CLI model selection

• Added GroqCloud preset model entries in src/components/ModelSelector.tsx (prefixed with groq:) so users can select Groq models from the built-in list.
• Updated model display-name normalization to also handle groq:-prefixed model IDs.

Environment + docs

• Added GROQ_API_KEY to env.example.
• Updated README.md prerequisites and .env example snippet to mention GroqCloud and the GROQ_API_KEY.
• Updated AGENTS.md provider list / env var list to include GroqCloud and added build/lint commands for local workflows.

Tooling scripts

• Added bun run build (bundles to dist/ with external packages) and bun run lint (alias to typecheck) in package.json.

How to test

• bun run build
• bun run lint
• bun run typecheck
• bun test

Notes

GroqCloud is wired via the OpenAI-compatible Chat Completions API (base URL https://api.groq.com/openai/v1), so it integrates cleanly via ChatOpenAI while keeping the provider routing model-prefix approach consistent with existing providers.