model-copilot

What problem does it solve?

It eliminates the friction of sending prompts from your workflow to GitHub Copilot-capable models, so you can quickly get high-quality responses without manually switching between providers or model endpoints.

Core Features & Use Cases

• Model routing for GitHub Copilot: Sends your prompt to selectable models (defaulting to gpt-4o) to match the job type (coding, reasoning, summarization).
• Copilot authentication via token: Uses GITHUB_COPILOT_TOKEN from ~/Jit/.env as a Bearer token for authorized access through the GitHub Models Marketplace API.
• Operational commands and visibility: Supports listing available models, checking status (token + connection), and returning results with attribution to the chosen model; also logs telemetry to token_log.jsonl.

Example use case: when you need a code review, you can route to claude-3.5-sonnet via Copilot; when you need deeper reasoning about logic or bug-hunting, you can route to o3-mini; when you need a fast summary, you can route to gpt-4o-mini.

Quick Start

Ask for a model-specific response by running: /model-copilot "Explain the architecture of the มนุษย์ Agent system".