What problem does it solve? Developers who want to embed GitHub Copilot's agentic capabilities into their own applications lack a structured way to manage sessions, tools, and model providers programmatically. This Skill provides the patterns and API references needed to integrate Copilot into Node.js, Python, Go, or .NET applications without reverse-engineering the CLI. ## Core Features & Use Cases - Session Management: Create, resume, and persist sessions with custom IDs, streaming responses, and infinite-session auto-compaction for long-running workflows. - Extensibility: Define custom tools, lifecycle hooks (pre/post tool use, prompt submission, error handling), custom agents, and connect MCP servers over stdio or HTTP. - Flexible Authentication & BYOK: Use GitHub tokens, OAuth apps, or bring your own keys for OpenAI, Azure OpenAI, Anthropic, or local Ollama providers without a Copilot subscription. - Use Case: Build a multi-tenant SaaS backend that runs a headless Copilot CLI server over TCP, isolates sessions per tenant, enforces tool permissions via hooks, and tracks token usage through events. ## Quick Start Ask the agent to create a Node.js script using the Copilot SDK that starts a session with the gpt-4.1 model, registers a custom weather tool, and prints the streamed response.