Microsoft announced October 28 that it is releasing the multi-model harness powering GitHub Copilot and other Copilot products through Microsoft Foundry, its managed infrastructure service.
The harness is an orchestration layer that provides AI agents with instructions, context, execution environment and controls. Originally built by GitHub’s engineering team, the harness now powers Copilot Code, Copilot Studio, the GitHub Copilot app, the CLI and VS Code, according to Microsoft’s announcement.
Through Foundry, developers and enterprise customers can build agents that reason, use tools and complete long-running tasks across different AI models without switching the harness or underlying infrastructure. The harness supports open standards including MCP and agent skills, as well as agent plugins.
Microsoft says the harness is designed to be fast, token-efficient and predictable. The company handles ongoing maintenance and improvements, with benefits accruing to both Microsoft’s own products and customer applications built on the harness.
The standardization reflects Microsoft’s adoption of GitHub’s engineering work across its product line. Microsoft describes the harness as having the highest satisfaction among developers, its “most demanding audience.”