Gives your agent the ability to define clear, unambiguous, and AI-ready specifications for schemas, tools, architecture, and more, ensuring a machine-readable project foundation.
When you need to define requirements and data contracts for a new system component
When you want to create a specification optimized for consumption by Generative AI
When you need to document architectural dependencies and acceptance criteria in a standard way
Identify the purpose of your spec, such as schema, tool, or infrastructure
Define the core requirements, constraints, and guidelines using precise language
Outline the interfaces, data contracts, and specific acceptance criteria (Given-When-Then)
Review and save the formatted spec file to your project's /spec/ directory
你
Create a technical spec for our new analytics tool.
Agent
I'll generate spec-tool-analytics.md in your /spec/ directory. I'll define clear REQ- codes for every requirement, set up data contracts for your event schemas, and include testable acceptance criteria so both humans and AI can verify the implementation.
赋予您的代理更新技术规范（包括需求、约束和针对 AI 使用优化的数据契约）的能力。
赋予你的智能体根据规范模板自动起草和管理 GitHub Pull Request (PR) 的能力，确保满足所有项目需求。
赋予你的智能体将技术规范文件转换为清晰、结构化的 GitHub Issue，以作为完整功能请求的能力。
赋予你的智能体根据规范审计代码库的能力，并自动为发现的任何未实现的需求创建 GitHub Issue。
赋予您的代理生成完整的 TypeSpec 声明式代理的能力，其中包含针对 Copilot 的指令和自定义功能。
