This repository is a guided workshop for demonstrating GitHub Copilot across repository context, skills, MCP, custom agents, Copilot CLI, governance, workflow automation, and operations.
概览
This repository is a guided workshop for demonstrating GitHub Copilot across repository context, skills, MCP, custom agents, Copilot CLI, governance, workflow automation, and operations. The detailed lecture has moved from one giant README into chapter files under docs/workshop. This root README is now the landing page and navigation hub. The workshop follows one connected engineering story: - — how Copilot is guided by repository instructions, product intent, specifications, and shared planning context. - — how Copilot gains capabilities through packaged local skills and connected tools. - — prompt files, custom agents, handoffs, and subagents working together. - — plan mode, execution, review, research, session history, and execution surfaces. - — scoped context, progressive reveal, compression, and OpenTelemetry-backed measurement. - — code generation is not the end; add cross-model review with the Critic agent. - — repository automation after merge.
README
以下为原文 README 节选,完整内容请查看原文。
GitHub Copilot Demo
This repository is a guided workshop for demonstrating GitHub Copilot across repository context, skills, MCP, custom agents, Copilot CLI, governance, workflow automation, and operations.
The detailed lecture has moved from one giant README into chapter files under docs/workshop. This root README is now the landing page and navigation hub.
Whole story at a glance
The workshop follows one connected engineering story:
- Shape Copilot behavior with repository context — how Copilot is guided by repository instructions, product intent, specifications, and shared planning context.
- Skills and MCP — how Copilot gains capabilities through packaged local skills and connected tools.
- Plan and specialize work in VS Code — prompt files, custom agents, handoffs, and subagents working together.
- Continue execution in Copilot CLI — plan mode, execution, review, research, session history, and execution surfaces.
- Save tokens with context hygiene — scoped context, progressive reveal, compression, and OpenTelemetry-backed measurement.
- Govern delivery with review, security, and hooks — code generation is not the end; add cross-model review with the Critic agent.
- Add workflow agents in GitHub Actions — repository automation after merge.
- Operate with SRE agents — closing the loop with operational thinking and OpenTelemetry observability.
Workshop chapters
| Chapter | Topic |
|---|---|
| 01 | Repository context, AGENTS.md, PRD, specs, and Spaces |
| 02 | Skills and MCP |
| 03 | VS Code prompt files, custom agents, handoffs, and subagents |
| 04 | Copilot CLI execution, sessions, research, memory, and execution surfaces |
| 05 | Token efficiency, context hygiene, and measurement lab results |
| 06 | Governance, review, security, hooks, and Critic agent |
| 07 | Workflow agents in GitHub Actions |
| 08 | SRE agents and operations |
| 09 | Optional demos and extensions |
Token-efficiency benchmark highlight
推荐工具
换一个关键词,或者移除筛选条件。
安装
npx skillfish add tkubica12/gh-copilot-demo