Skip to content

README.md

Latest commit

 

History

History
59 lines (40 loc) · 2.79 KB

README.md

File metadata and controls

59 lines (40 loc) · 2.79 KB

Agentic Workflows Workshop

Welcome to the Agentic Workflows workshop! In this hands-on session you will use GitHub Copilot's agentic workflow capability (gh aw) to automate real-world tasks directly inside your GitHub repository — no servers, no pipelines to maintain.

Workshop Overview

Agentic Workflows let you describe what you want done in plain English. Copilot figures out how to do it, runs the necessary tools, and posts the results back to GitHub. In this workshop you will bootstrap the tooling, build two practical workflows, and wire up a ChatOps slash command — all from the command line.

Learning Objectives

By the end of this workshop, you will be able to:

  1. Bootstrap Agentic Workflows — Install the gh aw extension and initialise it in a repository, and set up a daily digest for issues and pull requests.
  2. Build a Hacker News digest — Create a daily workflow that surfaces relevant Hacker News stories as GitHub issues.
  3. Use the ChatOps pattern — Implement a slash command that runs sentiment analysis on Hacker News story comments and replies inline.

Prerequisites

Before attending this workshop, please ensure you have:

  • A GitHub account with an active Copilot Pro, Pro+, Business, or Enterprise subscription
  • Git installed and configured
  • GitHub CLI (gh) installed and authenticated (gh auth login)
  • Basic comfort with a terminal

Note

If you are using Copilot Business or Copilot Enterprise, ensure your organisation admin has enabled Copilot Extensions and Agentic Workflows in the policy settings.

Workshop Exercises

Exercise Topic Duration
0. Prerequisites Setup & tooling 10 min
1. Quick Start gh aw init + daily issue/PR digest 20 min
2. Hacker News Digest Custom agentic workflow for HN stories 20 min
3. ChatOps Sentiment Analysis Slash command with gh aw ChatOps pattern 20 min
4. Review & Next Steps Recap and further reading 5 min

Tips for Success

  1. Be specific in your prompts — the more context you give, the better the agent performs.
  2. Read the generated workflow file — understanding what the agent writes helps you tune it.
  3. Iterate — if the first result isn't quite right, add constraints and re-run.
  4. Check the issue/PR it creates — agentic workflows post their output to GitHub, so watch for new issues.

Support

  • During the workshop: Raise your hand or use the chat to ask questions.
  • After the workshop: Open an issue in this repository.

Happy automating with GitHub Copilot! 🤖🚀