GitHub Actions related posts
View All Tags
If you've modernized a CI/CD pipeline in the last couple of years, you've heard the gospel of OpenID Connect: stop hardcoding long-lived AWS IAM keys in your repository secrets. Your workflow requests a short-lived token, the cloud verifies the claims, you deploy. It's a real upgrade. It's also where most teams stop thinking, and that's the problem.
Your laptop is about to stop being where work happens. As we get more competent with agents, local compute becomes the bottleneck, and the industry is converging on a new kind of cloud-native compute built specifically for autonomous agents.
Just like you use Copilot in VS Code or the Coding Agent from an issue, you can run those same agentic workflows triggered on GitHub events, inside Actions. Safe, secure, and ephemeral.
Actions isn't just your CI/CD pipeline. It's the execution layer for much of the GitHub platform — and it's purpose-built for the agentic era.
The ROI bar for automation just hit the floor. AI didn't just make coding faster, it made building automations so cheap that NOT automating is the expensive choice.
Discussions around AI frequently revolve around this end goal of automation. Where AI is an autonomous software engineer, a note taker, or even your personal chauffeur.
In reality most people are using AI as a tool to improve aspects of our workflows rather than building full blown automated solutions.
So it's 12:51 on a Wednesday afternoon and I've got nothing in the fridge. I've been using AI for everything so why isn't AI making me lunch? Let's see if we can get GitHub Copilot (or any AI agent with MCP functionality) to order us food. Seems like a simple task.
I've been experimenting using AI coding agents in GitHub Actions to automate workflows, generate code, and enhance my development process.
I've created a comprehensive GitHub Actions cheat sheet inspired by the Learn X in Y Minutes format.