If youāve been following me for a while, you may have seen my earlier post about GitHub Agentic Workflows. That got me wondering: can I do something similar from an Azure DevOps pipeline?
The short answer is yes, although the plumbing is different. GitHub Agentic Workflows compiles a Markdown definition into a guarded GitHub Actions workflow. Azure DevOps does not have that same compiler. Instead, an Azure Pipeline can install GitHub Copilot CLI, authenticate it non-interactively, and run a prompt using the CLIās programmatic mode.
Letās be honest - we all have a bunch of Azure Automation runbooks sitting in our subscriptions that we wrote back in 2018 and havenāt touched since. They mostly work (until they donāt), and nobody really remembers how they work (until something breaks at 2 AM).
Iām talking about the classics: VM start/stop schedules, backup validation scripts, certificate expiry checks, orphaned resource cleanup. All those little operational tasks that are too important to forget but too boring to do manually - also known as toil in the DevOps world.
If youāve been following my recent posts on agentic DevOps, you know Iām a big believer in letting AI handle the repetitive parts of infrastructure reviews. But hereās the thing - most teams are still doing manual PR reviews for Bicep (and Terraform) files, catching the same anti-patterns over and over: āHey, you forgot to enable HTTPS-only on that storage account.ā āThis VM SKU is way oversized for a dev environment.ā āDid you check if this violates our naming conventions?ā
When I was still on the Microsoft Technical Trainer, delivering classes or presenting to Executives or speaking at conferences on AI (M365 Copilot, Foundry and Github Copilot primarily), apart from the technical conversations, one other question that always popped up, was āgreat scenario Peter, but what does this actually cost?ā And while GitHub Enterprise with GitHub Copilot provides great insights, not all developers use GitHub Copilot organization-wide.
Thatās where about 4 months back, I compiled a basic GitHub Copilot Consumption Viewer terminal app together, sharing simple metrics on tokens, LLMs used, etc.
If youāve been following me for a while, you know I write a fair bit about Microsoft DevOps, Azure, and lately GitHub Copilot. What you probably donāt know is that Iāve been leaning on VS Codeās agent mode to help me stay on top of whatās trending. It scans GitHub changelog feeds, filters for the DevOps-relevant bits, overall looks at top DevOps online resources - which now feels like a good topic for another blog post - and hands me a shortlist of topics worth exploring. That partās been working well for months. (And yes, I do miss the good old RSS-feeds, although I know some sites still have themā¦)
You know Iām always interested when something new ships that makes automation (DevOps, right!) easier, especially when it involves AI agents. So when two GitHub changelog entries landed on the same day last week, I knew they were made for each other. GitHub Agentic Workflows is now in public preview, and at the exact same time, agentic workflows no longer need a personal access token.
Let me explain why this matters, what an agentic workflow actually looks like, and why ditching the PAT is a genuinely good security win.
Over the last few weeks, Iāve been deep into building (even more)custom agents for my day-to-day job and the broader content team Iām in. One thing that kept bugging me was the manual wiring, where you have to point each agent to the right MCP servers, skills, and tools. Itās a lot of YAML editing and context window planning. Apart from the time, effort and testing that is still required when authoring your own custom VSCode Agents and skills.
If you have been following me for a while, you know Iām a big fan of Azure reliability. It was the main topic I presented on several years ago (in the early days of Azure - if that sounds right?) and also mapped with a big part of my job as Azure Architect, consultant and trainer.
I got amazed at the end of 2021 by Azure Chaos Studio, a service that allows you to inject faults against your Azure workloads (preferably production!), to make them more stable, more reliable.
Iām usually pretty excited about AI and more specifically about Generative AI, especially with Microsoft Copilot, GitHub Copilot and Microsoft Foundry. I might be biased, but outside of my professional interactions with GenAI, Iām not into all the social media hypes around it (anyone remembers the Studio Gihbli hype from summer or any similar?)
With a few days off for Thanksgiving here in the USA, I wanted to spend a bit more time on updating my Spotify playlists. While not always perfect, at least a few of the ārecommended artistsā are closely in line with the artists and songs I like.
For about a year now, Iāve been teaching a lot on GitHub Copilot as part of my Microsoft role. Our program offers 2 different learning paths, one created by the Microsoft Content developers, AZ-2007, and the other one is managed by GitHub Content team, known as GH-300.
If you know my approach to teaching tech a bit, which a learner in my class lately called inspiring through technology, it means Iām trying to explain as much as possible through compelling, live demos. After walking learners through different GitHub Copilot features such as documenting/explaining code, generate application code (on different development frameworks), but also Azure CLI, CI/CD pipeline, Dockerfile, YAML, JSON and alike, I usually close with Agent Mode.
Cookie Consent
We use cookies to enhance your browsing experience and analyze site traffic.
Cookie Preferences
These cookies are required for the website to function and cannot be disabled.
These cookies help us understand how visitors interact with our website.
These cookies enable features like comments and embedded content.