I built something that still feels a bit surreal to watch in action: an orchestrator that chains together AI agents to handle the entire software development lifecycle. From planning to deployment, it’s agents all the way down.
Honestly, it’s not even anything too impressive or hard to build at all. Just some simple python to handle a workflow system, and to wrap either the claude cli tool or codex cli. The really crazy part is that once you have the idea, it takes an afternoon to get it up and running.
The SDLC Agents Orchestrator is a tool that automates software development workflows using AI agents. You define a workflow in YAML, point it at a task, and watch as AI agents:
- Plan the implementation
- Write the code
- Test it
- Review it
- Generate documentation
- Handle deployment
Each stage produces artifacts and reports that feed into the next stage. It supports multiple AI platforms (Claude, Codex), can resume from failures, and even allows human-in-the-loop interactions when you want to stay involved.
There’s something genuinely mesmerizing about watching it work. You kick off a workflow and then it’s just… going. Planning agent hands off to coding agent. Coding agent produces files. Testing agent finds issues and feeds them back. Review agent suggests improvements.
The first time I watched it complete a full cycle—planning a feature, implementing it, writing tests, and generating documentation—I had that uncanny valley feeling. This is what I do for a living. And here’s an automated system doing it, end-to-end, in a fraction of the time.
The logs stream by. Agents make decisions. Code appears. Tests run. Sometimes it fails and retries. Sometimes it asks for clarification. But mostly, it just… works.
Obviously, this type of system can have difficulties on gigantic enterprise repos that aren’t even designed well for humans to reason about, much less an AI Agent. But honestly, I think this kind of system is powerful enough where we’ll have to redesign those repos for the new world (or let them die off and just start new projects, more likely).
At some point in the future, whether it’s 1 year or 10, you will not be touching code. You may not even be looking at it. I know that’s unthinkable for many of you now, but it’s 100% coming, and the faster you realize it the better you can prepare yourself.
And that future always arrives faster than we expect. Building this orchestrator taught me that we’re not preparing for some distant sci-fi scenario. The tools exist now. They work now. The question isn’t “will this change software development?” but “how quickly will it change, and are we ready?”
The orchestrator is open source. Try it. Watch it work. Feel that same uncanny valley moment I did. Because ready or not, this is where we’re headed.
Check out the project: github.com/Chris-Cullins/agent_orchestrator