Seqr — Command Queue Runner

Inspiration

Seqr was born out of our team’s journey collaborating with AI coding agents, and quickly became a project we felt passionate about. As we encountered the quirks of command execution, we saw the need for a tool that could serve as both a guide and a guard, ensuring every command ran exactly as intended. The process was challenging, but it gave us more control and confidence in our workflows. This project has opened up new possibilities for our team, and we’re excited about how Seqr might be adopted in future work.

How it was built

Developing Seqr was a hands-on learning experience in automation, reliability, and transparency. We leaned into Go’s native features to build a CLI that’s both simple and robust, without relying on extra frameworks. Integrating releases and GitHub Actions was a new adventure for our team, and the timing was perfect to explore automated builds and releases. Setting these up made the project feel “real,” and we enjoyed discovering the power of the toolchain. Our build process tackled real-time monitoring, cross-platform support, and safe configuration handling, with additional features like concurrent command execution, status notifications, persistent logging, and live output streaming.

Working with Kiro

Kiro was incredibly helpful as an AI-based IDE throughout development. We especially enjoyed defining our own spec mode, something we’d wanted to try for a long time, but always outside the default settings. Since Kiro doesn’t natively support custom definitions, we had to get creative, adding verification steps to ensure changes matched our requirements. Our favorite part was designing a spec that was lean and manageable, rather than the heavy default. The workflow of defining a feature, implementing it, and then refining it (removing all the unnecessary bits added by Claude, for example) using vibe mode was genuinely rewarding.

What we learned

Through building Seqr, we discovered how much more effective our workflows could be with automation that we could actually trust and understand. We learned the importance of having process management that doesn’t get in the way, especially when working with unpredictable AI agents. Defining our own specs in Kiro, overcoming its limitations, and refining features in vibe mode showed us how much control and clarity we could achieve by tailoring tools to our needs. Seeing our ideas come to life—watching commands run and features work the way we intended—was incredibly rewarding, even when we had to fix bugs or clean up after the AI. These lessons are already shaping how we approach future projects.

Challenges

Some of the biggest hurdles were handling complex termination scenarios, supporting multiple configuration formats, and maintaining seamless cross-platform compatibility. Ensuring commands executed in the right order with the right environment, and providing real-time feedback, took careful design and thorough testing.

Reflections & Future Plans

We’re genuinely enthusiastic about Seqr’s potential. The project has already shaped how we think about automation, and we’re looking forward to revisiting it in the future to refine and expand its capabilities. The experience with Go, GitHub Actions, and Kiro has been transformative, and we’re excited to see how these tools will influence our team’s projects going forward.

Built With

• cli
• go