OUR WORK

Zedbee

From a shell script. To a published developer tool.

Visit www.zedbee.dev
www.zedbee.dev
Built
The tool, its identity, website, and documentation.
Used for
Repeatable code-quality and security checks in my development workflow.
Inspect it
Explore the real scan, source, package, and documentation below.
THE PRODUCT & THE WEBSITE

A tool to use. A website to explain it.

I built the tool, its visual identity, and the website that explains it. Here’s how those parts came together.

01

Make the results readable.

Each analyzer speaks its own language. I built adapters that turn their output into consistent findings, then designed a terminal interface that shows what’s running and what needs attention. Humans get a readable view; coding agents and automation can consume structured results.

A real Swarm scan from my development workflow.

02

Explain it before asking for the install.

The website has to explain the tool before asking someone to install it. I designed the black-and-yellow identity, product demonstration, and searchable documentation around that journey. Installation, configuration, and technical limits are easy to find. That same attention to content and navigation helps customers understand a business website.

The documentation, from first install to team configuration.

03

Handle the details that matter in daily use.

A check should be about the change someone is making. Swarm analyzes staged code or committed branch changes and separates new or worsened findings from unrelated existing issues. Incomplete checks stay visible. Working through those details is part of building software people can depend on, beyond what a polished screenshot can show.

Explaining how findings relate to a change. Detail view.

THE BUILD

Ready for someone else’s project.

I built setup to inspect an existing project and preview the changes before writing them. Teams can keep their project’s Prettier setup through an explicitly trusted formatting mode. Swarm brings formatting, lint, types, secrets, dependency checks, and other analysis into one workflow, with JSON and SARIF reports for agents and CI systems.

Building Swarm meant turning something that worked on my machine into something another person could install and understand. The interface, documentation, configuration, and underlying code all had to work together. I bring that experience to MAWD projects, whether a business needs a straightforward website or a custom integration behind it.

PUBLISHED & SOURCE AVAILABLE

You can look under the hood.

Swarm is published on npm as zedbee. Its source is publicly available on GitHub under the PolyForm Small Business License 1.0.0.

You can explore the code, read the documentation, or visit the live site to see the work beyond this case study.

THE ORIGIN

I wanted another set of eyes.

While building ResumeXRays in December 2025 and January 2026, I wanted a second set of eyes on the code AI was helping me write. Was it following my formatting and coding standards? Had a password or API key slipped into a file? I put together a shell script that ran several analysis tools and had AI review their results.

It did what I needed, but it was a quick solution for my own project. When I introduced agentic development to my team at work, I built a cleaner version. Zedbee Swarm grew out of those two experiences: a tool I could package, document, and make available to other developers.

Your business has a story, too.

Let’s bring it to life
NEXT PROJECTResumeXRays