None of these started as products. They started as us getting tired of doing the same fiddly thing by hand, building a small tool to stop, and then realizing other people probably hate that thing too.
There's a certain type of work that eats your day without ever showing up as progress: setting up a new laptop, wiring the same cloud plumbing again, reconciling five scanners that all speak different formats. We automated the ones that bugged us most, kept using them, and figured we'd share the ones that were genuinely useful. Consider it a small window into how we like to work.
The best tool is the one you actually reach for on a normal Tuesday. Everything here earned that by getting used before it got published.
The toolbox
Each of these does one job and tries to do it cleanly. They don't replace your project's own tooling—they take a repetitive step and make it boring.
- DevOps Setup — installs or updates a consistent workstation profile on macOS and Windows from one maintained package list per platform. New machine, or a team drifting apart config by config? One command instead of an afternoon of "wait, which version do you have?"
- AITO — sets up a token-efficient workflow for GitHub Copilot and Claude Code, then measures the result instead of promising you some invented saving. It cuts down repeated repository reading and oversized build output. If it doesn't actually help, the numbers will tell you—that's the point.
- Terraform ECS / Fargate — a ready-to-fork AWS reference stack that wires ECS Fargate, an application load balancer, ECR, EFS, and optional CloudFront and PostgreSQL together with sensible defaults. It's the "how do these pieces connect again?" starter so small container projects don't rebuild the relationships from scratch.
- CostHive — runs several open-source FinOps tools behind one read-only AWS scan and hands back a consolidated, dollar-ranked report. Instead of ten dashboards and a vague sense your bill is too high, you get a prioritized list of actions, estimated monthly savings, and a confidence level for each.
- SentryHive — runs established AWS security scanners through one read-only command and normalizes their findings into one client-ready report in HTML, PDF, Markdown, or JSON. Consultants and auditors stop reconciling toolchains and get one traceable document.
What they have in common
Look at that list and a pattern shows up, which is really just how we think about building anything:
- Read-only where it touches your account. A tool that scans your cloud shouldn't be able to break it.
- Honest output. Measured savings, confidence levels, traceable findings—no invented numbers to look impressive.
- One job, done cleanly. Small enough to understand, inspect, and trust before you run it on something that matters.
Why give them away
Partly because they're more useful to us out in the open—other people find the rough edges we missed. Partly because releasing them keeps us honest about quality; code you publish is code you clean up. And partly because it's the clearest way to show what we actually stand for, without a pitch deck. Fork them, read them, tear them apart, build something better. That's the whole idea.