Engineering for AI agents, built in the open.
schematize.org is where we develop our open-source tools and share technology research on how to build, secure and orchestrate AI agents that hold up in the real world.
Mission
Make agent-built software trustworthy.
AI agents can write a lot of code fast — but fast is worthless if it isn't correct, secure and maintainable. We build the open tools and the shared standards that turn agent output into software you can actually run in production, and we publish what we learn along the way.
The ecosystem
Where the work lives
Everything schematize builds in the open, organized across three places.
Skills
The catalog of engineering skills that make AI agents follow house standards — installable with one command.
Visitresearch.schematize.orgResearch
Papers, notes and findings from our technology research — published as we learn.
Visitlab.schematize.orgLab
Experiments and prototypes in the open — early ideas before they graduate into tools.
VisitBacked by schematize.net
Open work, commercial strength
The lab is funded by schematize's commercial arm. The same team ships open source, consults and teaches — so research becomes tools you can actually deploy.
What we work on
Open-source projects
The schematize CLI, the engineering skill catalog for Claude Code, and the tooling that makes it all reproducible — all readable, forkable and free.
Learn moreTechnology research
How to engineer with AI agents reliably and securely: orchestration, standards, evaluation and the security of AI-assisted delivery.
Learn moreHow we work
Open by default
Our tools are open-source and our findings are public. If it matters, you can read it, run it and build on it.
Rigorous
Standards over vibes: normative playbooks, real tests, and security treated as a first-class requirement — not an afterthought.
Practical
Research that ships. Everything we publish is meant to change how real teams build with agents, not to sit in a drawer.