AI-augmented development, documented in public.
Racecraft Lab is the technical innovation lab of Racecraft Systems. We build tools, systems, and public writing around one question:
How do technical teams use AI agents without accumulating verification debt?
Our answer is governed, spec-driven engineering: clear requirements, bounded agents, explicit review gates, human ownership, operational visibility, and honest retrospectives.
Racecraft is a platform for agentic software work, developed through real products instead of slideware:
- Methodology: specs, runbooks, Lab Notes, and review gates that turn AI assistance into repeatable engineering practice.
- Developer workflow: public and private Racecraft plugins that package planning, review, research, and automation patterns.
- Control plane: Mission Control, HAL, and OpenClaw integration work for coordinating agents, tasks, cost, and operational state.
- Product lines: FocusEngine first, then Racecraft Lab and GTO90, proving the method on real Mac, web, and creative software.
Some platform work stays private while it is hardened. The public repositories here are the visible edge of the lab: plugins, docs, forks, and experiments that make the method inspectable.
- racecraft-plugins-public - open plugins and skills for agentic development workflows.
- paddock - AI software factory control plane for experiments in AI agent orchestration, task routing, review states, and resource governance.
- openclaw - OpenClaw-facing experiments for personal AI assistant infrastructure.
- docs - documentation work connected to OpenClaw and the broader lab stack.
We write about verification debt, spec-driven development, agent orchestration, ethical AI for creative software, AWS full stack AI, and the practical tradeoffs that show up when AI agents touch real projects.
- Proof before pitch.
- Specs before code.
- Human ownership over AI output.
- Failures are data.
- Tools change; governed methodology should survive the churn.
If you are building with AI agents and care about reliability, traceability, and creative integrity, this is the lab notebook.
Watch the public repositories, read the docs, and follow the work as it moves from private hardening to public release.