Every corner of the stack

The wide bucket: CLIs, local environments, test runners, and code review.

Projects with public code, whether they are pre-1.0, pre-launch, or pre-announcement.

Docs, notes, and planning — the tools around the code rather than in it.

Hosted products where someone else carries the operational load.

Coding agents, inference runtimes, and the plumbing that keeps them reproducible.

CI, deploys, environments, and the tooling that keeps production boring.

Gateways, SDK generators, mock servers, and contract testing.

Storage engines, query layers, and the data path under everything else.

Component kits, headless primitives, and design token tooling.

The layer you build on, and the hardest one to change your mind about.

Scanning, secrets, and supply chain tooling that fails loudly by design.

Product analytics and telemetry, usually without a heavy browser tracker.