Second Brain — a knowledge hub over a Git repository
A read-and-write knowledge store that every AI surface we run can query, backed by an ordinary repository so the history is legible and the data is never trapped. The interesting part is the protocol, not the storage.
Not published
Keelstone — a multi-tenant foundation
The starting point we reach for when a system needs tenants, row-level security and an audit trail from the first commit rather than retrofitted at month six. Opinionated, small, and boring on purpose.
Not published
Pagination helpers for row-capped APIs
Small utilities that make a truncated response impossible to miss — cursor loops, exact counts, and a hard failure when a result set hits a provider’s ceiling instead of quietly returning a round number.
Not published