Downloads · 6 artifacts
Take the templates,
not just the theory.
Every artifact here is pulled from the systems the essays describe: the actual schemas, skill files, and loop skeletons my agents run on. Copy them, adapt them, ship them.
The actual CLAUDE.md running a portfolio of agent-operated businesses, sanitized only where keys, client names, and project ids lived: the root file plus the brain README it points at.
The annotated template for programming an agent's behavior with a document instead of fine-tuning: boot sequence, anatomy, and a correct-vs-wrong schema side by side.
A starter Postgres schema for a shared brain: one knowledge table plus one task queue, the single source of truth every agent reads first and writes to last.
A per-agent memory and lessons log so a fleet keeps what it learned across crashes, handoffs, and every agent that runs while you sleep.
The one decider, many disposable workers pattern as a copy-pasteable skeleton: atomic claims, worker caps, wall-clock timers, and one rolled-up note upward.
A kit for agents that write their own tools and reuse them forever: the registry row plus the write-test-register loop that keeps a tool.
Each template comes from an essay in the series, where the pattern behind it is explained in full.