Lore: Memory by Agents, for Agents

A field guide in five parts explaining lore — a per-project knowledge store for coding agents that treats memory as falsifiable claims, wires itself into the agent harness with hooks and ledgers, and audits its own advice.

This repo publishes the guide as a single self-contained HTML page.

Read it: https://rickarize.github.io/lore-explained/

Contents

• index.html — the full interactive guide (self-contained: inline CSS + JS, no external assets)

Parts

1. The Amnesiac Machine — why coding agents forget, and the design commitment lore follows from
2. The Harness — hooks, CLI, skills, agent roles, adapters, and the audit loop
3. Trust, but Verify — hard write-time blocks, soft protocol, and a measurement substrate agents never see
4. Is Any of This New? — what's actually novel
5. What Could Be Better — what remains unsolved

---

Originally authored as a Claude artifact.