Browse / LLM wiki / llm-wiki-loop
llm-wiki-loop
PALAN-K
The reference architecture for LLM-maintained knowledge vaults — schema, lifecycle loop, machine verification, self-installation
Open PALAN-K/llm-wiki-loop → Manually confirmed as a workspace.
Works with any agent
ships AGENTS.md, the cross-tool instruction file.
In its own words
The Production Framework for Self-Improving & Self-Organizing LLM Knowledge Vaults. Grounding Invariants • Event-Driven GC • Auto-Skillification • Multi-Agent 1-Click Injection
🌏 Read in: 🇺🇸 English • 🇯🇵 日本語 • 🇰🇷 한국어 | 🌐 Live Demo: EN • JA • KO | Zero-DB • 100% Markdown • 7 Agents • 25 Tests • v1.3.2
↑ One take — $ npx llm-wiki-loop init (0.8s, 7 agents) → $ check --strict (0 errors, 0 drifted) → $ scan:skills (1 candidate) — 22s, 860×484
⚡ Why Ordinary AI Notes Fail vs. llm-wiki-loop
4-Stage Self-Improving Loop — 1 sec ingest → 0-token verify → event GC & 2× skillify (proposal-first)
💡 What is llm-wiki-loop?
llm-wiki-loop transforms volatile AI outputs into an immutable, self-organizing second brain.
Most LLM notes gradually hallucinate, cite outdated assumptions, or bloat over time. llm-wiki-loop fixes this at the architectural layer:
Quoting 15% of the README. Read the rest on GitHub ↗
Read from the repository's git tree on 2026-08-17.