llm-wiki

A self-maintaining, continuously evolving knowledge base for Agentic AI, AI Security, LLMs, Physical AI, and Secure Coding — powered by a local LLM and driven by coding agents.

What Is This?

LLM Wiki replaces fragmented, stateless RAG workflows with a stateful knowledge compilation engine. Instead of searching raw documents at query time, every source is:

• Ingested — fetched from arXiv, CVE feeds, GitHub, or dropped manually
• Normalized — converted to structured Markdown with YAML frontmatter
• Extracted — a local LLM pulls out entities, claims, and relationships as JSON
• Integrated — knowledge is merged into a persistent, cross-linked wiki
• Queryable — the wiki is searched with BM25 semantic ranking

The result is a knowledge base that compounds over time — every new source enriches the existing structure rather than existing in isolation.

Setup

3. Configure your environment

Then edit .env with your local values:

Important: .env is gitignored. Never commit it. Use .env.example as the committed template.

4. Open as an Obsidian vault

Open Obsidian → "Open folder as vault" → select this directory.

Required Obsidian plugins (bundled in .obsidian/):

• Dataview — live frontmatter dashboards
• Obsidian Git — auto-commit after write cycles
• Marp — export wiki pages as slide decks

Wiki Page Format

Every wiki page uses YAML frontmatter compatible with Obsidian Dataview: