llm-wiki

Ingest fragmented sources into an interlinked local knowledge base for AI agents.

2|Updated May 25, 2026
One-click install
npx skills add https://github.com/XiaSanw/LLM-WIKI-Xiasanw --skill llm-wiki-xiasanw
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: llm-wiki
Source: https://github.com/XiaSanw/LLM-WIKI-Xiasanw/tree/main/skill
Command: npx skills add https://github.com/XiaSanw/LLM-WIKI-Xiasanw --skill llm-wiki-xiasanw

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

llm-wiki solves the challenge of turning scattered, ephemeral AI interactions into a durable, interlinked local knowledge base that agents can read and remember across sessions.

Core Features & Use Cases

  • Ingest various sources (web, social, PDFs) and convert to markdown pages
  • Build interconnected entities, topics, sources, and synthesis pages
  • Support multi-platform agent workflows with a shared core

Quick Start

Initialize a wiki and ingest your first source URL to start building.

Frequently Asked Questions about llm-wiki

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I build a local knowledge base from fragmented information for AI agents?▼

You can build a local knowledge base by using an ingestion workflow to convert scattered sources like web pages, social posts, and PDFs into interlinked markdown pages. This organizes fragmented information into a durable, growing structure that AI agents read and remember across sessions.

What is the best way to ingest web pages and PDFs into a markdown knowledge graph?▼

The best way to ingest web pages and PDFs into a markdown knowledge graph is using a batch-ingest workflow that converts external sources into structured pages. It enforces a strict source registry and automated quality checks, ensuring stable functionality and clear entity interlinking.

Can I use a shared knowledge base across multiple AI agent platforms like Claude Code and Codex?▼

Yes, you can use a shared knowledge base across Claude Code, Codex, and OpenClaw. The system uses a shared core with thin platform adapters, ensuring consistent workflows for initialization, ingestion, querying, and regression validation across all entry points.

How do I validate and maintain quality in an interlinked markdown knowledge base?▼

You validate and maintain quality in an interlinked markdown knowledge base by running automated lint and status checks. These quality checks enforce a strict contract via a source registry, ensuring stable core functionality and regression validation across all platform adapters.

Does an AI agent knowledge base support querying interconnected topics and synthesis pages?▼

Yes, an AI agent knowledge base supports querying interconnected topics and synthesis pages. The query workflow allows agents to read and retrieve interconnected entities, topics, and source data, turning ephemeral interactions into a durable memory structure.