wiki

Maintain a markdown knowledge base with interlinked topics and an index.

820|78|Updated Mar 20, 2026

Search Tags:#markdown #llm #navigation #knowledge #wiki #knowledge-management #cross-links

Ready to run

npx skills add https://github.com/SethGammon/Citadel --skill wiki-sethgammon

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: wiki
Source: https://github.com/SethGammon/Citadel/tree/main/skills/wiki
Command: npx skills add https://github.com/SethGammon/Citadel --skill wiki-sethgammon

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Markdown-first knowledge base where the LLM acts as librarian. Ingests raw sources, compiles and interlinks topic files, self-maintains an index. No vector DB or embeddings required -- uses LLM-native navigation over structured markdown up to ~400K words.

Core Features & Use Cases

Ingest raw sources into interlinked topic files and maintain a navigable index.
Answer questions by reasoning over the organized knowledge without embeddings.
Real-world use: build and maintain project documentation, meeting notes, and research into a single searchable wiki.

Quick Start

Initialize a wiki directory at the project root and begin ingesting sources with the /wiki workflow to build the knowledge base.

Frequently Asked Questions about wiki

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

FAQPage Schema

How do I build a knowledge base from markdown without using vector embeddings?▼

You can build a markdown-native knowledge base without vector embeddings by having the LLM ingest raw sources, compile interlinked topic files, and self-maintain a navigable index for reasoning over organized information.

How do I manage and navigate project documentation using LLM-native cross-links?▼

Manage project documentation by initializing a wiki directory where the LLM compiles interlinked markdown topics and maintains an index, enabling direct navigation and question answering across your knowledge base.

Can I use an LLM to answer questions over my project wiki without a vector database?▼

Yes, you can answer questions over your project wiki without a vector database because the LLM reasons directly over structured markdown files and a self-maintained index up to approximately 400,000 words.

What is the best way to maintain a searchable knowledge wiki for meeting notes and research?▼

The best way to maintain a searchable wiki is using lifecycle commands to ingest raw sources, compact existing topics, and rebuild the index, ensuring the LLM can navigate interlinked knowledge without embeddings.

Are there limitations to using markdown-based knowledge management for large documentation sets?▼

A key limitation is capacity, as this markdown-based knowledge management approach supports structured knowledge up to roughly 400,000 words, relying on LLM-native navigation rather than scalable vector search.