llm-wiki

Build and maintain an interlinked markdown knowledge base from ingested sources.

Updated May 28, 2026
One-click install
npx skills add https://github.com/patty-chow/the-stable --skill llm-wiki-patty-chow
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: llm-wiki
Source: https://github.com/patty-chow/the-stable/tree/main/skills/research/llm-wiki
Command: npx skills add https://github.com/patty-chow/the-stable --skill llm-wiki-patty-chow

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Traditional RAG rediscovers knowledge from scratch on every query, losing cross-references and synthesis. This Skill maintains a persistent, compounding markdown wiki where sources are ingested once, cross-linked, and kept current, so accumulated knowledge is immediately queryable. ## Core Features & Use Cases - Source Ingestion: Capture URLs, PDFs, and pasted text into an immutable raw/ layer with sha256 drift detection, then synthesize entity, concept, and comparison pages with wikilinks and provenance markers. - Query & Synthesis: Answer domain questions by reading the index and relevant pages, then file valuable answers back into the wiki as query or comparison pages. - Lint & Health Checks: Audit the wiki for orphan pages, broken wikilinks, stale content, contradictions, tag taxonomy violations, and source drift. - Use Case: A researcher tracking AI/ML developments ingests arxiv papers and articles weekly; the agent updates entity pages for models and labs, flags contradictions between sources, and keeps the Obsidian-compatible vault navigable. ## Quick Start Ask the agent to create a wiki for your research domain and ingest your first source URL into it.

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 markdown knowledge base with an AI agent?▼

Set the WIKI_PATH environment variable or use the default ~/wiki directory, then ask the agent to initialize a wiki for your domain. The agent creates SCHEMA.md, index.md, and log.md, then ingests sources into raw/ and synthesizes cross-linked entity and concept pages.

How is an LLM wiki different from RAG?▼

RAG retrieves and reprocesses raw documents on every query, while the wiki compiles knowledge once into interlinked markdown pages. Cross-references, contradictions, and synthesis persist across sessions, so queries read pre-digested pages instead of raw sources.

Does the llm-wiki skill work with Obsidian?▼

Yes, the wiki directory functions as an Obsidian vault out of the box. Wikilinks render as clickable links, YAML frontmatter powers Dataview queries, and raw/assets holds images. On headless servers, obsidian-headless can sync the vault via Obsidian Sync.

How do I check my markdown wiki for broken links and orphan pages?▼

Ask the agent to lint the wiki. The lint operation scans for orphan pages with no inbound wikilinks, broken links to nonexistent pages, index completeness, frontmatter validation, stale content, contradictions, and source drift via sha256 comparison.

What are the limitations of an agent-maintained markdown wiki?▼

The agent must orient by reading SCHEMA.md, index.md, and the log before every session, or it risks creating duplicate pages. Large wikis over 100 pages require search passes beyond the index, and pages over 200 lines must be manually split to stay scannable.