llm-wiki

Build and maintain an interlinked Markdown knowledge base with tracked provenance.

Updated May 15, 2026
One-click install
npx skills add https://github.com/cabezno/bmb-encover-agent --skill llm-wiki-cabezno
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: llm-wiki
Source: https://github.com/cabezno/bmb-encover-agent/tree/main/skills/research/llm-wiki
Command: npx skills add https://github.com/cabezno/bmb-encover-agent --skill llm-wiki-cabezno

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes assets (resource) components.

What problem does it solve?

It solves the problem of turning scattered articles, notes, and research into a durable, cross-referenced knowledge base that compounds over time instead of requiring repeated re-RAG from scratch.

Core Features & Use Cases

  • Persistent interlinked wiki: Maintains a Markdown knowledge base with entities, concepts, comparisons, and saved query results.
  • Source ingestion with provenance: Ingests sources into a structured three-layer layout and tracks provenance so claims can be traced back.
  • Orientation + consistency guardrails: Orients on existing SCHEMA.md, index.md, and recent log.md to prevent duplicates and schema drift.
  • Query + lint loop: Supports answering questions using the wiki and auditing it for orphans, broken links, frontmatter validity, staleness, contradictions, and raw drift.

Quick Start

Ask your agent to initialize a new wiki at the configured path, define the domain in SCHEMA.md, and then ingest your first set of sources to produce the initial index and log.

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 that compounds over time instead of re-running RAG?

You build a markdown knowledge base by compiling sources once into a persistent directory with SCHEMA.md, index.md, and log.md, maintaining interlinked pages with tracked provenance to avoid re-RAG from scratch.

What is the best way to ingest URLs and PDFs into a wiki while tracking provenance?

Ingest URLs and PDFs by processing them into a structured three-layer layout that tracks provenance for synthesized claims, ensuring every piece of information can be traced back to its original source.

How do I lint a markdown wiki for broken links and schema drift?

Lint a markdown wiki by auditing it for orphans, broken wikilinks, frontmatter validity, staleness, contradictions, and raw drift, using orientation on existing SCHEMA.md to prevent schema drift.

Can I query an existing markdown knowledge base for research questions?

Yes, you can query an existing markdown knowledge base by having the agent orient on the current index and schema, then answering research questions using the interlinked pages and saved query results.

Does this knowledge base approach require a specific folder structure?

Yes, the knowledge base requires a persistent wiki directory with SCHEMA.md, index.md, log.md, and layer folders, along with strict frontmatter validation and cross-linking via wikilinks to function correctly.

Why use a self-compounding wiki instead of standard RAG for research workflows?

Use a self-compounding wiki instead of standard RAG to turn scattered articles into a durable, cross-referenced knowledge base that maintains consistency and tracked provenance without repeated re-ingestion.