Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
Donald Thompson avatar

Azkg

  • 2 repo stars
  • Updated October 19, 2025
  • witt3rd/claude-plugins

Maintain a Zettelkasten knowledge graph automatically with an MCP server, specialized agents, and workflow hooks.

About

AZKG is an agent-maintained Zettelkasten knowledge graph backed by an MCP server, specialized agents, and workflow hooks that keep notes and links current automatically. A developer uses it to give an agent persistent, structured memory and to automate note-taking and knowledge organization across sessions.

  • Agent-maintained Zettelkasten knowledge graph
  • Bundled MCP server
  • Workflow hooks and specialized agents

Azkg by the numbers

  • Data as of Jul 7, 2026 (Skillselion catalog sync)
/plugin marketplace add witt3rd/claude-plugins
/plugin install azkg@witt3rd-claude-plugins

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
repo stars2
Last updatedOctober 19, 2025
Repositorywitt3rd/claude-plugins

What it does

Maintain a Zettelkasten knowledge graph automatically with an MCP server, specialized agents, and workflow hooks.

README.md

Agentic-ZKG Plugin

Agent-maintained Zettelkasten Knowledge Graph for Claude Code

🚧 Status: In Development - Phase 1 implementation underway

Overview

This plugin implements the Agentic-ZKG paradigm - agent-maintained knowledge graphs with conversational interface.

Features

19 Slash Commands

Core Knowledge Management (6):

  • /create-note [topic] - Research and create with auto-linking
  • /search-notes [query] - Find notes by keyword
  • /expand-graph [note] - Discover missing relationships
  • /learning-path [target] - Generate prerequisite sequence
  • /graph-note [file] - View relationships
  • /refresh-topic [file] - Update with latest info

Content Ingestion (4):

  • /process-pdf <pdf_url_or_path> - Extract PDF content and create knowledge graph notes
  • /scrape-url <url> - Scrape web content and create knowledge graph notes
  • /youtube-transcript <youtube_url> - Download YouTube transcripts and create notes
  • /telegram-next-message - Process Telegram saved messages for knowledge capture

Maintenance (3):

  • /conform-note - Restructure notes to question-oriented format
  • /rename-note - Rename notes and update wikilinks
  • /graph-validate - Run validation checks

Analysis (2):

  • /graph-stats - Display knowledge graph statistics
  • /graph-moc - Display MOC information

Granular Operations (2):

  • /graph-add-relationship - Add typed bidirectional relationships
  • /update-note - Update note metadata

Shared Methodology:

  • All content ingestion commands use question-oriented content synthesis
  • Follows methodology in _shared_content_synthesis.md
  • Ensures consistent note structure across all sources

MCP Server

Exposes knowledge graph as MCP resources and tools:

  • Resources: kg://notes, kg://note/{file}, kg://relationships/{file}
  • Tools: search_knowledge, find_related, get_learning_path

Architecture

Markdown-first - No JSON graph, markdown is source of truth:

  • Wikilinks [[note]] → implicit relationships
  • YAML frontmatter → metadata
  • "Related Concepts" sections → typed relationships
  • MOC files → thematic organization

Installation

/plugin marketplace add witt3rd/claude-plugins
/plugin install azkg@witt3rd

Links

License

MIT

Command Documentation

For complete command specifications, implementation details, and usage examples, see:

COMMANDS.md - Comprehensive command reference with:

  • All 19 command specifications
  • Tier organization (Core, Content Ingestion, Maintenance, Analysis, Granular)
  • Tool usage (Read, Write, Edit, Grep, Glob, Bash)
  • Rationalization and recommendations
  • Future command proposals

Quick Reference

Commands are organized by usage frequency:

Tier 1 - Core (Daily): create-note, search-notes, expand-graph, learning-path, graph-note, refresh-topic

Tier 1.5 - Content Ingestion (As Needed): process-pdf, scrape-url, youtube-transcript, telegram-next-message

Tier 2 - Maintenance (Weekly): conform-note, rename-note, graph-validate

Tier 3 - Analysis (Monthly): graph-stats, graph-moc

Tier 4 - Granular (Rarely): graph-add-relationship, update-note

Related skills

AI & Agent Buildingagentsautomation

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.