playbooks

home / skills / laurigates / claude-plugins / docs-knowledge-graph

docs-knowledge-graph skill

/documentation-plugin/skills/docs-knowledge-graph

This skill builds a comprehensive knowledge graph from Obsidian vault z/ docs, enabling semantic search and pattern recognition across technical domains.

npx playbooks add skill laurigates/claude-plugins --skill docs-knowledge-graph

Review the files below or copy the command above to add this skill to your agents.

Files (1)
SKILL.md
1.8 KB
---
model: opus
created: 2025-12-16
modified: 2025-12-16
reviewed: 2025-12-16
allowed-tools: Read, Glob, Task, TodoWrite
argument-hint: [vault-directory]
description: Build comprehensive knowledge graph from Obsidian vault documentation
name: docs-knowledge-graph
---

# Build Knowledge Graph

Build a comprehensive knowledge graph from technical documentation files in an Obsidian vault's z/ directory.

## Usage

```bash
claude build-knowledge-graph
```

## What it does

This command systematically processes all files in the z/ directory of your current Obsidian vault and commits their summaries and relationships to the memory system using the memory-keeper agent. It:

1. **Scans** all documentation files in the z/ directory
2. **Processes** them in logical groups (AI/Automation, Infrastructure, DevOps, Security, etc.)
3. **Extracts** key entities, relationships, and technical specifications
4. **Builds** a comprehensive knowledge graph for semantic search and pattern recognition
5. **Preserves** architectural decisions, integration points, and operational context

## Output

- Complete knowledge graph stored in memory with group_id "technical-docs"
- Rich entity relationships for cross-referencing technologies and projects
- Enables intelligent queries about system dependencies and technical patterns

## Requirements

- Must be run from an Obsidian vault directory containing a z/ subdirectory
- Memory-keeper agent must be available
- Files should be technical documentation in Markdown format

## Example Use Cases

- Building institutional knowledge from years of technical documentation
- Creating searchable relationships between projects and technologies
- Preserving architectural decisions and their rationale
- Enabling intelligent queries about system dependencies

Overview

This skill builds a comprehensive knowledge graph from technical documentation stored in an Obsidian vault z/ directory. It extracts entities, relationships, and architectural context and commits them to a memory system for semantic search and intelligent queries. The graph preserves decisions, integration points, and operational patterns across your documentation.

How this skill works

The command scans all Markdown files under the vault's z/ directory, groups documents by domain (AI/Automation, Infrastructure, DevOps, Security, etc.), and extracts key entities and relationships. It summarizes content, identifies dependencies and design rationales, and writes the resulting nodes and edges into the memory-keeper agent under the group_id "technical-docs". The stored graph enables cross-referencing, pattern recognition, and semantic queries against your institutional knowledge.

When to use it

  • After consolidating or importing legacy technical docs into an Obsidian vault
  • When you need a single source for system dependencies and architectural rationale
  • Before onboarding new engineers who need context across projects
  • To enable semantic search and cross-referencing across years of documentation
  • When preparing for audits, migrations, or major architectural changes

Best practices

  • Keep technical files in z/ as standard Markdown with clear headings and metadata
  • Organize files into logical folders or tags (AI, Infrastructure, DevOps, Security) to improve grouping
  • Run the build process from the vault root to ensure all z/ files are discovered
  • Verify memory-keeper agent availability and credentials before running
  • Review and curate extracted summaries and relationships periodically to maintain accuracy

Example use cases

  • Generate an organizational map of services, libraries, and their dependencies for a large platform
  • Surface historic architectural decisions and their rationales during a migration planning phase
  • Create semantic queries to find all projects that use a particular technology or pattern
  • Preserve cross-project integration points for faster incident response and root cause analysis
  • Build searchable institutional knowledge for onboarding and training programs

FAQ

Where does the graph get stored?

The generated graph is committed to the memory-keeper agent under the group_id "technical-docs" for semantic retrieval.

What files are processed?

All Markdown technical documentation files inside the vault's z/ directory are scanned and processed.