Deep Dive · Knowledge Graph

Connecting Knowledge

22 relation types, path finding and context traversal — how the graph works.

5 Categories, 22 Relation Types

Each relation has a direction and an automatic reverse. 'A depends on B' also creates 'B required by A'.

🔗

Structural

depends_on ↔ required_by
part_of ↔ contains

⚡

Causal

caused_by ↔ causes
fixed_by ↔ fixes

⚙️

Semantic

implements, supersedes
related_to

📝

Knowledge

validates, contradicts
documents, extends

⏩

Workflow

precedes ↔ follows
blocks ↔ blocked_by

Finding Paths — How Does A Connect to B?

The graph finds the shortest path between any two nodes. This reveals connections you wouldn't see immediately.

You: "Context Memory: How does the cache bug connect to the book structure?"

Claude: 🛤️ Path (3 hops):
  📌 Cache Bug
  ↓ caused by
  📌 Multi-Tenant Decision
  ↓ required by
  📌 API Architecture
  ↓ documents
  📌 Book Structure

Context Retrieval — Everything in N Hops

Before working on a topic, load all related context. Depth 1 = direct neighbors. Depth 2 = their neighbors too.

You: "Context Memory: Show me everything connected to the DB decision."

Claude: 🌐 Context for dec-001 (2 hops):
  ●○ arch-001: REST API (architecture)
  ●○ bug-001: Race Condition (bug)
  ●○ pat-001: Repository Pattern (pattern)
  ●● cfg-001: Redis Cache (config) ← 2 hops
  ●● les-001: N+1 Query Fix (lesson) ← 2 hops

Mermaid — See Your Graph

Export as Mermaid diagram for visual rendering. Color-coded by type, different arrow styles per relation. Open in Obsidian, GitHub, or mermaid.live.