name	find-clusters
description	Find knowledge clusters (groups of highly connected notes). Triggers on "clusters", "knowledge clusters", "note clusters", "find clusters".
auto_trigger	true
trigger_keywords	clusters, knowledge clusters, note clusters, find clusters, show clusters, topic clusters, communities, note groups, knowledge areas, related groups, knowledge domains, topic areas, note neighborhoods, groupings, what topics exist
allowed-tools	mcp__flywheel__get_backlinks, mcp__flywheel__get_forward_links, mcp__flywheel__find_hub_notes

Knowledge Clusters

Name: find-clusters
Author: bencassie

Find groups of notes that are densely connected (knowledge clusters).

Purpose

Clusters are groups of notes that:

Link heavily to each other
Form cohesive topics or projects
Represent distinct areas of knowledge

Identifying clusters helps:

Understand vault organization
Find topic boundaries
Discover implicit categories
Plan folder restructuring

What is a Cluster?

A cluster is 3+ notes where:

Each note links to at least 2 others in the group
The group is more connected internally than externally
Forms a cohesive topic/project

Process

1. Find Hub Notes

hubs = find_hub_notes(min_links=10)
  → Start with highly connected notes

2. Expand from Each Hub

for hub in hubs:
  cluster = [hub]

  // Get all notes linking to/from hub
  connections = get_backlinks(hub) + get_forward_links(hub)

  // Find notes that also link to each other
  for note in connections:
    shared_links = count_shared_connections(note, cluster)
    if shared_links >= 2:
      cluster.append(note)

3. Score Cluster Cohesion

cohesion = internal_links / (internal_links + external_links)
  → Higher cohesion = tighter cluster

4. Report Results

Knowledge Clusters
═══════════════════════════════════════════════
Found 12 knowledge clusters

🎯 Top Clusters by Size:

**Cluster 1: Data Platform** (23 notes, 89% cohesion)
Hub: [[Main Project]]
Notes: [[Database]], [[API]], [[Pipeline]], [[Data Quality]], ...

**Cluster 2: Development Tools** (18 notes, 76% cohesion)
Hub: [[Dev Environment]]
Notes: [[Git]], [[VS Code]], [[GitHub]], [[Python]], [[Node.js]], ...

**Cluster 3: Work Process** (15 notes, 82% cohesion)
Hub: [[Workflow]]
Notes: [[Deployments]], [[Releases]], [[Testing]], [[Change Management]], ...

📊 Cluster Statistics:
   • Total clusters: 12
   • Average size: 14.3 notes
   • Average cohesion: 78%
   • Unclustered notes: 200 (20% of vault)

💡 Insights:
   • 3 major topic areas dominate vault
   • Data platform cluster is largest and most cohesive
   • Consider creating MOC (Map of Content) for each cluster
   • Unclustered notes may need topic assignment

═══════════════════════════════════════════════

Cluster Actions

For each cluster:

Create MOC: Make a Map of Content page linking to all cluster notes
Tag consistently: Apply cluster tag to all members
Move to folder: Consider organizing by cluster
Strengthen links: Add missing connections within cluster

MOC Example:

# Data Platform MOC

Core Platform:
- [[Main Project]]
- [[Database]]
- [[Architecture]]

Data Sources:
- [[Source A]]
- [[Source B]]
- [[Data Quality]]

[... organized cluster view]

Version: 1.0.0

find-clusters

Install Skill

SKILL.md