mcpField
V
vector-cluster

vector-cluster

github.com/ruvnetData & databases80.0k728 installsUpdated today
Repository

About vector-cluster

vector-cluster is a skill for grouping vectors by semantic similarity with ruvector. It runs graph clustering on a code graph and provides cluster assignments, community labels, and edges for review and storage.

Key features

  1. 01Runs spectral and Louvain graph clustering
  2. 02Produces cluster assignments and community labels
  3. 03Reports graph edges for review
  4. 04Stores results in a namespace

Use cases

  1. 01Discover themes in vector collections
  2. 02Identify outliers in a graph
  3. 03Organize clusters for later review

MCP support

Uses a Ruflo MCP memory tool to store clustering results in the vector-clusters namespace.

Install

Copy the command and run it in a terminal.

$npx skills add https://github.com/ruvnet/ruflo --skill vector-cluster

After installing, type/vector-clusterin Claude Code to confirm it loaded.

Details

From the project README

Vector Cluster

Cluster vectors in a namespace by semantic similarity using ruvector.

When to use

Use this skill when you have a collection of embeddings and want to discover natural groupings. Clustering reveals themes, identifies outliers, and helps organize large vector collections.

Steps

  1. Ensure [email protected] is available: 
    npm ls ruvector 2>/dev/null | grep '0.2.25' || npm install [email protected]
  2. Run clustering — in [email protected] the only working clustering is via hooks graph-cluster (spectral/Louvain over a code graph). The top-level cluster command is reserved for distributed cluster ops and is currently "Coming Soon" upstream. 
    npx -y [email protected] hooks graph-cluster <files...>
npx -y [email protected] hooks graph-mincut <files...>
  3. Review output — JSON with cluster assignments, community labels, and edges. If you see "graph.nodes is not iterable", run hooks init first to seed the graph state.
  4. Store results: mcp__plugin_ruflo-core_ruflo__memory_store({ key: "clusters-PROJECT-TIMESTAMP", value: "CLUSTER_ASSIGNMENTS", namespace: "vector-clusters" })

Interpreting results

  • High cohesion (>0.85): tight, well-defined cluster
  • Medium cohesion (0.6-0.85): related but diverse content
  • Low cohesion (<0.6): loose grouping, try higher resolution
  • Outliers: novel or anomalous files worth investigating

Caveats

  • cluster --namespace ... --k N and cluster --density are not valid in [email protected] — those flags fall through to the distributed-cluster command, which only accepts --status, --join, --leave, --nodes, --leader, --info.
  • For namespaced k-means over arbitrary embeddings, run k-means in your own code against vectors stored in AgentDB.

Comments

Markdown supported