LinkMind for Openclaw

LinkMind is a local knowledge connection engine that transforms heterogeneous content into a unified network of explainable nodes and evidence-based answers.

harrylabsj
v1.0.0
Apr 5, 2026
0
138
0

Install & Download

1. ClawHub CLI

The fastest way to install a skill directly from the registry.

npx clawhub@latest install linkmind-pro

2. Manual Installation

Copy the skill folder to one of these locations

Global
~/.openclaw/skills/
Workspace
<project>/skills/

Priority: Workspace > Local > Bundled

3. Prompt Installation

Copy this prompt to OpenClaw to install it automatically.

Help me install linkmind-pro using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).

Prefer to download?

Get the raw skill files in a ZIP archive.

What is LinkMind?

LinkMind serves as a sophisticated local knowledge hub designed to bridge the gap between raw data and structured intelligence. By ingesting local text files, it breaks down content into semantically friendly fragments and extracts concept nodes, effectively building a manageable knowledge graph on your local machine.

As a core component of Openclaw Skills, LinkMind focuses on explainability and evidence. It doesn't just provide answers; it returns evidence fragments and related concept nodes, ensuring every piece of information is traceable back to its source. The Phase 2 update introduces modular storage and embedding abstractions, allowing users to choose between simple JSON storage or high-performance SQLite backends.

LinkMind Use Cases

  • Building a personal second brain or knowledge base from local markdown notes.
  • Performing hybrid searches that combine exact keyword matching with semantic vector similarity.
  • Extracting structured concept networks from unstructured technical documentation.
  • Locally hosting an AI-ready research assistant with zero external dependencies.

How LinkMind Works

  1. Normalization: The Ingest Normalizer converts raw files (like Markdown) into standard Document entities.
  2. Fragmentation: The Unit Builder segments documents into semantic fragments and extracts key concept names.
  3. Linking: The Link Builder establishes connections between fragments and concepts to create a relationship graph.
  4. Embedding: Text fragments are converted into vectors using a chosen provider (Mock or OpenAI-compatible).
  5. Retrieval: When a query is issued, the system performs a dual-layer search using keyword matching and cosine similarity.
  6. Composition: The Answer Composer assembles the final response, including summaries, evidence fragments, and statistical overviews.

LinkMind Setup

Navigate to the skill directory and build the project using npm:

cd skills/linkmind

# Install dependencies (required for SQLite)
npm install better-sqlite3

# Build the skill
npm run build

# Run the CLI to see available commands
node dist/index.js --help

LinkMind Data Schema & Taxonomy

LinkMind organizes data into four primary tiers to ensure high retrieval accuracy within Openclaw Skills:

Entity Description Key Fields
Document The original source file. id, title, sourceUri, status
Fragment A semantic slice of a document. text, summary, conceptNames
Concept A unique normalized idea/entity. name, normalizedName, salience
LinkEdge The relationship between nodes. fromId, toId, score

LinkMind Advanced Features

  • Hybrid Retrieval: Merges keyword matching with vector-based rankers for superior search precision.
  • Storage Adapters: Switch between JsonStorageAdapter for zero-config use and SqliteStorageAdapter for high-performance batch writes.
  • OpenAI Compatible: Seamlessly connect to OpenAI, vLLM, or Ollama for local or cloud-based embeddings.
  • Evidence-Based Answers: Every query return includes a list of evidence fragments to verify AI-generated summaries.

SKILL.md


Loading

Related Openclaw Skills

METADATA

Github Stars: 0
forks: 0

Featured*