MindGraph integrates Obsidian-style wikilinks and learnable MindSkills into your workspace to create a connected, intelligent knowledge base.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install mindgraph
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install mindgraph 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 MindGraph?
MindGraph is a powerful knowledge management and process engine designed for AI agents. By implementing Obsidian-style [[wikilinks]] as a standard practice for all workspace files, it allows for the creation of a dense web of information connecting people, projects, tools, and concepts. This Openclaw Skills extension ensures that knowledge is never isolated, transforming flat markdown files into a navigable and queryable graph.
Beyond simple linking, MindGraph introduces the concept of MindSkills. These are repeatable, learned processes that the agent can execute to perform specific analytical tasks—such as SEO validation or competitor analysis. The results are automatically saved as graph-connected markdown files, complete with metadata and bidirectional links, ensuring that every insight is contextualized within the broader workspace environment.
MindGraph Use Cases
- Creating a structured knowledge base where people, projects, and concepts are interconnected via [[wikilinks]].
- Learning and executing repeatable frameworks like the knockout test or SEO validator.
- Querying the workspace to find backlinks, forward links, and bidirectional connections for deep context retrieval.
- Maintaining workspace hygiene by identifying orphan files, dead links, and unconnected nodes.
- Proactively suggesting specialized workflows based on the user's current task or subject matter.
How MindGraph Works
- The agent uses [[wikilinks]] as a standard syntax whenever writing or editing markdown files in the workspace.
- A Python indexing script scans the files to build a comprehensive map of relationships, stored in a mindgraph.json file.
- The user or agent triggers specific Graph Commands to query connections, visualize trees, or identify disconnected content.
- For repeatable tasks, the agent 'learns' a MindSkill by creating a process definition in a dedicated directory.
- Upon execution, the agent follows the defined PROCESS.md, generates a result file with YAML metadata, and updates the global graph index.
MindGraph Setup
To begin using this skill, ensure the mindgraph script is in the correct directory and initialize your index. Use the following commands to manage your Openclaw Skills workspace graph:
# Build or rebuild the graph index
python3 skills/mindgraph/scripts/mindgraph.py index
# List all currently learned MindSkills
python3 skills/mindgraph/scripts/mindgraph.py skills
MindGraph Data Schema & Taxonomy
MindGraph organizes data through a centralized index and a structured directory for learned processes. The taxonomy is designed for high discoverability within the Openclaw Skills ecosystem:
| Component | Location | Description |
|---|---|---|
| Graph Index | mindgraph.json |
The root-level JSON file mapping all wikilink connections. |
| MindSkills | skills/mindgraph/mindskills/ |
Directory for process definitions and historical results. |
| Result Files | .../results/<subject>.md |
Markdown files containing task outcomes and YAML frontmatter. |
| Core Script | skills/mindgraph/scripts/ |
The Python logic for indexing and querying the knowledge graph. |
MindGraph Advanced Features
- Case-insensitive link resolution across file basenames, project directories, and result files.
- YAML alias support allowing multiple names to resolve to a single knowledge node.
- ASCII tree visualization for exploring multi-level relationship hierarchies.
- Automated health checks including detection of 'lonely' files and broken links.
- Extensible MindSkill architecture allowing users to teach the agent new, complex workflows on the fly.
SKILL.md
Loading
Related Openclaw Skills
Free Groq Voice Recognition
A high-performance, cost-free voice recognition skill leveraging Groq's Whisper API for near-instant speech-to-text in 50+ languages.
Intent Guardian
An intelligent focus companion that monitors desktop activity to maintain a real-time task stack and remind you of forgotten work after interruptions.
Tavily Search
A high-performance web search skill optimized specifically for AI agents to retrieve clean, relevant real-time data.
ADMET Prediction
Accelerate drug discovery by predicting ADMET properties and safety risks for chemical compounds using advanced ML models.
XMTP Agents
A comprehensive toolkit for building, configuring, and extending event-driven messaging agents on the XMTP network.
XMTP CLI
A powerful command-line interface for interacting with the XMTP network to manage secure messaging, groups, and debugging workflows.
Featured*
