clawphify 0.1.0

OpenClaw framework knowledge graph injector - new Claw projects instantly understand the framework structure

clawphify 🦞

OpenClaw Framework Knowledge Graph Injector

🍴 Fork Notice: This is a specialized fork of graphify by @safishamsi. While graphify is a general-purpose knowledge graph builder, clawphify focuses specifically on pre-built OpenClaw framework knowledge for instant project understanding. New OpenClaw projects immediately understand the framework structure—no need to RAG the entire source code.

Why clawphify?

When you start a new project based on OpenClaw:

Without clawphify:

1. AI reads OpenClaw source files to understand structure
2. Consumes thousands of tokens
3. May miss architectural patterns

With clawphify:

1. Pre-built knowledge graph loaded instantly
2. Zero token cost for framework understanding
3. Structured, accurate framework knowledge

Quick Start

# Install
pip install clawphify

# Show built-in OpenClaw knowledge
clawphify info

# Initialize your project with framework graph
clawphify init

# Query the framework
clawphify query "how does channel routing work?"

What You Get

Pre-built Framework Knowledge

Core System:

• Gateway architecture and session management
• Pi agent and multi-agent routing
• Channel ecosystem (20+ channels)
• Web Control UI

Source Structure:

src/agents     → Agent system, tools, sandbox
src/gateway    → Gateway core, routing
src/channels   → Channel implementations  
src/cli        → CLI commands
src/plugin-sdk → Plugin development SDK
src/security   → Tokens, allowlists
src/tasks      → Task management

CLI Commands:

• openclaw onboard --install-daemon
• openclaw dashboard
• openclaw agent
• openclaw nodes

Channels:

• Built-in: Discord, Telegram, WhatsApp, Signal, Slack, iMessage, WebChat
• Plugin: Matrix, Zalo, Nostr, Twitch, IRC, Mattermost, Feishu

Knowledge Communities

The framework is pre-clustered into 8 communities:

1. Core System - gateway, agent, config
2. Channel Ecosystem - all channel integrations
3. Agent System - runtime, routing, skills
4. CLI & Interface - commands, TUI, web UI
5. Plugin System - SDK, extensions
6. Infrastructure - utils, security, sessions
7. Media Processing - images, audio, docs
8. Documentation - all docs hubs

Usage Patterns

New OpenClaw Project

# In your new project directory
clawphify init

# Now ask framework questions:
# "What files should I modify to add a channel?"
# "How does authentication work?"
# "Where is the session config?"

OpenClaw Development

If you're modifying OpenClaw itself, use graphify directly:

graphify . --update

This builds a project-specific graph including your changes.

Installation

pip install clawphify

# Install skill for your AI assistant
clawphify install                    # Claude Code
clawphify install --platform claw    # OpenClaw
clawphify install --platform codex   # Codex

Commands

Command	Description
clawphify info	Show built-in OpenClaw framework info
clawphify init	Copy built-in graph to clawphify-out/
clawphify query "Q"	Query the framework graph
clawphify install	Install skill for AI assistant
clawphify claude install	Configure for Claude Code
clawphify claw install	Configure for OpenClaw

Example Queries

# Architecture questions
clawphify query "how does gateway manage sessions?"
clawphify query "what is the relationship between channels and plugins?"

# Implementation guidance
clawphify query "where should I add authentication logic?"
clawphify query "how do I create a new channel?"

# Configuration help
clawphify query "what are the key config settings?"
clawphify query "how do I configure WhatsApp channel?"

Integration with graphify

clawphify uses graphify's graph format. You can use graphify tools:

# Merge with project-specific knowledge
graphify . --update --graph clawphify-out/openclaw_graph.json

# Use graphify's query on clawphify graph
graphify query "channels" --graph clawphify-out/openclaw_graph.json

How It Works

1. Built-in Graph: Contains OpenClaw v3.22.0 framework structure
2. JSON Format: Compatible with graphify's graph.json format
3. Node Types: system, module, feature, command, config, docs
4. Edges: EXTRACTED (from source) and INFERRED (relationships)
5. Communities: Pre-clustered using Leiden algorithm

Graph Contents

• 50+ nodes: Core components, modules, features, docs
• 30+ edges: Relationships and dependencies
• 8 communities: Logical groupings
• Quick reference: Getting started, key directories, config essentials

Requirements

• Python 3.10+
• networkx
• tree-sitter

License

MIT - Same as OpenClaw and graphify

Credits

• Built on graphify by @safishamsi
• OpenClaw framework by the OpenClaw team
• Graph extraction from OpenClaw v3.22.0

---

For new Claw: Know thy framework instantly. 🦞