ClaudeClaw logo

ClaudeClaw

Persistent agent orchestrator for Claude Code.
Multi-channel. Sandboxed. 8K lines you can actually read.

<10ms cold start ~8K lines No Docker
$ git clone https://github.com/sbusso/claudeclaw.git && cd claudeclaw
$ claude
> /setup
ClaudeClaw architecture overview

How does it compare?

Three approaches to the same problem — different trade-offs.

ClaudeClaw NanoClaw OpenClaw
Codebase ~8K lines ~8K lines ~400K lines
Security model Kernel-enforced sandbox Docker isolation Opt-in sandboxing
Cold start <10ms ~2-5s ~2-5s native
Docker required No Yes Optional
Auditability Fits in context window Read in 8 minutes Open source, complex
Configuration Code changes (git revert) Code changes JSON config + CLI
Extensions Pluggable system None Skills ecosystem
Memory Structured + searchable Basic MEMORY.md + SOUL.md

ClaudeClaw is built on NanoClaw, adding sandbox isolation, extensions, structured memory, and multi-channel routing.

How it works

Single process. Channels in, sandboxed agents out.

ClaudeClaw message flow

Sandbox-first security

The kernel enforces isolation, not your application code.

Kernel isolation

Seatbelt (macOS) / bubblewrap (Linux). Not app-level permission checks — OS-level enforcement.

Network restriction

Agents can only reach api.anthropic.com. Credentials passed directly — nowhere to exfiltrate.

Zero overhead

<10ms cold starts. No VMs, no containers, no daemon. Just an ordinary process with restrictions.

Per-group boundaries

Each group's agent sees only its own files. Family chat can't read your work directory. Kernel-enforced.

Sandbox security model

A tool, not a platform

Small enough to understand

~8K lines. One process. The entire codebase fits in Claude's context window for self-improvement.

Built for one person

Not an enterprise framework. Software that fits each user's exact needs. Fork it, modify it, own it.

Code is configuration

No YAML sprawl. Want different behavior? Tell Claude to change the code. Every change is a commit you can revert.

Features

Multi-channel routing

Slack, WhatsApp, Telegram, Discord, Gmail. Channels self-register — the orchestrator connects whichever ones have credentials.

Multi-channel routing

Structured memory

Daily logs, topic files, long-term CLAUDE.md, searchable archive. Auto-saved before context compaction. Upgradeable to QMD semantic search.

Structured memory system

Three trigger modes

Channel messages (@mention), scheduled tasks (cron), webhook triggers (HMAC-SHA256). CI failures and monitoring alerts can invoke agents directly.

Three trigger modes

Cost tracking & limits

Per-run logging with token usage and estimated cost. Per-group budget caps. Model selection per group — use Haiku for triage, Opus for deep work.

Cost tracking dashboard

Composable extensions

Install only what you use. No merge conflicts. Core updates are git pull.

$ claude
> /install-extension slack
Cloning claudeclaw-slack... compiling... installing deps... copying skills...
> /install-extension triage
Cloning claudeclaw-triage... compiling... installing deps... copying skills...
Channel

claudeclaw-slack

Socket Mode, threads, reactions

Agent

claudeclaw-triage

Bug triage + SWE task queue

Planned

claudeclaw-discord

Discord channel

Planned

claudeclaw-gmail

Gmail channel

Planned

claudeclaw-memory

QMD semantic memory

Planned

claudeclaw-obsidian

Obsidian vault sync

Planned

claudeclaw-tui

Terminal dashboard

Get started in 30 seconds

macOS or Linux. Node.js 20+. Claude Code.

$ git clone https://github.com/sbusso/claudeclaw.git
$ cd claudeclaw
$ claude
> /setup
View on GitHub →