ZeroClaw is a Rust-based autonomous AI agent infrastructure that acts as the runtime OS for agentic workflows, with swappable subsystems and CLI tooling for configuration, deployment, and migration from other frameworks.

How do I bootstrap and start using ZeroClaw?

Install ZeroClaw, run zeroclaw onboard --interactive to configure providers and channels, edit config.toml, then start with zeroclaw daemon or use zeroclaw agent for quick tests.

How can I migrate from another agent framework?

Use zeroclaw migrate openclaw --dry-run to preview changes, then proceed with migration steps to move providers, channels, and memory backends into ZeroClaw’s config and runtime.

zeroclaw

npx machina-cli add skill Lu1sDV/skillsmd/zeroclaw --openclaw

Files (1)

SKILL.md

6.0 KB

ZeroClaw

Rust-based autonomous AI assistant infrastructure. Runtime OS for agentic workflows — deploy anywhere, swap anything. Runs on <5MB RAM, <10ms startup, ~8.8MB binary.

Quick Reference

Command	Purpose
`zeroclaw onboard --interactive`	First-time setup wizard
`zeroclaw agent -m "msg"`	Single message
`zeroclaw agent`	Interactive REPL
`zeroclaw daemon`	Full autonomous runtime + channels
`zeroclaw gateway`	Webhook server (default :42617)
`zeroclaw status`	Check daemon & config
`zeroclaw doctor`	System diagnostics
`zeroclaw channel doctor`	Channel health check
`zeroclaw channel bind-telegram <id>`	Bind Telegram chat
`zeroclaw providers`	List available providers
`zeroclaw auth login --provider <p> --device-code`	OAuth login
`zeroclaw auth setup-token --provider <p>`	Token auth
`zeroclaw service install`	Install as system service
`zeroclaw migrate openclaw --dry-run`	Preview migration
`zeroclaw completions <shell>`	Shell completions

Installation

# Homebrew
brew install zeroclaw

# Bootstrap (local clone)
git clone https://github.com/zeroclaw-labs/zeroclaw.git && cd zeroclaw && ./bootstrap.sh

# Remote bootstrap
curl -fsSL https://raw.githubusercontent.com/zeroclaw-labs/zeroclaw/main/scripts/bootstrap.sh | bash

# From source
cargo build --release --locked && cargo install --path . --force --locked

# Pre-built binaries: GitHub Releases (Linux x86_64/aarch64/armv7, macOS, Windows)
# Bootstrap flags: --prefer-prebuilt, --prebuilt-only, --docker

Build requirements: 2GB+ RAM+swap, 6GB+ disk. Runtime: <5MB.

Configuration

Edit ~/.zeroclaw/config.toml. Core sections:

[providers]
# OpenAI, Anthropic, OpenRouter, Google, Mistral, xAI, DeepSeek, 70+ more
# Set via env: OPENAI_API_KEY, ANTHROPIC_API_KEY, etc.

[memory]
backend = "sqlite"  # "postgres", "markdown", "lucid", "none"
# Built-in vector search (BLOBs + cosine similarity) + FTS5 keyword search
# No external vector DB needed — hybrid merge of vector + BM25

[runtime]
kind = "native"  # "docker" for sandboxed execution

[channels]
# CLI, Telegram, Discord, Slack, Matrix, Signal, IRC, Email, iMessage,
# WhatsApp, Mattermost, Lark, DingTalk, QQ, Linq, Nostr, Webhook

[security]
# allowlists, pairing, rate_limits, filesystem_scoping

[autonomy]
level = "full"                          # "supervised" (default) or "full"
max_actions_per_hour = 999999           # ⚠️ MUST be > 0 or daemon silently fails
max_cost_per_day_cents = 999999         # same
require_approval_for_medium_risk = false
block_high_risk_commands = false

Architecture

Trait-driven — every subsystem is swappable:

Layer	Swappable Options
Providers	OpenAI, Anthropic, OpenRouter, Google, custom endpoints
Channels	18+ (CLI, Telegram, Discord, Slack, Matrix, etc.)
Memory	SQLite (default), PostgreSQL, Markdown, Lucid, none
Tools	Shell, file ops, Git, HTTP, screenshots, browser
Runtime	Native, Docker sandbox

Authentication

OPENAI_API_KEY=sk-... zeroclaw agent                              # env var
zeroclaw auth login --provider openai-codex --device-code         # OAuth device flow
zeroclaw auth paste-token --provider anthropic --profile default  # token paste

Profiles stored encrypted at ~/.zeroclaw/auth-profiles.json (key: ~/.zeroclaw/.secret_key). Manage with auth use, auth status, auth refresh (see Quick Reference).

WARNING (Feb 2026): Claude Code OAuth tokens (Free/Pro/Max) are restricted to Claude Code/Claude.ai only. Do NOT use them via ZeroClaw — violates Anthropic ToS.

Memory System

Built-in full-stack search engine (no external vector DB):

Vector: Embeddings as BLOBs in SQLite, cosine similarity
Keyword: FTS5 virtual tables, BM25 scoring
Hybrid: Weighted merge of vector + keyword results
Embeddings: OpenAI, custom URL, or noop
Chunking: Line-based markdown with heading preservation
Cache: SQLite LRU embedding cache
Reindex: Atomic FTS5 rebuild + missing vector re-embedding

The agent manages memory automatically via built-in tools.

Common Patterns

Deploy as daemon with Telegram

zeroclaw onboard --interactive          # configure provider + channel
zeroclaw channel bind-telegram 123456   # bind chat ID
zeroclaw daemon                         # start autonomous runtime

Webhook gateway

zeroclaw gateway --port 8080            # or --port 0 for random

Docker sandbox

# config.toml
[runtime]
kind = "docker"

Migration from OpenClaw

zeroclaw migrate openclaw --dry-run     # preview changes
zeroclaw migrate openclaw               # execute migration

Troubleshooting

Issue	Fix
Build OOM	Use `--prefer-prebuilt` or add swap (need 2GB+)
Channel not responding	`zeroclaw channel doctor`
Auth expired	`zeroclaw auth refresh --provider <p>`
Config issues	`zeroclaw doctor`
Gateway unreachable	Check port binding, firewall, pairing config
Memory search poor	Check embedding provider config, run reindex
Config schema JSON parse fails	`zeroclaw config schema` mixes INFO log lines with JSON — save to file and strip the log line first
Config changes not applied	`zeroclaw service restart` (or `zeroclaw service stop && zeroclaw service start`) then `zeroclaw doctor`

Security

Supports gateway pairing, Docker sandboxing, allowlists (tools/files/channels), rate limiting, filesystem scoping, and encrypted secrets at rest.

Source

git clone https://github.com/Lu1sDV/skillsmd/blob/main/zeroclaw/SKILL.mdView on GitHub

Overview

ZeroClaw is a Rust-based autonomous AI agent infrastructure that serves as the runtime OS for agentic workflows. It is designed to deploy anywhere with a tiny footprint: <5MB RAM, startup <10ms, and an ~8.8MB binary. Its trait-driven, swappable architecture (providers, channels, memories, runtime) is paired with CLI tooling, Docker/native runtimes, and migration paths from other agent frameworks.

How This Skill Works

ZeroClaw uses a trait-driven architecture where subsystems such as providers, channels, memory backends, tools, and runtime are swappable. Core configuration lives in ~/.zeroclaw/config.toml, and you interact with the system using commands like onboard, agent, daemon, and channel bind-telegram. It supports multiple channels, authentication flows, and migration workflows to move from other frameworks.

When to Use It

You are setting up a new ZeroClaw deployment: selecting providers, channels, and memory backends via config.toml.
You want to run a full autonomous runtime with channels and system services (daemon and gateway) for ongoing agent operations.
You need to migrate from another agent framework (e.g., using zeroclaw migrate openclaw --dry-run) to ZeroClaw.
You are diagnosing infra issues or channel health using doctor and channel doctor utilities.
You want to test or iterate quickly with an interactive session or a bound channel (e.g., zeroclaw agent or zeroclaw gateway).

Quick Start

Step 1: Install zeroclaw (brew install zeroclaw; or bootstrap from source as described in the docs).
Step 2: Run onboarding to configure initial providers, channels, and auth: zeroclaw onboard --interactive.
Step 3: Edit ~/.zeroclaw/config.toml to set [providers], [memory], [runtime], and [channels], then start the daemon with zeroclaw daemon and test with zeroclaw agent or zeroclaw gateway.

Best Practices

Start with a minimal, secure provider setup and gradually add channels; keep sensitive keys in environment variables or encrypted auth profiles.
Define memory backends explicitly (sqlite by default) and align with your data needs (BM25-like search, FTS).
Edit ~/.zeroclaw/config.toml carefully and validate with zeroclaw status before starting the daemon.
Run onboarding first (zeroclaw onboard --interactive) to complete initial setup, then bind channels (e.g., zeroclaw channel bind-telegram) for real use.
Regularly run zeroclaw doctor and zeroclaw channel doctor to detect misconfigurations and health issues early.

Example Use Cases

Deploy a OpenAI provider with a Telegram channel and run as a system service using zeroclaw service install.
Migrate an existing OpenAI-based agent setup from another framework via zeroclaw migrate openclaw --dry-run, then execute migration steps.
Operate a local sandboxed workflow by running zeroclaw daemon with Docker runtime for isolated channel execution.
Bind a Telegram chat and test one-shot messages with zeroclaw agent -m 'hello' for quick validation.
Configure OAuth login with zeroclaw auth login --provider openai-codex --device-code and manage profiles in ~/.zeroclaw/auth-profiles.json.

Frequently Asked Questions

Add this skill to your agents