Install OpenClaw too difficult for you?

Try shipclaw.app and deploy OpenClaw in 30 seconds.

2026/02/11

OpenClaw Cheatsheet 2026

Complete reference guide for OpenClaw — 150+ CLI commands, configuration, workspace management, and troubleshooting

Getting Started

Everything you need to install, configure, and connect your first OpenClaw agent — from CLI basics and global flags to channel pairing and model authentication.

Core CLI Commands

Gateway Operations

openclaw gateway                        # Run WebSocket Gateway server
openclaw gateway --port 8080 --bind 0.0.0.0 --token <tok>
openclaw gateway start | stop | restart # Manage via launchd/systemd

Channel Management

openclaw channels login                 # WhatsApp QR pairing
openclaw channels add --channel <ch> --token <tok>  # Add bot
openclaw channels status --probe        # Check channel health

Configuration & Setup

openclaw onboard                        # Interactive setup wizard
openclaw onboard --install-daemon       # Setup + install daemon
openclaw doctor                         # Health checks & quick fixes
openclaw doctor --deep --yes            # Deep scan with auto-fix
openclaw config get <key>               # Read config (JSON5)
openclaw config set <key> <value>       # Write config
openclaw config unset <key>             # Remove config key

Model Management

openclaw models list --all              # View all available models
openclaw models set <model>             # Set primary model
openclaw models set-image <model>       # Set default image model
openclaw models status --probe          # Live probe auth profiles
openclaw models auth setup-token        # Preferred Anthropic auth
openclaw models auth add --provider <p> # Add provider API key
openclaw models fallbacks add <model>   # Add to fallback chain
openclaw models fallbacks remove <model>
openclaw models aliases add <alias> <model>

Memory & Search

openclaw memory status                  # Check memory system
openclaw memory index --all             # Reindex all memory files
openclaw memory search "query"          # Semantic vector search

Diagnostics

openclaw logs --follow                  # Tail Gateway logs (colorized)
openclaw logs --json                    # JSON output (one event/line)
openclaw logs --limit 200               # Limit log lines
openclaw status --all --deep            # Full diagnosis
openclaw sessions --json                # List conversation sessions

Security & Advanced

openclaw security audit                 # Audit config vulnerabilities
openclaw security audit --fix           # Auto-fix issues
openclaw reset --scope <scope>          # Reset: config | credentials | sessions | full

Global Flags

Flag	Description
`--dev`	Isolate state under `~/.openclaw-dev`, shift ports
`--profile <name>`	Isolate state under `~/.openclaw-<name>`
`--no-color`	Disable ANSI colors (respects `NO_COLOR=1`)
`--update`	Shorthand for `openclaw update` (source installs only)
`-V, --version, -v`	Print version and exit

Channel Setup

openclaw channels login
# Scan the QR code displayed in the terminal

# Get a bot token from @BotFather first
openclaw channels add --channel telegram --token $TELEGRAM_BOT_TOKEN

# Create a bot in Discord Developer Portal
openclaw channels add --channel discord --token $DISCORD_BOT_TOKEN

# Configure a Slack App with bot token
openclaw channels add --channel slack

macOS native bridge — no additional token required. Enabled automatically on macOS.

Google Chat — Service account setup
Signal — Linked device connection
MS Teams — Bot registration

openclaw channels add --channel <platform> --token <token>
openclaw channels status --probe  # Verify connectivity

Models & Authentication

Auth Setup

Set Up Auth Token (Recommended)

openclaw models auth setup-token   # Preferred Anthropic auth

Or Add Provider Key

openclaw models auth add --provider openai
openclaw models auth add --provider anthropic

Verify

openclaw models status --probe     # Live probe configured profiles

Model Configuration

openclaw models list --all          # View available models
openclaw models set claude-sonnet   # Set primary model
openclaw models set-image dall-e-3  # Set image model
openclaw models fallbacks add gpt-4 # Add fallback
openclaw models aliases add fast claude-haiku  # Create alias

Failover Schedule

Model failover timing: 1 min → 5 min → 1 hour

Workspace & Memory

OpenClaw stores agent personality, user preferences, and long-term knowledge in Markdown workspace files and a hybrid vector/BM25 memory system.

Workspace Files

All workspace files are located at ~/.openclaw/workspace/:

File	Purpose
AGENTS.md	Operating instructions for the agent
SOUL.md	Persona, tone, and boundaries
USER.md	User information and preferences
IDENTITY.md	Agent name, emoji, theme
MEMORY.md	Curated long-term memory (DM only)
TOOLS.md	Local tool notes
HEARTBEAT.md	Heartbeat checklist
BOOT.md	Startup checklist
`memory/YYYY-MM-DD.md`	Daily append-only logs

Tip

AGENTS.md and SOUL.md are the most important files — they define how your agent behaves and responds.

Memory System

Daily Logs

memory/YYYY-MM-DD.md — Append-only logs. Reads today + yesterday at session start.

Long-Term Memory

MEMORY.md — Curated facts, loaded only in main DM session.

Vector Search

memory_search tool provides semantic search (~400 tokens per result).

Providers — Auto-selects from: local GGUF → OpenAI → Gemini → Voyage

Search Configuration

Feature	Details
Hybrid Search	Default 0.7 vector / 0.3 BM25 weighting
QMD Backend	Optional experimental mode with BM25 + vectors + reranking

openclaw memory index --all       # Reindex everything
openclaw memory search "query"    # Semantic search

Sessions & Commands

Control how conversations are scoped, reset, and managed through built-in slash commands and text-to-speech output.

Sessions

DM Scope Options

Scope	Description
`main`	Single shared session (default)
`per-peer`	One session per contact
`per-channel-peer`	One per contact per channel
`per-account-channel-peer`	Fully isolated sessions

Reset Modes

Mode	Description
`daily`	Resets at 4am local time (default)
`idle`	Resets after idle timeout

Configuration

{
  "session": {
    "dmScope": "main",
    "reset": {
      "idleMinutes": 30            // Sliding idle window
    },
    "resetByType": { ... },        // Override for dm, group, thread
    "resetByChannel": { ... },     // Per-channel override (highest priority)
    "identityLinks": { ... },      // Map provider:id to canonical identity
    "sendPolicy": { ... },         // Block delivery for specific types
    "store": "~/.openclaw/agents/{agentId}/sessions/sessions.json"
  }
}

Security

Use per-channel-peer for multi-user inboxes to prevent context leakage between users.

Slash Commands

Session Management

Command	Description
`/status`	Session health, context usage, credential status
`/context list`	What's in context window (largest contributors)
`/context detail`	Full system prompt and injected workspace files
`/model <model>`	Switch model for session
`/model list`	List available models
`/compact [instructions]`	Summarize older context, free window space
`/new [model]`	Start fresh session (optional model switch)
`/reset`	Alias for `/new`
`/stop`	Abort current run, clear queued followups

Delivery & Output

Command	Description
`/send on\|off\|inherit`	Override delivery for this session
`/tts on\|off`	Toggle text-to-speech
`/think`	Toggle reasoning mode
`/verbose`	Toggle verbose mode

Configuration

Command	Description
`/config`	Persisted config changes
`/debug`	Runtime-only overrides (requires `commands.debug: true`)

Text-to-Speech

Provider	Characteristics
ElevenLabs	Premium ultra-realistic, higher latency
OpenAI	Standard fast, high-quality voices
Edge TTS	Free, no API key, multi-language support

Enable auto-TTS:

{
  "messages": {
    "tts": {
      "auto": "always"
    }
  }
}

Agent Architecture

Scale beyond a single agent — route messages to specialized agents, extend capabilities with skills, delegate tasks to sub-agents, and keep everything alive with heartbeats.

Multi-Agent Routing

Routing Precedence

Priority	Binding	Description
1 (highest)	`peer`	Exact DM/group id
2	`guildId`	Discord server
3	`teamId`	Slack workspace
4	`accountId`	Account-level
5	`channel`	Channel type
6 (lowest)	`default`	Final fallback agent

Management

openclaw agents add <name>                # Add new agent
openclaw agents list --bindings           # List with routing bindings
openclaw agents delete <name>             # Remove agent

Each agent gets:

Isolated workspace
Per-agent auth profiles
Session store at ~/.openclaw/agents/<id>/sessions/

Skills System

Skill Precedence

Agent Skills (Highest)

<workspace>/skills/ — Per-agent, takes priority over all others.

Managed/Local Skills

~/.openclaw/skills/ — Shared across workspaces.

Bundled Skills (Lowest)

Shipped with OpenClaw — default capabilities.

ClawHub Registry

clawhub install <slug>                    # Install from ClawHub
clawhub update --all                      # Update all installed
clawhub sync --all                        # Scan and publish updates

SKILL.md Format

---
name: my-skill
description: "What this skill does"
metadata: { "openclaw": { "requires": { ... } } }
---

Sub-Agents

Sub-agents enable parallel work for research/long tasks without blocking the main conversation.

Own session key with optional sandbox isolation
Auto-announce results to requester chat channel
Auto-archive after 60 minutes (configurable)

Commands

Command	Description
`/subagents list`	List active sub-agents
`/subagents stop <id\|#\|all>`	Stop sub-agent runs
`/subagents log <id\|#>`	View sub-agent logs
`/subagents info <id\|#>`	Show run metadata
`/subagents send <id\|#> <msg>`	Send message to sub-agent

Spawn Tool — The sessions_spawn tool accepts:

{
  "task": "Research topic X",
  "label": "research-x",
  "model": "claude-sonnet",
  "thinking": true,
  "runTimeoutSeconds": 300,
  "cleanup": true
}
// Returns: { status, runId, childSessionKey }

Heartbeat System

{
  "heartbeat": {
    "every": "30m",                 // Interval (1h for Anthropic OAuth)
    "target": "last",               // last | none | <channel id>
    "to": null,                     // Optional recipient override
    "model": null,                  // Model override for heartbeat runs
    "prompt": null,                 // Custom prompt body
    "activeHours": {
      "start": "09:00",
      "end": "22:00",
      "timezone": "America/New_York"
    }
  }
}

Contract

Agent replies HEARTBEAT_OK if nothing needs attention. OK-only replies are automatically stripped and dropped.

Automation & Tools

Automate browser tasks, schedule cron jobs, and wire up lifecycle hooks to customize how your agent responds to events.

Browser Automation

openclaw browser start | stop             # Headless instance
openclaw browser tabs                     # List open pages
openclaw browser open <url>               # Open URL in new tab
openclaw browser navigate <url>           # Navigate current tab
openclaw browser screenshot               # Capture active view
openclaw browser click | type | press     # DOM interactions
openclaw browser evaluate <js>            # Run JavaScript
openclaw browser pdf                      # Export page as PDF

Cron Jobs

openclaw cron list                        # View scheduled jobs
openclaw cron add                         # Create new job
openclaw cron edit <id>                   # Edit existing job
openclaw cron enable | disable <id>       # Toggle job
openclaw cron run <id>                    # Manual trigger
openclaw cron runs                        # View run history

Hooks & Automation

Bundled Hooks

Hook	Description
`session-memory`	Save session context to memory on `/new`
`command-logger`	Log all commands to audit file
`boot-md`	Run `BOOT.md` on gateway start
`soul-evil`	Swap `SOUL.md` during purge window

Hook Commands

openclaw hooks list                       # List discovered hooks
openclaw hooks enable <name>              # Enable a hook
openclaw hooks disable <name>             # Disable a hook
openclaw hooks info <name>                # Show hook details
openclaw hooks check                      # Check eligibility

Event Types

Event	Trigger
`command:new`	When `/new` is issued
`command:reset`	When `/reset` is issued
`command:stop`	When `/stop` is issued
`gateway:startup`	After channels start
`agent:bootstrap`	Before workspace files injected

Operations & Security

Advanced commands for plugin management, logging, sandbox isolation, and system diagnostics.

Power Commands

Plugin & System Management

openclaw plugins list | enable | disable          # Manage plugins
openclaw approvals get | set | allowlist           # Execution approval policy
openclaw sandbox list | recreate | explain         # Inspect/rebuild sandbox
openclaw system event --text "X"                   # Queue system event
openclaw system heartbeat enable | disable | last  # Control heartbeat
openclaw update status | wizard                    # Release channel updates

Node & Device Management

openclaw nodes list | pending | approve            # Remote node management
openclaw devices list | approve | rotate | revoke  # Device token lifecycle
openclaw directory peers | groups list             # Resolve IDs

Monitoring & Dashboard

openclaw health --json --verbose        # Direct gateway health probe
openclaw dashboard                      # Open Control UI
openclaw tui --url <url> --token <tok>  # Remote terminal UI

Logging & Diagnostics

openclaw logs --follow                    # Tail logs (colorized in TTY)
openclaw logs --json                      # Line-delimited JSON
openclaw logs --limit 200                 # Limit output
openclaw channels logs --channel whatsapp # Channel-specific logs

OpenTelemetry Export

{
  "diagnostics": {
    "otel": {
      "enabled": true
    }
  }
}

Sandboxing

Mode Options

Mode	Description
`"off"`	No sandboxing, tools run on host
`"non-main"`	Sandbox only non-main sessions (default)
`"all"`	Every session runs in sandbox

Scope Options

Scope	Description
`"session"`	One container per session (default)
`"agent"`	One container per agent
`"shared"`	One container for all sandboxed sessions

Workspace Access

Access	Description
`"none"`	Tools see sandbox workspace only (default)
`"ro"`	Read-only mount at `/agent`
`"rw"`	Read/write mount at `/workspace`

Setup:

scripts/sandbox-setup.sh  # Creates openclaw-sandbox:bookworm-slim image

Reference

Quick-access paths and common troubleshooting steps.

Key Paths

Path	Description
`~/.openclaw/openclaw.json`	Main configuration file
`~/.openclaw/workspace/`	Default agent workspace
`~/.openclaw/agents/<id>/`	Per-agent state directory
`~/.openclaw/agents/<id>/sessions/`	Session store & transcripts
`~/.openclaw/credentials/`	OAuth/API keys
`~/.openclaw/memory/<agentId>.sqlite`	Vector index store
`/tmp/openclaw/openclaw-YYYY-MM-DD.log`	Gateway log file

Tip

Use --dev or --profile <name> to isolate all state under a separate directory for testing.

Troubleshooting

Universal Fix

When in doubt, run openclaw doctor --deep --yes — it performs health checks, quick fixes, and system service scans.

All Posts

Related Use Cases

Productivity

Autonomous Project Management with Subagents

Decentralized project coordination where subagents work autonomously on tasks, coordinating through shared STATE.yaml files rather than a central orchestrator.

AdvancedPro Plan

Social Media

Daily Reddit Digest

Run a daily digest of the top performing posts from your favourite subreddits, with preference learning over time.

BeginnerStarter Plan

Social Media

Daily YouTube Digest

Start your day with a personalized summary of new videos from your favorite YouTube channels — no more missing content from creators you actually want to follow.

BeginnerStarter Plan

Join the community

Subscribe to our newsletter for the latest news and updates

Install OpenClaw too difficult for you?

Try shipclaw.app and deploy OpenClaw in 30 seconds.

Deploy in 30s Homepage

2026/02/11

OpenClaw Cheatsheet 2026

Complete reference guide for OpenClaw — 150+ CLI commands, configuration, workspace management, and troubleshooting

Getting Started

Everything you need to install, configure, and connect your first OpenClaw agent — from CLI basics and global flags to channel pairing and model authentication.

Core CLI Commands

Gateway Operations

openclaw gateway                        # Run WebSocket Gateway server
openclaw gateway --port 8080 --bind 0.0.0.0 --token <tok>
openclaw gateway start | stop | restart # Manage via launchd/systemd

Channel Management

openclaw channels login                 # WhatsApp QR pairing
openclaw channels add --channel <ch> --token <tok>  # Add bot
openclaw channels status --probe        # Check channel health

Configuration & Setup

openclaw onboard                        # Interactive setup wizard
openclaw onboard --install-daemon       # Setup + install daemon
openclaw doctor                         # Health checks & quick fixes
openclaw doctor --deep --yes            # Deep scan with auto-fix
openclaw config get <key>               # Read config (JSON5)
openclaw config set <key> <value>       # Write config
openclaw config unset <key>             # Remove config key

Model Management

openclaw models list --all              # View all available models
openclaw models set <model>             # Set primary model
openclaw models set-image <model>       # Set default image model
openclaw models status --probe          # Live probe auth profiles
openclaw models auth setup-token        # Preferred Anthropic auth
openclaw models auth add --provider <p> # Add provider API key
openclaw models fallbacks add <model>   # Add to fallback chain
openclaw models fallbacks remove <model>
openclaw models aliases add <alias> <model>

Memory & Search

openclaw memory status                  # Check memory system
openclaw memory index --all             # Reindex all memory files
openclaw memory search "query"          # Semantic vector search

Diagnostics

openclaw logs --follow                  # Tail Gateway logs (colorized)
openclaw logs --json                    # JSON output (one event/line)
openclaw logs --limit 200               # Limit log lines
openclaw status --all --deep            # Full diagnosis
openclaw sessions --json                # List conversation sessions

Security & Advanced

openclaw security audit                 # Audit config vulnerabilities
openclaw security audit --fix           # Auto-fix issues
openclaw reset --scope <scope>          # Reset: config | credentials | sessions | full

Global Flags

Flag	Description
`--dev`	Isolate state under `~/.openclaw-dev`, shift ports
`--profile <name>`	Isolate state under `~/.openclaw-<name>`
`--no-color`	Disable ANSI colors (respects `NO_COLOR=1`)
`--update`	Shorthand for `openclaw update` (source installs only)
`-V, --version, -v`	Print version and exit

Channel Setup

openclaw channels login
# Scan the QR code displayed in the terminal

# Get a bot token from @BotFather first
openclaw channels add --channel telegram --token $TELEGRAM_BOT_TOKEN

# Create a bot in Discord Developer Portal
openclaw channels add --channel discord --token $DISCORD_BOT_TOKEN

# Configure a Slack App with bot token
openclaw channels add --channel slack

macOS native bridge — no additional token required. Enabled automatically on macOS.

Google Chat — Service account setup
Signal — Linked device connection
MS Teams — Bot registration

openclaw channels add --channel <platform> --token <token>
openclaw channels status --probe  # Verify connectivity

Models & Authentication

Auth Setup

Set Up Auth Token (Recommended)

openclaw models auth setup-token   # Preferred Anthropic auth

Or Add Provider Key

openclaw models auth add --provider openai
openclaw models auth add --provider anthropic

Verify

openclaw models status --probe     # Live probe configured profiles

Model Configuration

openclaw models list --all          # View available models
openclaw models set claude-sonnet   # Set primary model
openclaw models set-image dall-e-3  # Set image model
openclaw models fallbacks add gpt-4 # Add fallback
openclaw models aliases add fast claude-haiku  # Create alias

Failover Schedule

Model failover timing: 1 min → 5 min → 1 hour

Workspace & Memory

OpenClaw stores agent personality, user preferences, and long-term knowledge in Markdown workspace files and a hybrid vector/BM25 memory system.

Workspace Files

All workspace files are located at ~/.openclaw/workspace/:

File	Purpose
AGENTS.md	Operating instructions for the agent
SOUL.md	Persona, tone, and boundaries
USER.md	User information and preferences
IDENTITY.md	Agent name, emoji, theme
MEMORY.md	Curated long-term memory (DM only)
TOOLS.md	Local tool notes
HEARTBEAT.md	Heartbeat checklist
BOOT.md	Startup checklist
`memory/YYYY-MM-DD.md`	Daily append-only logs

Tip

AGENTS.md and SOUL.md are the most important files — they define how your agent behaves and responds.

Memory System

Daily Logs

memory/YYYY-MM-DD.md — Append-only logs. Reads today + yesterday at session start.

Long-Term Memory

MEMORY.md — Curated facts, loaded only in main DM session.

Vector Search

memory_search tool provides semantic search (~400 tokens per result).

Providers — Auto-selects from: local GGUF → OpenAI → Gemini → Voyage

Search Configuration

Feature	Details
Hybrid Search	Default 0.7 vector / 0.3 BM25 weighting
QMD Backend	Optional experimental mode with BM25 + vectors + reranking

openclaw memory index --all       # Reindex everything
openclaw memory search "query"    # Semantic search

Sessions & Commands

Control how conversations are scoped, reset, and managed through built-in slash commands and text-to-speech output.

Sessions

DM Scope Options

Scope	Description
`main`	Single shared session (default)
`per-peer`	One session per contact
`per-channel-peer`	One per contact per channel
`per-account-channel-peer`	Fully isolated sessions

Reset Modes

Mode	Description
`daily`	Resets at 4am local time (default)
`idle`	Resets after idle timeout

Configuration

{
  "session": {
    "dmScope": "main",
    "reset": {
      "idleMinutes": 30            // Sliding idle window
    },
    "resetByType": { ... },        // Override for dm, group, thread
    "resetByChannel": { ... },     // Per-channel override (highest priority)
    "identityLinks": { ... },      // Map provider:id to canonical identity
    "sendPolicy": { ... },         // Block delivery for specific types
    "store": "~/.openclaw/agents/{agentId}/sessions/sessions.json"
  }
}

Security

Use per-channel-peer for multi-user inboxes to prevent context leakage between users.

Slash Commands

Session Management

Command	Description
`/status`	Session health, context usage, credential status
`/context list`	What's in context window (largest contributors)
`/context detail`	Full system prompt and injected workspace files
`/model <model>`	Switch model for session
`/model list`	List available models
`/compact [instructions]`	Summarize older context, free window space
`/new [model]`	Start fresh session (optional model switch)
`/reset`	Alias for `/new`
`/stop`	Abort current run, clear queued followups

Delivery & Output

Command	Description
`/send on\|off\|inherit`	Override delivery for this session
`/tts on\|off`	Toggle text-to-speech
`/think`	Toggle reasoning mode
`/verbose`	Toggle verbose mode

Configuration

Command	Description
`/config`	Persisted config changes
`/debug`	Runtime-only overrides (requires `commands.debug: true`)

Text-to-Speech

Provider	Characteristics
ElevenLabs	Premium ultra-realistic, higher latency
OpenAI	Standard fast, high-quality voices
Edge TTS	Free, no API key, multi-language support

Enable auto-TTS:

{
  "messages": {
    "tts": {
      "auto": "always"
    }
  }
}

Agent Architecture

Scale beyond a single agent — route messages to specialized agents, extend capabilities with skills, delegate tasks to sub-agents, and keep everything alive with heartbeats.

Multi-Agent Routing

Routing Precedence

Priority	Binding	Description
1 (highest)	`peer`	Exact DM/group id
2	`guildId`	Discord server
3	`teamId`	Slack workspace
4	`accountId`	Account-level
5	`channel`	Channel type
6 (lowest)	`default`	Final fallback agent

Management

openclaw agents add <name>                # Add new agent
openclaw agents list --bindings           # List with routing bindings
openclaw agents delete <name>             # Remove agent

Each agent gets:

Isolated workspace
Per-agent auth profiles
Session store at ~/.openclaw/agents/<id>/sessions/

Skills System

Skill Precedence

Agent Skills (Highest)

<workspace>/skills/ — Per-agent, takes priority over all others.

Managed/Local Skills

~/.openclaw/skills/ — Shared across workspaces.

Bundled Skills (Lowest)

Shipped with OpenClaw — default capabilities.

ClawHub Registry

clawhub install <slug>                    # Install from ClawHub
clawhub update --all                      # Update all installed
clawhub sync --all                        # Scan and publish updates

SKILL.md Format

---
name: my-skill
description: "What this skill does"
metadata: { "openclaw": { "requires": { ... } } }
---

Sub-Agents

Sub-agents enable parallel work for research/long tasks without blocking the main conversation.

Own session key with optional sandbox isolation
Auto-announce results to requester chat channel
Auto-archive after 60 minutes (configurable)

Commands

Command	Description
`/subagents list`	List active sub-agents
`/subagents stop <id\|#\|all>`	Stop sub-agent runs
`/subagents log <id\|#>`	View sub-agent logs
`/subagents info <id\|#>`	Show run metadata
`/subagents send <id\|#> <msg>`	Send message to sub-agent

Spawn Tool — The sessions_spawn tool accepts:

{
  "task": "Research topic X",
  "label": "research-x",
  "model": "claude-sonnet",
  "thinking": true,
  "runTimeoutSeconds": 300,
  "cleanup": true
}
// Returns: { status, runId, childSessionKey }

Heartbeat System

{
  "heartbeat": {
    "every": "30m",                 // Interval (1h for Anthropic OAuth)
    "target": "last",               // last | none | <channel id>
    "to": null,                     // Optional recipient override
    "model": null,                  // Model override for heartbeat runs
    "prompt": null,                 // Custom prompt body
    "activeHours": {
      "start": "09:00",
      "end": "22:00",
      "timezone": "America/New_York"
    }
  }
}

Contract

Agent replies HEARTBEAT_OK if nothing needs attention. OK-only replies are automatically stripped and dropped.

Automation & Tools

Automate browser tasks, schedule cron jobs, and wire up lifecycle hooks to customize how your agent responds to events.

Browser Automation

openclaw browser start | stop             # Headless instance
openclaw browser tabs                     # List open pages
openclaw browser open <url>               # Open URL in new tab
openclaw browser navigate <url>           # Navigate current tab
openclaw browser screenshot               # Capture active view
openclaw browser click | type | press     # DOM interactions
openclaw browser evaluate <js>            # Run JavaScript
openclaw browser pdf                      # Export page as PDF

Cron Jobs

openclaw cron list                        # View scheduled jobs
openclaw cron add                         # Create new job
openclaw cron edit <id>                   # Edit existing job
openclaw cron enable | disable <id>       # Toggle job
openclaw cron run <id>                    # Manual trigger
openclaw cron runs                        # View run history

Hooks & Automation

Bundled Hooks

Hook	Description
`session-memory`	Save session context to memory on `/new`
`command-logger`	Log all commands to audit file
`boot-md`	Run `BOOT.md` on gateway start
`soul-evil`	Swap `SOUL.md` during purge window

Hook Commands

openclaw hooks list                       # List discovered hooks
openclaw hooks enable <name>              # Enable a hook
openclaw hooks disable <name>             # Disable a hook
openclaw hooks info <name>                # Show hook details
openclaw hooks check                      # Check eligibility

Event Types

Event	Trigger
`command:new`	When `/new` is issued
`command:reset`	When `/reset` is issued
`command:stop`	When `/stop` is issued
`gateway:startup`	After channels start
`agent:bootstrap`	Before workspace files injected

Operations & Security

Advanced commands for plugin management, logging, sandbox isolation, and system diagnostics.

Power Commands

Plugin & System Management

openclaw plugins list | enable | disable          # Manage plugins
openclaw approvals get | set | allowlist           # Execution approval policy
openclaw sandbox list | recreate | explain         # Inspect/rebuild sandbox
openclaw system event --text "X"                   # Queue system event
openclaw system heartbeat enable | disable | last  # Control heartbeat
openclaw update status | wizard                    # Release channel updates

Node & Device Management

openclaw nodes list | pending | approve            # Remote node management
openclaw devices list | approve | rotate | revoke  # Device token lifecycle
openclaw directory peers | groups list             # Resolve IDs

Monitoring & Dashboard

openclaw health --json --verbose        # Direct gateway health probe
openclaw dashboard                      # Open Control UI
openclaw tui --url <url> --token <tok>  # Remote terminal UI

Logging & Diagnostics

openclaw logs --follow                    # Tail logs (colorized in TTY)
openclaw logs --json                      # Line-delimited JSON
openclaw logs --limit 200                 # Limit output
openclaw channels logs --channel whatsapp # Channel-specific logs

OpenTelemetry Export

{
  "diagnostics": {
    "otel": {
      "enabled": true
    }
  }
}

Sandboxing

Mode Options

Mode	Description
`"off"`	No sandboxing, tools run on host
`"non-main"`	Sandbox only non-main sessions (default)
`"all"`	Every session runs in sandbox

Scope Options

Scope	Description
`"session"`	One container per session (default)
`"agent"`	One container per agent
`"shared"`	One container for all sandboxed sessions

Workspace Access

Access	Description
`"none"`	Tools see sandbox workspace only (default)
`"ro"`	Read-only mount at `/agent`
`"rw"`	Read/write mount at `/workspace`

Setup:

scripts/sandbox-setup.sh  # Creates openclaw-sandbox:bookworm-slim image

Reference

Quick-access paths and common troubleshooting steps.

Key Paths

Path	Description
`~/.openclaw/openclaw.json`	Main configuration file
`~/.openclaw/workspace/`	Default agent workspace
`~/.openclaw/agents/<id>/`	Per-agent state directory
`~/.openclaw/agents/<id>/sessions/`	Session store & transcripts
`~/.openclaw/credentials/`	OAuth/API keys
`~/.openclaw/memory/<agentId>.sqlite`	Vector index store
`/tmp/openclaw/openclaw-YYYY-MM-DD.log`	Gateway log file

Tip

Use --dev or --profile <name> to isolate all state under a separate directory for testing.

Troubleshooting

Universal Fix

When in doubt, run openclaw doctor --deep --yes — it performs health checks, quick fixes, and system service scans.

All Posts

Related Use Cases

Productivity

Autonomous Project Management with Subagents

Decentralized project coordination where subagents work autonomously on tasks, coordinating through shared STATE.yaml files rather than a central orchestrator.

AdvancedPro Plan

Social Media

Daily Reddit Digest

Run a daily digest of the top performing posts from your favourite subreddits, with preference learning over time.

BeginnerStarter Plan

Social Media

Daily YouTube Digest

Start your day with a personalized summary of new videos from your favorite YouTube channels — no more missing content from creators you actually want to follow.

BeginnerStarter Plan

Join the community

Subscribe to our newsletter for the latest news and updates

OpenClaw Cheatsheet 2026

No DM Reply

Silent in Group

Auth Expired

Gateway Down

Memory Not Indexing

Context Full

Channel Disconnected

Session Issues

Author

Categories

Related Use Cases

Autonomous Project Management with Subagents

Daily Reddit Digest

Daily YouTube Digest

Newsletter

OpenClaw Cheatsheet 2026

No DM Reply

Silent in Group

Auth Expired

Gateway Down

Memory Not Indexing

Context Full

Channel Disconnected

Session Issues

Author

Categories

Related Use Cases

Autonomous Project Management with Subagents

Daily Reddit Digest

Daily YouTube Digest

Newsletter