| name | openclaw-master-ops |
| description | The definitive OpenClaw operations & expertise skill. Combines CLI mastery, gateway administration, security auditing, troubleshooting, channel management, agent/node orchestration, and production best practices. Includes automated release tracking with SQLite database for breaking change detection. Use for any OpenClaw operations task: setup, debugging, security hardening, performance tuning, multi-node deployments, channel configuration, model routing, cron automation, plugin management, incident response, and version change management.
|
OpenClaw Master Operations
The comprehensive OpenClaw operations skill — combining CLI mastery, gateway administration, security hardening, troubleshooting workflows, and production best practices.
When to Use This Skill
Use this skill whenever the user needs to:
- Set up OpenClaw — Installation, initial config, onboarding, multi-node deployments
- Operate the Gateway — Start/stop/restart, bind modes, auth, discovery, Tailscale exposure
- Debug issues — Connection failures, auth errors, channel problems, agent crashes, exec failures
- Harden security — Audit config, fix permissions, tighten group policies, SecretRef management
- Manage channels — Telegram, WhatsApp, Discord, Slack, Signal setup and troubleshooting
- Configure agents/models — Model routing, sandbox settings, tool profiles, ACP harness setup
- Automate operations — Cron jobs, scheduled tasks, watchdogs, health monitoring
- Manage plugins/skills — Install, update, publish to ClawHub, troubleshoot plugin conflicts
- Respond to incidents — Gateway crashes, session corruption, credential expiry, approval failures
Core Principles
1. Safety First
- Back up before destructive actions — Always backup
openclaw.json, credentials/, sessions/ before --fix operations
- Prefer
trash over rm — Use recoverable deletion when cleaning up state
- Verify before applying — Show users what will change before running
--fix or --repair commands
- SecretRef awareness — Know when config values are secret-managed vs plaintext
2. Diagnostic Discipline
Follow this triage sequence for any OpenClaw issue:
1. Check gateway status → openclaw gateway status
2. Run health check → openclaw doctor
3. Inspect logs → openclaw logs --recent 50
4. Verify auth → openclaw gateway query whoami
5. Check channels → openclaw channels list
6. Test exec → openclaw exec "echo test"
7. Security audit → openclaw security audit
3. Output Hygiene
- Prefer
--json for automation — Machine-readable output for scripts/CI
- Use
--no-color in logs — Clean output for cron/webhook contexts
- Redact sensitive data — Never log tokens, passwords, or credential paths
Command Reference
Gateway Operations
openclaw gateway start
openclaw gateway stop
openclaw gateway restart
openclaw gateway status
openclaw gateway run --bind loopback --auth token
openclaw gateway query whoami
openclaw gateway query sessions
openclaw gateway query channels
openclaw gateway query nodes
openclaw gateway discover
openclaw gateway discover --wide
Health & Diagnostics
openclaw doctor
openclaw doctor --repair
openclaw doctor --repair --non-interactive
openclaw doctor --deep
openclaw doctor --generate-gateway-token
Security Operations
openclaw security audit
openclaw security audit --deep
openclaw security audit --json | jq '.summary'
openclaw security audit --fix
openclaw security audit --fix --json
openclaw secrets list
openclaw secrets get <name>
openclaw secrets set <name> <value>
openclaw secrets delete <name>
Channel Management
openclaw channels list
openclaw channels list --json
openclaw channels status telegram
openclaw channels status whatsapp
openclaw qr show telegram
openclaw qr show whatsapp
openclaw channels repair telegram
openclaw channels repair whatsapp
Node & Agent Management
openclaw nodes list
openclaw nodes status
openclaw node pair
openclaw node unpair <node-id>
openclaw pairing list
openclaw agents list
openclaw agent status <agent-id>
openclaw agent restart <agent-id>
openclaw acp list
openclaw acp status
Session Management
openclaw sessions list
openclaw sessions list --active-minutes 30
openclaw sessions list --kinds subagent
openclaw sessions history <session-key>
openclaw sessions history <session-key> --limit 50 --include-tools
openclaw sessions send <session-key> "message here"
openclaw sessions kill <session-key>
Cron & Automation
openclaw cron list
openclaw cron list --json
openclaw cron add "0 6 * * *" "openclaw doctor --repair"
openclaw cron remove <job-id>
openclaw cron run <job-id>
openclaw heartbeat status
openclaw heartbeat run
Plugin & Skill Management
clawhub install <skill-slug>
clawhub update <skill-slug>
clawhub update --all
clawhub list
clawhub search "<query>"
openclaw plugins list
openclaw plugins install <plugin-name>
openclaw plugins update <plugin-name>
openclaw plugins remove <plugin-name>
Logs & Debugging
openclaw logs --recent 100
openclaw logs --follow
openclaw logs --level debug
openclaw logs gateway --recent 50
openclaw logs agent <agent-id> --recent 50
openclaw logs export ./logs-backup.tar.gz
Config Management
openclaw config list
openclaw config get <key>
openclaw config get gateway.auth.token
openclaw config set <key> <value>
openclaw config set gateway.mode local
openclaw config delete <key>
openclaw backup create
openclaw backup list
openclaw backup restore <backup-id>
System Operations
openclaw system info
openclaw system status
openclaw system resources
openclaw update
openclaw update --preview
openclaw reset --keep-channels
openclaw reset --all
Troubleshooting Playbooks
Gateway Won't Start
lsof -i :18789
netstat -tlnp | grep 18789
openclaw gateway stop --force
openclaw config get gateway.mode
openclaw config get gateway.bind
openclaw doctor --repair
openclaw logs gateway --recent 100
"Pairing Required" Errors
openclaw pairing list
openclaw node pair
openclaw config delete nodes.paired
openclaw node pair
openclaw qr show
Channel Auth Failures
openclaw channels status <channel>
ls -la ~/.openclaw/credentials/
openclaw channels auth <channel>
openclaw qr show <channel>
Exec Approval Timeouts
openclaw config get gateway.nodes.allowCommands
openclaw config get gateway.nodes.denyCommands
openclaw approvals list
openclaw config set gateway.nodes.allowCommands '["ls", "cat", "grep"]'
openclaw gateway restart
Session Corruption
openclaw sessions list
openclaw sessions kill <session-key>
openclaw doctor --repair
openclaw reset --sessions-only
Model/API Errors
openclaw config get agents.defaults.model
openclaw infer --model qwen/qwen3.5-plus "Hello"
openclaw models list
openclaw secrets set OPENAI_API_KEY <new-key>
Security Hardening Checklist
Run these periodically (weekly for production):
openclaw security audit --deep
openclaw security audit --fix
openclaw secrets list --show-metadata
openclaw config get channels.telegram.allowFrom
openclaw config get channels.whatsapp.allowFrom
openclaw config get channels.telegram.groupPolicy
openclaw config get channels.whatsapp.groupPolicy
openclaw plugins list
clawhub list
openclaw sessions list --active-minutes 1440
openclaw update --check
Production Best Practices
1. Use SecretRef for Sensitive Values
{
"gateway.auth.token": {
"$secretRef": "gateway-token"
},
"channels.telegram.token": {
"$secretRef": "telegram-bot-token"
}
}
2. Enable Sandbox Mode
openclaw config set agents.defaults.sandbox.mode on
openclaw config set agents.defaults.sandbox.docker true
3. Configure Session DM Scope
For shared inboxes:
openclaw config set session.dmScope per-channel-peer
4. Set Up Monitoring Cron
openclaw cron add "*/30 * * * *" "openclaw doctor --non-interactive"
openclaw cron add "0 3 * * 0" "openclaw security audit --deep --json > ~/security-audit.json"
5. Enable Logging Redaction
openclaw config set logging.redactSensitive tools
6. Configure Model Routing
openclaw config set agents.defaults.model.primary qwen/qwen3.5-plus
openclaw config set agents.defaults.model.fallback zai/glm-5
openclaw config set agents.defaults.model.reasoning qwen/qwen3.5-plus
Quick Reference: Command Families
| Family | Commands | Use When |
|---|
| Gateway | gateway start/stop/status/query | Managing the WebSocket server |
| Health | doctor, health | Diagnosing issues |
| Security | security audit, secrets | Hardening config |
| Channels | channels list/status/repair, qr | Channel setup/debug |
| Nodes | nodes list, node pair, pairing | Multi-node setups |
| Agents | agents list, agent status, acp | Agent management |
| Sessions | sessions list/history/send/kill | Session orchestration |
| Cron | cron list/add/run/remove | Automation |
| Plugins | plugins list/install/update | Plugin management |
| Skills | clawhub install/update/search | Skill management |
| Config | config list/get/set/delete | Config changes |
| Logs | logs --recent/--follow | Debugging |
| System | system info/status/update | System ops |
Escalation Paths
When standard troubleshooting fails:
- Check official docs —
openclaw docs or https://docs.openclaw.ai
- Search ClawHub —
clawhub search "<topic>" for relevant skills
- Check GitHub issues — https://github.com/openclaw/openclaw/issues
- Discord community — https://discord.gg/clawd
- Enable debug logging —
openclaw config set logging.level debug
Notes
- Profile isolation — Use
--profile <name> to isolate state: openclaw --profile dev gateway
- Dev mode — Use
--dev flag for isolated dev environment: openclaw --dev gateway
- Container targeting — Use
--container <name> to target specific containers
- JSON output — Add
--json for machine-readable output in all commands
- No color — Add
--no-color or set NO_COLOR=1 for clean logs
Release Tracking
This skill includes automated OpenClaw release tracking to stay current with frequent updates.
Database
Location: skills/openclaw-master-ops/release-history.db
Tracks:
- Release versions and changelogs
- Breaking changes (auto-detected)
- Skill knowledge verification status
Commands
cd /home/marius/.openclaw/workspace/skills/openclaw-master-ops
python3 scripts/release-tracker.py sync
python3 scripts/release-tracker.py history
python3 scripts/release-tracker.py breaking
python3 scripts/release-tracker.py skill-update
python3 scripts/release-tracker.py export
Automated Sync (Cron)
openclaw cron add "0 3 * * 0" "cd ~/.openclaw/workspace/skills/openclaw-master-ops && python3 scripts/release-tracker.py sync"
openclaw cron add "0 8 * * *" "cd ~/.openclaw/workspace/skills/openclaw-master-ops && python3 scripts/release-tracker.py skill-update"
When OpenClaw Updates
After running openclaw update:
- Run sync —
python3 scripts/release-tracker.py sync
- Check breaking —
python3 scripts/release-tracker.py breaking
- Review skill updates —
python3 scripts/release-tracker.py skill-update
- Update skill if needed — Edit SKILL.md based on breaking changes
- Republish —
clawhub publish ... --version 1.X.0 --changelog "Updated for OpenClaw <version>"
Breaking Change Detection
Auto-detects these patterns in changelogs:
removed → feature_removal
deprecated → deprecation
breaking → breaking_change
migration → migration_required
changed default → default_change
config schema → config_change
command removed → command_removal
See references/release-tracking.md for full workflow.
Related Skills
deep-research — For researching OpenClaw topics, comparing configurations
plugin-publisher — For creating and publishing OpenClaw plugins
npd-validator — For validating new OpenClaw feature ideas