How do I install the Openclaw CLI skill?

Run npx skillmds@latest add modbender/openclaw-cli in your terminal (requires Node.js), paste this page's agent-chat prompt into Claude, Cursor, or any MCP-connected agent, or download the SKILL.md file and copy it into your agent's skills directory.

Is Openclaw CLI safe to use?

This skill has not completed SkillMD's automated safety review yet. Independent scanners report: SkillSpector: PASS, Skill Scanner: PASS. Capability flags: docs only. SkillMD never runs a skill's scripts for you; review the SKILL.md before installing.

Which AI agents work with Openclaw CLI?

This skill is tagged as working with Claude Code, Claude.ai, OpenAI Codex. SKILL.md is an open format, so most agents that read a skills directory can load it too.

Is Openclaw CLI free to use?

Yes. Installing skills from SkillMD is free, and the skill stays under its author's original license.

Who published Openclaw CLI?

modbender (@modbender) published this skill. Their other Agent Skills are listed on their SkillMD profile.

Openclaw CLI by modbender — Agent Skill

Name: Openclaw CLI
Rating: 4.2 (5 reviews)
Author: modbender

Openclaw CLI

Operate and troubleshoot the OpenClaw CLI across setup, gateway/node lifecycle, channel login, messaging, agent turns, models, plugins, and system health. Use when asked to choose, run, debug, or explain `openclaw` commands, global flags (`--dev`, `--profile`), or command families such as `gateway`, `agent`, `channels`, `message`, `doctor`, `status`, and `update`.

modbender Updated Sep 9, 2026 12 repo stars

File contents

OpenClaw CLI

Run OpenClaw commands safely and efficiently. Choose the right command family, execute with the correct profile context, and verify postconditions.

Execution Workflow

Clarify the target state. Ask what should change and what must remain untouched.
Select runtime scope first. Use default profile unless isolation is requested. Use:

openclaw --dev ... for isolated dev state under ~/.openclaw-dev.
openclaw --profile <name> ... for named isolated state under ~/.openclaw-<name>.

Choose the command family. Use references/command-map.md to route the request quickly.
Expand command details before running risky operations. Run openclaw <command> --help for starred families and confirm flags.
Prefer machine-readable output when automation is needed. Use --json where available, then parse/verify.
Verify outcomes explicitly. Check with openclaw status, openclaw health, openclaw nodes status --json, or command-specific follow-up.

Safety Rules

Require explicit user confirmation before reset, uninstall, destructive --force flows, or credential-clearing operations.
Prefer non-destructive diagnostics first: status, health, doctor, logs.
Keep profile usage consistent across a workflow. Do not mix default and --dev/--profile commands accidentally.
For gateway issues, diagnose before restart unless restart is explicitly requested.

Triage Sequence

For generic "OpenClaw not working" issues:

Run openclaw status.
Run openclaw health.
Run openclaw doctor.
Check openclaw gateway ..., openclaw node ..., or openclaw nodes ... based on where failure appears.
Escalate to targeted commands in references/command-map.md.

Resources

references/command-map.md: Command families, routing guidance, and practical recipes.

---name: openclaw-clidescription: Operate and troubleshoot the OpenClaw CLI across setup, gateway/node lifecycle, channel login, messaging, agent turns, models, plugins, and system health. Use when asked to choose, run, debug, or explain `openclaw` commands, global flags (`--dev`, `--profile`), or command families such as `gateway`, `agent`, `channels`, `message`, `doctor`, `status`, and `update`.--- # OpenClaw CLI Run OpenClaw commands safely and efficiently. Choose the right command family, execute with the correct profile context, and verify postconditions. ## Execution Workflow 1. Clarify the target state. Ask what should change and what must remain untouched. 2. Select runtime scope first. Use default profile unless isolation is requested. Use: - `openclaw --dev ...` for isolated dev state under `~/.openclaw-dev`. - `openclaw --profile <name> ...` for named isolated state under `~/.openclaw-<name>`. 3. Choose the command family. Use `references/command-map.md` to route the request quickly. 4. Expand command details before running risky operations. Run `openclaw <command> --help` for starred families and confirm flags. 5. Prefer machine-readable output when automation is needed. Use `--json` where available, then parse/verify. 6. Verify outcomes explicitly. Check with `openclaw status`, `openclaw health`, `openclaw nodes status --json`, or command-specific follow-up. ## Safety Rules - Require explicit user confirmation before `reset`, `uninstall`, destructive `--force` flows, or credential-clearing operations. - Prefer non-destructive diagnostics first: `status`, `health`, `doctor`, `logs`. - Keep profile usage consistent across a workflow. Do not mix default and `--dev`/`--profile` commands accidentally. - For gateway issues, diagnose before restart unless restart is explicitly requested. ## Triage Sequence For generic "OpenClaw not working" issues: 1. Run `openclaw status`. 2. Run `openclaw health`. 3. Run `openclaw doctor`. 4. Check `openclaw gateway ...`, `openclaw node ...`, or `openclaw nodes ...` based on where failure appears. 5. Escalate to targeted commands in `references/command-map.md`. ## Resources - `references/command-map.md`: Command families, routing guidance, and practical recipes.

Frequently asked questions