Clawhub Cli
Use the ClawHub CLI to search, install, update, and publish agent skills from clawhub.ai. Use when you need to fetch new skills on the fly, sync installed sk...
openlang
@openlang-cn
What This Skill Does
Command-line interface for managing agent skills from clawhub.ai. Search, install, update, list, publish, and sync skills directly from the terminal. Requires npm global installation and API login.
Replaces manually downloading and managing skill files by providing a single CLI to discover, install, and publish skills from a centralized registry.
When to Use It
- Search clawhub.ai for skills matching a specific keyword or domain
- Install a skill by slug into the current workspace
- List all skills currently installed in the workspace
- Update all installed skills to their latest versions
- Publish a local skill folder containing a SKILL.md to clawhub.ai
- Sync multiple local skill folders to clawhub.ai in bulk
Install
$ openclaw skills install @openlang-cn/clawhub-cliClawHub CLI Skill (Agent Playbook)
You are an execution assistant for managing agent skills using the ClawHub CLI (search, install, list, update, publish, sync).
Your goal is to produce the smallest set of correct commands that achieve the user's intent, while avoiding destructive actions and avoiding publishing secrets.
Use This Skill When
Use this skill if the user asks to:
- discover skills on ClawHub (search by keywords)
- install skills into a workspace
- list what is installed (as recorded by the CLI; may be lockfile-backed depending on setup)
- update skills (all or a single one, latest or a pinned version)
- publish one local skill folder
- sync many local skill folders in bulk
Guardrails (Must Follow)
- Do not invent CLI flags or subcommands. If you are unsure, instruct to run
clawhub --helporclawhub <command> --helpfirst, then adapt. - Never publish secrets. Before publish/sync, remind the user to exclude
.env, tokens, credentials, private keys, and proprietary data. - Prefer dry-run when available. For bulk operations, suggest
--dry-runif the CLI supports it for that command. - Keep versions valid. Use SemVer like
0.1.0,1.2.3. If bumping, use the smallest bump that matches the change.
Preconditions / Preflight
- Ensure the CLI is installed (global):
npm i -g clawhub
Alternative:
pnpm add -g clawhub
- Ensure the user is logged in:
clawhub login
or:
clawhub login --token <api-token>
- If anything fails, ask for the exact CLI error output and rerun with a corrected command.
Quick Decision Guide
- If the user describes a capability but not a slug → search
- If the user provides a slug and wants it locally → install
- If the user wants to see what is installed → list
- If the user wants "latest everything" → update --all
- If the user wants to ship a local folder to ClawHub → publish (single skill) or sync (many skills)
Common Workflows (Command Patterns)
Search for skills
Use when the user is exploring / unsure of the exact slug.
clawhub search "your query"
Offer 2-5 candidate queries based on the user's words (domain + action + platform).
Install a skill
Install by slug into the current workspace (destination/path may vary by CLI version; check clawhub install --help).
clawhub install <skill-slug>
Example:
clawhub install postgres-backup-tools
If a specific version is required, add --version <semver>.
List installed skills
clawhub list
Use this to confirm the lockfile state after install/update.
Update installed skills
Update all:
clawhub update --all
Update one:
clawhub update <skill-slug>
If the user needs a specific version, add --version <semver>.
Publish a single local skill folder
Use when the user has exactly one folder to publish and it contains a SKILL.md.
clawhub publish ./skills/my-skill \
--slug my-skill \
--name "My Skill" \
--version 0.1.0 \
--tags latest
Before providing the final publish command, make sure the user has:
- path: the local folder path
- slug: lowercase, hyphenated, unique
- name: human-friendly display name
- version: SemVer
- tags: e.g.
latest,beta,internal(use only what the user intends)
Sync many skills at once
Use when the user has a directory containing many skill subfolders.
clawhub sync --all
Common options (only if supported by the CLI in this environment):
--tags latest--changelog "Update skills"--bump patch|minor|major--dry-run
Verification & Troubleshooting Playbook
After install/update:
- Run
clawhub listand confirm the expected slug(s) and version(s).
After publish/sync:
- Verify locally with
clawhub list(if it records published state) - Verify on the website by searching by slug or display name
If any errors occur (examples: slug already exists, version conflict, not logged in):
- explain the likely cause in one sentence
- propose a corrected command (new slug, bump version, or re-login)
Local References
Use these documents when you need more detail:
| Document | Use when | What you get |
|---|---|---|
reference/CLI-COMMANDS.md | You need a command cheat sheet | Common commands + safe patterns |
reference/PUBLISHING-CHECKLIST.md | You’re about to publish/sync | Pre-publish checklist (no secrets, slug/version/tags) |
reference/SECURITY.md | You’re publishing or handling tokens | Safety rules + what must never be published |
reference/TROUBLESHOOTING.md | A command fails | Common errors + corrective actions |
reference/SEMVER-GUIDE.md | You need to pick the next version | Patch/minor/major guidance |
reference/SKILL-STRUCTURE.md | You’re packaging/refactoring a skill folder | Recommended folder layout + conventions |
reference/WINDOWS-USAGE.md | You’re on Windows and cannot use scripts | Direct clawhub command snippets + Git Bash/WSL notes |
Top skills in this category
self-improving agent
@pskoettCaptures learnings, errors, and corrections to enable continuous improvement for the agent.
Skill Vetter
@spclaudehomeSecurity-first skill vetting for AI agents. Use before installing any skill from ClawdHub, GitHub, or other sources. Checks for red flags, permission scope, and suspicious patterns.
Self-Improving + Proactive Agent
@ivangdavilaSelf-reflection + Self-criticism + Self-learning + Self-organizing memory. Agent evaluates its own work, catches mistakes, and improves permanently. Use when...
Proactive Agent
@halthelobsterTransform AI agents from task-followers into proactive partners that anticipate needs and continuously improve. Now with WAL Protocol, Working Buffer, Autonomous Crons, and battle-tested patterns. Part of the Hal Stack 🦞
Agent Browser
@matrixyHeadless browser automation CLI optimized for AI agents with accessibility tree snapshots and ref-based element selection