MCP.Market

create-claw skill

by openclaw·openclaw/openclaw·393k stars·MIT

Turn this agent into a shareable Claw for a person to review.

A100/100content scan

Is the create-claw skill safe?

Clean: nothing in its files matched our rules. We read 3 files in the folder on 2026-10-25.

No findings.

Install the create-claw skill

A skill is a folder. Copy it into your agent's skills folder and the agent loads it when the task matches its description. 

git clone --depth 1 https://github.com/openclaw/openclaw.git /tmp/openclaw
mkdir -p ~/.claude/skills
cp -r /tmp/openclaw/skills/create-claw ~/.claude/skills/create-claw
available in every project

In the Claude apps, zip the folder and upload it from the Skills settings. The folder on GitHub

The instructions your agent would load

SKILL.md as published, without the frontmatter. Read it on GitHub

Create a Claw

The Control UI invokes this manual skill in the agent's existing main conversation. It prepares a private draft before sending the message and attaches that request's draftId as working context. Use the draftId attached to this turn and edit only .openclaw/claw-drafts/.json; never search for a latest draft or use a shared pointer. The attached context is request-specific, so overlapping Claw creations stay isolated. Keep internal paths and IDs out of your reply. Work in this same conversation; the Control UI opens package review when the draft is ready and owns package assembly and publication.

  1. Capture the reusable agent. Read the current draft JSON. Use this conversation and the agent's durable instructions to identify its coherent reusable role, input, job, output, and authority. Preserve useful behavior and settings; leave out personal context. If the responsibilities do not form one coherent Claw, ask which responsibility to share before writing. Complete when the four scope fields describe an agent a new person can use.
  2. Inventory only relevant material. Select portable workspace instructions and files that implement that role. Locate this SKILL.md and use its scripts/cli.mjs for inventory so the CLI matches the running OpenClaw installation; invoke it with an absolute path when outside the skill directory. Run node scripts/cli.mjs skills list --agent --json and node scripts/cli.mjs plugins list --enabled --json. Compare the results with tools actually available in this session and capabilities used in the role. Inspect a candidate with node scripts/cli.mjs plugins inspect --json for its source and exact version. Declare a skill or plugin only when you can ground it in a current ClawHub package reference and exact version; plugin availability alone is not a requirement. Put package references in packages, MCP requirements in mcpServers, and useful schedules in cronJobs. Their declarations are dependencies, not files copied into the archive. Complete when every included file and declaration has a reason tied to the reusable role.
  3. Author the proposal. Follow the draft format. Rewrite instructions for a new owner. Include input, job, output, and authority in the draft scope, and describe the portable behavior in CLAW.md. Place only newly authored portable text in files, mapping included workspace files by source in the manifest. Exclude raw conversations, memories, USER.md, credentials, secret values, private URLs, local paths, and personal facts. Never copy the source workspace tree. Complete when the JSON at this turn's draft path is valid and contains no private source material.
  4. Hand off for review. Briefly tell the person what the Claw does, which optional files/dependencies you included, and any capability you could not port. Do not print internal draft paths or IDs, generate a review link, upload, or publish. The Control UI opens a review of the exact built archive; the person can remove optional items and explicitly publish there. Complete when the review opens.

More skills from openclaw/openclaw

  • A1passwordSet up and use 1Password CLI for sign-in, desktop integration, and reading or injecting secrets.
  • Aacp-routerRoute plain-language requests for Claude Code, Cursor, Copilot, OpenClaw ACP, OpenCode, Gemini CLI, Grok Build, Qwen, Kiro, Kimi, iFlow, Factory Droid, Kilocode, or explicit ACP harness work into either OpenClaw ACP runtime sessions or direct acpx-driven sessions ("telephone game" flow). For coding-agent thread requests, read this skill first, then use only `sessions_spawn` for thread creation. Codex chat binding defaults to the native Codex app-server plugin unless ACP is explicit or background spawn needs ACP.
  • Aadd-model-providerAdd and live-prove a model provider with non-interactive config one-liners, without exposing credentials.
  • Aagent-transcriptRequested GitHub PR/issue agent transcripts: redact, trim, preview, and insert safely.
  • Aapple-notesCreate, view, edit, delete, search, move, or export Apple Notes via the memo CLI on macOS.
  • Aapple-remindersList, add, edit, complete, or delete Apple Reminders and reminder lists via remindctl.
  • Aauto-qaRun OpenClaw-wide autonomous QA and live/stress campaigns across independent subsystem lanes, with verified fixes and a resumable evidence report.
  • AautoreviewStructured code review through the shared OpenClaw agent-skills installation.
  • Abear-notesCreate, search, and manage Bear notes via grizzly CLI.
  • Ablock-kitUse proactively for structured or interactive Slack replies, and when asked to author or validate native Slack Block Kit JSON.
  • AblogwatcherMonitor blogs and RSS/Atom feeds for updates using the blogwatcher CLI.
  • AblucliBluOS CLI (blu) for discovery, playback, grouping, and volume.

All agent skills → · MCP servers