CommunityCoding & Developmentgithub.com

node-connect

Diagnose OpenClaw Control UI browser and native Android, iOS, or macOS node connection failures across route, auth, pairing, QR/setup-code, and reconnect states.

What is node-connect?

node-connect is a Claude Code agent skill that diagnose OpenClaw Control UI browser and native Android, iOS, or macOS node connection failures across route, auth, pairing, QR/setup-code, and reconnect states.

Works with~Claude Code~Codex CLI~Cursor
npx skills add https://github.com/clawdbot/clawdbot/tree/main/skills/node-connect

Installed? Explore more Coding & Development skills: steipete/bluebubbles, steipete/eightctl, steipete/blucli · View all 6 →

Ask in your favorite AI

Open a new chat with this agent skill pre-loaded.

Documentation

Node Connect

Goal: fix one exact client against one exact Gateway, then prove that client's fresh connection.

1. Lock the target

Record the target environment/profile, OpenClaw binary, config/state root, Gateway URL/port, and service before changing anything.

  • Use the named deployment wrapper/profile for every config, log, service, device, and node command.
  • Never drift to a bare openclaw, proof environment, or similarly named deployment.
  • If the global executable is stale, invoke the target through its owner.
  • Verify status, config, logs, and process identity describe the same Gateway.

Identify the target Gateway before changing pairing or auth.

2. Classify the client

Classify from the request and Gateway log before choosing commands:

  • Control UI browser: the user says browser, dashboard, Control UI, or webchat; logs show client=openclaw-control-ui or mode=webchat.
  • Native mobile/node: the official app shows Connect, Scan QR, or setup code; logs/request metadata show a native client or role=node.

A phone can be either client. Do not use openclaw qr or openclaw nodes status for a phone browser; those belong to native mobile/node pairing.

3. Observe the failing attempt

Run these through the locked target:

openclaw gateway status --deep
openclaw logs --follow --json
openclaw devices list
openclaw config get gateway.mode
openclaw config get gateway.bind
openclaw config get gateway.remote.url
openclaw config get gateway.auth.mode
openclaw config get gateway.auth.allowTailscale
openclaw config get gateway.tailscale.mode

Have the client retry once while logs are live. Correlate its client ID, mode, platform, address, auth result, device ID, user, and close code. Ignore other paired devices.

Interpret the first failed transition:

  • No matching Gateway attempt: route, DNS, TLS, origin, or proxy problem.
  • token_missing, password_missing, mismatch, or Tailscale identity failure: auth failed before pairing, so an empty pending list is expected.
  • pairing required: route and auth succeeded; approve the exact pending request.
  • authenticated user connected / webchat connected: match the exact client; if followed by 1006, preserve auth/pairing and inspect lifecycle, transport, proxy, or reconnect.
  • 1006 without either application-level connected log does not prove auth/pairing; inspect the earlier handshake transition.

4. Prove the route

Choose one topology: same machine, LAN, tailnet, or public reverse proxy. Do not mix them.

  • Browser Control UI needs HTTPS or localhost for browser device identity. A remote plain-HTTP Tailnet/LAN URL is not a valid substitute.
  • gateway.tailscale.mode=off means OpenClaw is not managing Serve/Funnel. It does not prove that Tailscale or an externally managed Serve route is absent.
  • When Tailscale is involved, inspect live state rather than inferring it from OpenClaw config:
tailscale status --json
tailscale serve status --json

Match the client's URL to the listener/proxy route reaching the locked Gateway.

5A. Control UI browser lane

Restore browser auth before looking for a pairing request:

  • For token/password auth, enter the credential in Control UI settings. Never put permanent secrets in chat, logs, or URLs.
  • Prefer openclaw dashboard on the Gateway host for a one-time signed handoff. Use --no-open only when the operator can retrieve that host's clipboard, and keep the host browser/clipboard outside agent tooling. Never capture dashboard --json: it can expose the handoff and shared credentials. Never relay, rewrite, or send a loopback handoff URL to a remote phone.
  • For Tailscale Serve, verify the live route and forwarded identity. Enable gateway.auth.allowTailscale only for that intended trust boundary. Verified Tailscale Control UI auth with browser device identity can skip pairing.

After auth succeeds:

  • Retry; never infer pairing from the pre-auth attempt. Token/password auth can reveal pairing required; verified Tailscale identity can skip it.
  • If logs say pairing required, re-list devices and approve the exact request ID.
  • A successful verified-Tailscale connection may correctly create no pending or paired-device row.
  • After a repeated 1006, preserve auth/pairing and inspect reconnect evidence.

5B. Native mobile/node lane

Inspect the native route through the locked target without exposing the setup credential:

openclaw qr --json | jq '{gatewayUrl, gatewayUrls, auth, access, accessDowngraded, urlSource}'

For a CLI controlling a remote Gateway, add --remote before --json; it selects gateway.remote.url and remote credentials. If the redaction filter is unavailable, do not run raw QR JSON in agent-visible output.

Verify gatewayUrl and urlSource. The setup code is password-equivalent: have the operator copy it from Control UI → Devices → Pair device, or run openclaw qr --setup-code-only in a terminal outside agent tooling and paste it directly into the official app. Never relay it through agent/chat/tool output. Generate a fresh code after a URL/auth fix or expiry.

If the app reports pairing required:

openclaw devices list
openclaw devices approve --latest   # preview only; exits without approval
openclaw devices approve <requestId>
openclaw nodes status

--latest only previews the current request; never treat it as approval. Re-list immediately before the exact-ID command because retries can supersede the request. Never approve by position, age, or similarity.

6. Correlate and finish

Before approval, match available request, device/public-key, client, mode/role, platform, address, user, and retry-time facts.

Declare success only after a new attempt made after the final change proves all applicable checks:

  • the exact client reaches the locked Gateway;
  • intended auth succeeds;
  • the browser completes initial requests, or the native node appears in openclaw nodes status;
  • approval used the exact request ID; and
  • no immediate auth, pairing, or reconnect failure follows.

A QR/setup code, launched browser, empty pending list, approval, paired-device count, or Tailscale ping proves only one transition.

Report the diagnosis, chosen route/auth lane, exact-client evidence, and any remaining failed transition.

Individual skills in this repo

This repo contains 20 individual skills — each has its own dedicated page.

1password

Set up and use 1Password CLI for sign-in, desktop integration, and reading or injecting secrets.

acp-router

Route plain-language requests for Claude Code, Cursor, Copilot, OpenClaw ACP, OpenCode, Gemini CLI, 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.

add-model-provider

Add and live-prove a model provider with non-interactive config one-liners, without exposing credentials.

agent-transcript

Requested GitHub PR/issue agent transcripts: redact, trim, preview, and insert safely.

apple-notes

Create, view, edit, delete, search, move, or export Apple Notes via the memo CLI on macOS.

apple-reminders

List, add, edit, complete, or delete Apple Reminders and reminder lists via remindctl.

auto-qa

Run OpenClaw-wide autonomous QA and live/stress campaigns across independent subsystem lanes, with verified fixes and a resumable evidence report.

autoreview

Structured code review through the shared OpenClaw agent-skills installation.

bear-notes

Create, search, and manage Bear notes via grizzly CLI.

blacksmith-testbox

Run Blacksmith Testbox for CI-parity checks, secrets, hosted services, migrations, or builds local cannot reproduce.

block-kit

Use proactively for structured or interactive Slack replies, and when asked to author or validate native Slack Block Kit JSON.

blogwatcher

Monitor blogs and RSS/Atom feeds for updates using the blogwatcher CLI.

blucli

BluOS CLI (blu) for discovery, playback, grouping, and volume.

bluebubbles

Send and manage iMessages via BlueBubbles, including attachments, tapbacks, edits, replies, and groups.

browser-automation

Use when controlling web pages with the OpenClaw browser tool, especially multi-step flows, login checks, tab management, or recovery from stale refs/timeouts.

camsnap

Capture frames or clips from RTSP/ONVIF cameras and local webcams, including USB pan/tilt/zoom control.

canvas

Present HTML on connected OpenClaw node canvases, navigate/eval/snapshot, and debug canvas host URLs.

channel-message-flows

Use when running QA Lab channel message flow evidence.

clawdtributor

Clawtributor PRs here, last week or another window: discover conversation refs, recheck GitHub, rank by impact.

clawhub

Search ClawHub for plugins by default, or skills when explicitly requested; install, verify, update, uninstall, publish, or sync skills.

Related Skills