OpenClaw gateway.bind_no_auth: What This Critical Finding Means and How to Fix It

Published 2026-10-20 · Updated 2026-10-20 · By The Is My Agent Safe team

Short answer

Your Gateway listens beyond this machine, but the audit found no Gateway token, password or trusted-proxy auth. Fix it first: put gateway.bind back to loopback, or turn on token auth and firewall the port if you need remote access. Then run openclaw security audit --deep again.

  • OpenClaw rates gateway.bind_no_auth critical. It fires when gateway.bind isn't loopback and no shared secret or trusted-proxy auth is set.
  • Quickest fix: stop the Gateway, set gateway.bind to loopback, start it again and re-run the audit.
  • Need remote access? OpenClaw's docs prefer an SSH tunnel or Tailscale Serve, with the Gateway kept on loopback.
  • If the Gateway may have been reachable, follow OpenClaw's full rollback plan, which includes rotating its token or password and any affected integration keys.

Paste your audit: run openclaw security audit --deep and paste the output into our free check for a red/amber/green result with the most important finding first (Limits of the check).

What should I do first?

Put the Gateway back on your own machine unless you need remote access:

openclaw gateway stop
openclaw config set gateway.bind loopback
openclaw config validate
openclaw gateway start
openclaw security audit --deep

OpenClaw's config docs say config set needs the Gateway stopped, which is why the commands stop it first. With loopback, the default, "only local clients can connect". When the audit comes back without gateway.bind_no_auth, you're done with this finding.

If you need to reach the Gateway from other devices, read If you need remote access below before you change anything else.

What does gateway.bind_no_auth mean?

OpenClaw's security audit checks page lists it as critical: "Remote bind without shared secret". The settings it points to are gateway.bind and gateway.auth.*.

In OpenClaw 2026.10.5's source, the finding appears when both of these are true:

The detail line names your bind value, for example gateway.bind="lan" but no gateway.auth token/password is configured. OpenClaw's own suggested fix is: "Set gateway.auth (token recommended) or bind to loopback."

Why does OpenClaw rate it critical?

One port (18789 by default) carries its WebSocket, HTTP APIs and the Control UI, according to OpenClaw's network exposure docs. The same page says the lan, tailnet and custom bind values "make the Gateway reachable by more clients" and should only be used "with gateway auth (shared token/password, or a correctly configured trusted proxy) and a real firewall". It also says to "never expose the Gateway unauthenticated on 0.0.0.0".

OpenClaw's audit triage order puts "Public network exposure (LAN bind, Funnel, missing auth)" second, with the instruction "fix immediately".

If you need remote access

OpenClaw's exposure runbook says to "prefer the narrowest pattern that satisfies the workflow". In order:

  1. Loopback plus an SSH tunnel. Keep gateway.bind on loopback and tunnel 127.0.0.1:18789. The runbook recommends this for personal use, admin access and debugging.
  2. Loopback plus Tailscale Serve. The Gateway stays on loopback and Tailscale handles access. The network exposure docs say to prefer this over a LAN bind.
  3. A tailnet or LAN bind. Only on a private network with known devices, with Gateway auth, a firewall allowlist and no public port-forward.

If you keep a non-loopback bind, turn on token auth. The docs show this shape:

{ gateway: { auth: { mode: "token", token: "your-token" } } }

openclaw doctor --generate-gateway-token can generate a token for you. Then firewall the port to a tight list of source IP addresses. The runbook says to "avoid direct public port-forwarding to the Gateway".

What if the Gateway was already reachable?

The finding describes your config. It doesn't mean anyone connected. If the port could be reached from other machines, follow OpenClaw's rollback plan in full. It also turns off direct messages on your channels, locks down exec tools and removes "*" and unexpected senders from allowlists. The Gateway steps are:

  1. Stop any public forwarding, Tailscale Funnel or reverse proxy routes.
  2. Set gateway.bind back to loopback.
  3. Rotate Gateway tokens or passwords and any affected integration credentials.
  4. Review recent audit logs, run history, tool calls and config changes.
  5. Re-run openclaw security audit --deep.

To test from outside whether your agent answers, use our guide to checking if your agent is exposed.

Could it be a false alarm?

The check reads your config, so first confirm what the config says:

openclaw config get gateway

If gateway.bind really isn't loopback and you can't see a token, password or trusted-proxy setup, the finding is accurate. If you're sure auth is set and the finding still appears, our guide to audit false alarms shows how to check a finding and report it. Don't suppress a critical finding you haven't confirmed. For the other gateway findings, see every audit finding, explained.

What our free check covers

The free check at /check is paste-only. We check the text you paste: your openclaw security audit output. The check never connects to your Gateway or probes your network, so it can't tell you whether your Gateway port is reachable from outside. Nothing is installed into the agent you're checking. Green means no issues found in the checks we ran.

The free check puts the most important finding first, in plain English. The full report explains every finding and lists the fixes in order. The check doesn't detect malicious skills. We don't review the code inside skills or MCP servers ourselves, beyond what OpenClaw's own audit reports. The check is free, and a full report is $29. See Limits of the check.

Frequently asked questions

What does gateway.bind_no_auth mean?

OpenClaw's audit found that gateway.bind is set to something other than loopback (such as lan, tailnet or custom), and no Gateway token, password or trusted-proxy auth is configured. OpenClaw rates it critical.

How do I fix gateway.bind_no_auth?

If you don't need remote access, stop the Gateway, run openclaw config set gateway.bind loopback, start it again and re-run openclaw security audit --deep. If you do need remote access, keep the Gateway on loopback and use an SSH tunnel or Tailscale Serve, or turn on token auth and firewall the port.

Does this finding mean someone got in?

No. It describes your config, not an event. If the port could be reached from other machines, follow the full rollback plan in OpenClaw's exposure runbook. Its steps include putting the Gateway back on loopback, turning off direct messages on your channels, locking down exec tools, rotating the Gateway token or password and affected integration keys, removing "*" and unexpected senders from allowlists, and reviewing recent logs, tool calls and config changes.

Can your free check tell me if my Gateway is reachable?

No. We check the text you paste, which is where gateway.bind_no_auth comes from. The check never connects to your Gateway or probes your network, so it can't tell you whether the port is reachable. Our exposure guide shows how to test that yourself.

Sources

  1. OpenClaw docs: security audit checks (checkId catalog)
  2. OpenClaw docs: network exposure (bind, firewall, Gateway auth)
  3. OpenClaw docs: Gateway exposure runbook
  4. OpenClaw docs: running the security audit
  5. OpenClaw docs: openclaw config (CLI)
  6. OpenClaw docs: managing the Gateway service (CLI)
  7. OpenClaw source v2026.10.5: src/security/audit-gateway-config.ts (where the check is defined)