OpenClaw gateway.control_ui.allowed_origins_required: What It Means and How to Fix It

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

Short answer

Your Control UI is served on a non-loopback bind, but the Gateway has no list of browser origins it should accept. Fix it first: if you don't use the Control UI from other devices, put gateway.bind back to loopback. Otherwise set gateway.publicOrigin to the HTTPS address you open it from, leave allowedOrigins unset, and run openclaw security audit --deep again.

  • OpenClaw rates allowed_origins_required critical. It fires when gateway.bind isn't loopback, the Control UI is on, no allowed origins are in effect, and Host-header fallback is off.
  • Fix: set gateway.publicOrigin (used as the allowlist when allowedOrigins is omitted), or set gateway.controlUi.allowedOrigins to the full origins you use.
  • allowed_origins_wildcard means allowedOrigins contains "*". host_header_origin_fallback means a dangerous fallback setting is on. Both are critical on a non-loopback bind.
  • An empty list, allowedOrigins: [], replaces the publicOrigin default, so the finding stays.

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?

If you only open the Control UI on this machine, put the Gateway back on loopback:

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

If you open it from other devices, tell the Gateway which address you use. Set gateway.publicOrigin and leave allowedOrigins unset:

openclaw gateway stop
openclaw config set gateway.publicOrigin https://control.example.com   # your own HTTPS origin
openclaw config get gateway.controlUi                                  # check allowedOrigins isn't set to [] or ["*"]
openclaw config validate
openclaw gateway start
openclaw security audit --deep

OpenClaw's Gateway config docs say publicOrigin "supplies the browser-origin allowlist when gateway.controlUi.allowedOrigins is omitted". It must be an HTTPS origin "without a path, query, or credentials". HTTP is accepted only for literal loopback hosts.

To list origins yourself instead, set an explicit list:

openclaw config set gateway.controlUi.allowedOrigins '["https://control.example.com"]' --strict-json

OpenClaw's config docs say config set needs the Gateway stopped, so stop it first, as above. When the audit comes back without this finding, you're done with it.

If you don't use the Control UI at all, you can turn it off with gateway.controlUi.enabled: false. OpenClaw's Control UI docs say that stops serving dashboard pages "while bots and existing Gateway connections keep running". In 2026.10.5's source, this check only runs when the Control UI is enabled.

What does gateway.control_ui.allowed_origins_required mean?

OpenClaw's security audit checks page lists it as critical: "Non-loopback Control UI without an effective browser-origin allowlist".

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

One detail catches people out. The config docs say "an explicit list, including [], replaces this default". So allowedOrigins: [] stops publicOrigin from being used, and the finding stays.

OpenClaw's own suggested fix is:

“Set gateway.publicOrigin with gateway.controlUi.allowedOrigins omitted, or configure a list of full trusted origins (for example https://control.example.com).”

The wildcard and Host-header variants

Two related checkIds cover the same setting from other angles. In 2026.10.5's source, both are critical when gateway.bind isn't loopback, and warn on loopback.

checkIdWhat it meansFix
gateway.control_ui.allowed_origins_wildcardallowedOrigins includes "*". OpenClaw: this "disables origin allowlisting and should be treated as an intentional allow-all policy".Replace "*" with the exact origins you use. The network exposure docs say ["*"] "should be avoided outside tightly controlled local testing".
gateway.control_ui.host_header_origin_fallbackdangerouslyAllowHostHeaderOriginFallback is true. OpenClaw: it "weakens DNS rebinding protections".openclaw config set gateway.controlUi.dangerouslyAllowHostHeaderOriginFallback false, then set explicit origins.

The second one turns off the first: with the Host-header fallback on, allowed_origins_required doesn't fire. That doesn't mean the setup is fine. You'll see host_header_origin_fallback instead.

Why does OpenClaw rate it critical?

A non-loopback bind means other machines can reach the Gateway's port. The browser-origin allowlist decides which websites a browser may use to talk to it. OpenClaw's network exposure docs say "non-loopback Control UI deployments need configured browser origins".

A non-loopback bind usually shows up with other gateway findings too, such as gateway.bind_no_auth when no Gateway auth is set. See our guide to gateway.bind_no_auth and every audit finding, explained.

Could it be a false alarm?

Confirm what the config says:

openclaw config get gateway

If bind isn't loopback and neither publicOrigin nor controlUi.allowedOrigins is set, the finding matches your config.

One case can look like a false alarm. OpenClaw's config docs say private same-origin loads, from loopback, RFC1918 or link-local addresses, .local, .ts.net or Tailscale CGNAT hosts, "are accepted without enabling Host-header fallback". So the Control UI can work from your LAN or Tailnet address with no allowlist. The 2026.10.5 check doesn't look at where you open it from. It only reads the bind, the Control UI setting and the origin config, so it still reports the finding. Setting publicOrigin or an explicit list clears it.

If you see origin not allowed in the browser after a change, OpenClaw's troubleshooting docs say that public or cross-origin deployments "need an entry in gateway.controlUi.allowedOrigins".

Our guide to audit false alarms covers how to check any finding. Don't suppress a critical finding you haven't confirmed.

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 loads your Control UI, so it can't tell which websites can reach it. 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.control_ui.allowed_origins_required mean?

OpenClaw's audit found the Control UI enabled on a non-loopback bind "without an effective browser-origin allowlist". OpenClaw rates it critical.

How do I fix gateway.control_ui.allowed_origins_required?

If you don't open the Control UI from other devices, stop the Gateway, set gateway.bind to loopback, start it and re-run the audit. If you do, set gateway.publicOrigin to the HTTPS origin you use (for example https://control.example.com) and leave gateway.controlUi.allowedOrigins unset, or set allowedOrigins to an explicit list of full origins.

What does allowed_origins_wildcard mean?

gateway.controlUi.allowedOrigins includes "*". OpenClaw says that "disables origin allowlisting". It's critical on a non-loopback bind and warn on loopback. Replace "*" with the exact origins you use.

What does host_header_origin_fallback mean?

gateway.controlUi.dangerouslyAllowHostHeaderOriginFallback is true. OpenClaw's checks table calls it a "DNS rebinding hardening downgrade". It's critical on a non-loopback bind and warn on loopback. Set it to false and configure explicit origins.

Can your free check see my Control UI?

No. We check the text you paste, which is where this finding comes from. The check never connects to your Gateway or loads your Control UI.

Sources

  1. OpenClaw docs: security audit checks (checkId catalog)
  2. OpenClaw docs: Gateway configuration (publicOrigin, controlUi.allowedOrigins)
  3. OpenClaw docs: network exposure (HSTS and origin notes)
  4. OpenClaw docs: Control UI
  5. OpenClaw docs: troubleshooting agent replies and the Control UI (origin not allowed)
  6. OpenClaw docs: openclaw config (CLI)
  7. OpenClaw source v2026.10.5: src/security/audit-gateway-config.ts (where the checks are defined)