TypeError: Cannot read properties of undefined (reading 'trim') | Friends of the Crustacean 🦞🤝 | Page 1

wet jungle Apr 16, 2026, 4:15 AM

#

i can never finish the onboarding now because of this

half spearBOT Apr 16, 2026, 4:16 AM

#

that’s frustrating. First thing I need is your exact version, because this exact trim onboarding crash has had fixes recently.

Please paste:

openclaw --version
openclaw status --all

Why I’m asking:

there are recent fixes for onboarding/setup crashes with Cannot read properties of undefined (reading 'trim')
if you’re already on a newer build, the next likely cause is stale bootstrapping/session state from a failed earlier run

Relevant docs:

FAQ: https://docs.openclaw.ai/help/faq
Bootstrapping: https://docs.openclaw.ai/start/bootstrapping

#

Paste those two outputs and I’ll tell you the next exact step.

wet jungle Apr 16, 2026, 4:19 AM

#

PS C:\Users\unkil> openclaw --version
OpenClaw 2026.4.14 (323493f)
PS C:\Users\unkil> openclaw status --all

OpenClaw status --all

Overview
┌─────────────────┬────────────────────────────────────────────────────────────────────────────────────────────────────┐
│ Item │ Value │
├─────────────────┼────────────────────────────────────────────────────────────────────────────────────────────────────┤
│ Version │ 2026.4.14 │
│ OS │ windows 10.0.26200 (x64) │
│ Node │ 24.14.1 │
│ Config │ ~.openclaw\openclaw.json │
│ Dashboard │ http://127.0.0.1:18789/ │
│ Tailscale │ off │
│ Channel │ stable (default) │
│ Update │ pkg │
│ Gateway │ local · ws://127.0.0.1:18789 (local loopback) · unreachable │
│ Security │ Run: openclaw security audit --deep │
│ Gateway self │ unknown │
│ Gateway service │ Scheduled Task not installed │
│ Node service │ Scheduled Task not installed │
│ Agents │ 0 total · 0 bootstrapping · 0 active · 0 sessions │
│ Secrets │ none │
└─────────────────┴────────────────────────────────────────────────────────────────────────────────────────────────────┘

Channels
┌──────────┬─────────┬────────┬────────────────────────────────────────────────────────────────────────────────────────┐
│ Channel │ Enabled │ State │ Detail │
├──────────┼─────────┼────────┼────────────────────────────────────────────────────────────────────────────────────────┤
└──────────┴─────────┴────────┴────────────────────────────────────────────────────────────────────────────────────────┘

Agents
┌────────────┬────────────────┬──────────┬──────────┬──────────────────────────────────────────────────────────────────┐
│ Agent │ Bootstrap file │ Sessions │ Active │ Store │
├────────────┼────────────────┼──────────┼──────────┼──────────────────────────────────────────────────────────────────┤
└────────────┴────────────────┴──────────┴──────────┴──────────────────────────────────────────────────────────────────┘

Diagnosis (read-only)

Gateway connection details:
Gateway target: ws://127.0.0.1:18789
Source: local loopback
Config: C:\Users\unkil.openclaw\openclaw.json
Bind: loopback

✗ Config: C:\Users\unkil.openclaw\openclaw.json
✓ Secret diagnostics (0)
✓ Restart sentinel: none
✓ Port 18789
! Tailscale: off · unknown
✓ Plugin compatibility (none)
! Channel issues skipped (gateway unreachable)

Gateway health:
gateway unreachable

Pasteable debug report. Auth tokens redacted.
Troubleshooting: https://docs.openclaw.ai/troubleshooting

half spearBOT Apr 16, 2026, 4:19 AM

#

perfect, that helps.

You’re on a recent build, so this is less likely to be the older onboarding trim bug. The bigger clue is this part:

Gateway ... unreachable
Gateway service: Scheduled Task not installed
Agents: 0 total

So onboarding is not even reaching the point where the agent/bootstrap starts.

Try this in PowerShell, in order:

openclaw doctor
openclaw gateway status
openclaw gateway run

#

A few notes:

Gateway service not installed is okay if you run it manually
Gateway unreachable is not okay, that’s the blocker
keep openclaw gateway run open, then retry onboarding in another window

If openclaw gateway run crashes or prints the same trim error, paste that output here.
If doctor reports a config problem, paste that too, since status --all only shows ✗ Config without the reason.

Docs:

troubleshooting: https://docs.openclaw.ai/help/faq
gateway runbook: https://docs.openclaw.ai/gateway
windows notes: https://docs.openclaw.ai/platforms/windows

If you want, after you paste doctor + gateway run, I’ll narrow it down fast.

wet jungle Apr 16, 2026, 4:20 AM

#

well i just did a fresh reinstall of openclaw thats why there is nothing

half spearBOT Apr 16, 2026, 4:21 AM

#

yep, that explains the empty agents/sessions 👍

So the reinstall explains the blank state, but not the trim crash itself. For that, I need the actual failing output, not the post-failure status snapshot.

Please paste these:

openclaw doctor

Then run the same onboarding flow that crashes, and paste the full error output from the terminal.

If you’re using the dashboard/UI, do this first so we can catch the real log:

openclaw gateway run

#

Then retry onboarding and paste whatever shows up in that terminal when it blows up.

If you want, also tell me which command/path you’re using to onboard:

openclaw onboard
openclaw configure
dashboard/control UI

That’ll narrow this down a lot faster.

wet jungle Apr 16, 2026, 4:30 AM

#

◇ Select channel (QuickStart)
│ Discord (Bot API)
│
◇ Discord bot token ───────────────────────────────────────────────────────────────────────╮
│ │
│ 1) Discord Developer Portal -> Applications -> New Application │
│ 2) Bot -> Add Bot -> Reset Token -> copy token │
│ 3) OAuth2 -> URL Generator -> scope 'bot' -> invite to your server │
│ Tip: enable Message Content Intent if you need message text. (Bot -> Privileged Gateway │
│ Intents -> Message Content Intent) │
│ Docs: discord │
│ │
├───────────────────────────────────────────────────────────────────────────────────────────╯
│
◇ How do you want to provide this Discord bot token?
│ Enter Discord bot token
│
◇ Enter Discord bot token
│ i removed so you dont see
│
◇ Configure Discord channels access?
│ Yes
│
◇ Discord channels access
│ Open (allow all channels)
TypeError: Cannot read properties of undefined (reading 'trim')
PS C:\Users\unkil>

PS C:\Users\unkil> openclaw gateway run

🦞 OpenClaw 2026.4.14 (323493f) — Alexa, but with taste.

│
◇
00:24:45 [gateway] loading configuration…
00:24:45 [gateway] resolving authentication…
00:24:45 [gateway] starting...
00:25:10 [gateway] starting HTTP server...
00:25:10 [canvas] host mounted at http://127.0.0.1:18789/__openclaw__/canvas/ (root C:\Users\unkil.openclaw\canvas)
00:25:10 [health-monitor] started (interval: 300s, startup-grace: 60s, channel-connect-grace: 120s)
00:25:10 [gateway] agent model: openai/gpt-5.4
00:25:10 [gateway] ready (5 plugins: acpx, browser, device-pair, phone-control, talk-voice; 24.7s)
00:25:10 [gateway] log file: C:\Users\unkil\AppData\Local\Temp\openclaw\openclaw-2026-04-16.log
00:25:10 [gateway] starting channels and sidecars...
00:25:10 [hooks] loaded 4 internal hook handlers
00:25:10 [plugins] embedded acpx runtime backend registered (cwd: C:\Users\unkil.openclaw\workspace)
00:25:11 (node:5828) [DEP0190] DeprecationWarning: Passing args to a child process with shell option true can lead to security vulnerabilities, as the arguments are not escaped, only concatenated.
(Use node --trace-deprecation ... to show where the warning was created)
00:25:11 [browser/server] Browser control listening on http://127.0.0.1:18791/ (auth=token)
00:25:12 [heartbeat] started
00:25:13 [plugins] embedded acpx runtime backend ready
00:28:03 [ws] closed before connect conn=19c5640e-eb6d-41e7-9cba-77c0505751d2 peer=127.0.0.1:55837->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a code=1008 reason=connect failed

half spearBOT Apr 16, 2026, 4:33 AM

#

yep, this looks like a wizard bug in 2026.4.14, not something wrong with your token.

Also, this line is probably a side issue, not the cause of onboarding failing:

code=1008 reason=connect failed

That looks like the local canvas/control UI websocket, not the Discord setup crash.

Fastest ways forward

Option 1, best bet: update and retry
There are onboarding/setup fixes in the newer codepath.

openclaw update

#

If you’re okay trying beta:

openclaw update --channel beta

Then retry the Discord QuickStart.

Option 2, unblock yourself right now: skip the buggy wizard
Configure Discord manually instead of using the TUI prompt.

In your config, set:

{
  channels: {

#

    discord: {
      enabled: true,
      groupPolicy: "open",
      token: {
        source: "env",
        provider: "default",
        id: "DISCORD_BOT_TOKEN",
      },
    },
  },
}

Then locally set your env var, start the gateway, and DM the bot for pairing. Do not paste the token here.

Docs:

#

If you want, I can help you with the exact config file location/edit flow on Windows next.

wet jungle Apr 16, 2026, 4:50 AM

#

ok that worked, how do i set the embedding model to be this one
provider: lmstudio
model: text-embedding-embeddinggemma-300m-qat

half spearBOT Apr 16, 2026, 4:52 AM

#

nice, glad that unblocked it 🎉

Set it under agents.defaults.memorySearch:

{
  agents: {
    defaults: {
      memorySearch: {
        provider: "lmstudio",
        model: "text-embedding-embeddinggemma-300m-qat",
      },
    },
  },
}

#

If you want CLI commands instead:

openclaw config set agents.defaults.memorySearch.provider lmstudio
openclaw config set agents.defaults.memorySearch.model text-embedding-embeddinggemma-300m-qat

Then restart the gateway and reindex memory, because changing embedding models changes the vector space:

openclaw memory index --force

Docs:

#

If this fixed it, please right-click the solution message → Apps → ✅ Mark Solution so Answer Overflow can index it.

wet jungle Apr 16, 2026, 4:52 AM

#

{
"gateway": {
"auth": {
"mode": "token",
"token": "i removed it"
},
"mode": "local",
"port": 18789,
"bind": "loopback",
"tailscale": {
"mode": "off",
"resetOnExit": false
},
"controlUi": {
"allowInsecureAuth": true
}
},
"wizard": {
"lastRunAt": "2026-04-16T04:45:22.299Z",
"lastRunVersion": "2026.4.15-beta.1",
"lastRunCommand": "onboard",
"lastRunMode": "local"
},
"meta": {
"lastTouchedVersion": "2026.4.15-beta.1",
"lastTouchedAt": "2026-04-16T04:45:39.992Z"
},
"agents": {
"defaults": {
"workspace": "C:\Users\unkil\.openclaw\workspace",
"models": {
"lmstudio/google/gemma-4-e4b": {}
},
"model": {
"primary": "lmstudio/google/gemma-4-e4b"
}
}
},
"update": {
"channel": "beta"
},
"session": {
"dmScope": "per-channel-peer"
},
"tools": {
"profile": "coding",
"web": {
"search": {
"provider": "ollama",
"enabled": true
}
}
},
"models": {
"mode": "merge",
"providers": {
"lmstudio": {
"baseUrl": "http://localhost:1234/v1",
"api": "openai-completions",
"auth": "api-key",
"apiKey": "LM_API_TOKEN",
"models": [
{
"id": "google/gemma-4-e4b",
"name": "Gemma 4 E4B",
"reasoning": true,
"input": [
"text",
"image"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 131072,
"contextTokens": 64000,
"maxTokens": 8192
}
]
}
}
},
"auth": {
"profiles": {
"lmstudio:default": {
"provider": "lmstudio",
"mode": "api_key"
}
}
},
"channels": {
"discord": {
"enabled": true,
"token": "i removed it ",
"groupPolicy": "allowlist",
"guilds": {}
}
},
"plugins": {
"entries": {
"ollama": {
"enabled": true
},
"lmstudio": {
"enabled": true
}
}
},
"hooks": {
"internal": {
"enabled": true,
"entries": {
"boot-md": {
"enabled": true
},
"bootstrap-extra-files": {
"enabled": true
},
"command-logger": {
"enabled": true
},
"session-memory": {
"enabled": true
}
}
}
}
}

half spearBOT Apr 16, 2026, 4:53 AM

#

yep, in your current config it should go right here under agents.defaults:

{
  "agents": {
    "defaults": {
      "workspace": "C:\\Users\\unkil\\.openclaw\\workspace",
      "models": {
        "lmstudio/google/gemma-4-e4b": {}
      },
      "model": {
        "primary": "lmstudio/google/gemma-4-e4b"
      },
      "memorySearch": {
        "provider": "lmstudio",
        "model": "text-embedding-embeddinggemma-300m-qat"

#

      }
    }
  }
}

Then restart and rebuild the index:

openclaw memory index --force

One other thing I noticed: right now you have

#

"channels": {
  "discord": {
    "enabled": true,
    "groupPolicy": "allowlist",
    "guilds": {}
  }
}

That means server channels are still blocked. If you wanted open server access, change that to:

"groupPolicy": "open"

If you want, paste the updated redacted config and I’ll sanity check it.

wet jungle Apr 16, 2026, 5:40 AM

#

01:39:40 [tools] exec failed: elevated is not available right now (runtime=direct).
Failing gates: allowFrom (tools.elevated.allowFrom.<provider> / agents.list[].tools.elevated.allowFrom.<provider>)
Context: provider=discord session=agent:main:discord:direct:486637630394466342
Fix-it keys:

tools.elevated.enabled
tools.elevated.allowFrom.<provider>
agents.list[].tools.elevated.enabled
agents.list[].tools.elevated.allowFrom.<provider> raw_params={"command":"openclaw memory index","elevated":true,"pty":true}

how do i let my agent do it

half spearBOT Apr 16, 2026, 5:41 AM

#

the key detail is this part:

runtime=direct

So your agent is already on the host. It does not need elevated mode for openclaw memory index.

Best fix

Turn elevated off for that session, then ask again:

/elevated off

Or make that permanent in config:

#

{
  agents: {
    defaults: {
      elevatedDefault: "off"
    }
  }
}

If you really want to allow elevated anyway

Add a tight Discord allowlist for your user id only:

{
  tools: {

#

    elevated: {
      enabled: true,
      allowFrom: {
        discord: ["486637630394466342"]
      }
    }
  }
}

Then restart the gateway.

A couple important notes:

Do not use "*" here.
Per-agent agents.list[].tools.elevated... is optional and only makes it more restrictive.

#

Since you’re runtime=direct, I’d still recommend elevatedDefault: "off". It avoids this exact failure mode.

Docs:

If you want, I can show you the exact merged config block based on the JSON you pasted.

wet jungle Apr 16, 2026, 5:43 AM

#

PS C:\Users\unkil> openclaw sandbox explain --session agent:main:discord:slash:486637630394466342

🦞 OpenClaw 2026.4.15-beta.1 (943cb47)
I'll do the boring stuff while you dramatically stare at the logs like it's cinema.

Effective sandbox:
agentId: main
sessionKey: agent:main:discord:slash:486637630394466342
mainSessionKey: agent:main:main
runtime: direct
mode: off scope: agent
workspaceAccess: none workspaceRoot: C:\Users\unkil.openclaw\sandboxes

Sandbox tool policy:
allow (default): exec, process, read, write, edit, apply_patch, image, sessions_list, sessions_history, sessions_send, sessions_spawn, sessions_yield, subagents, session_status
deny (default): browser, canvas, nodes, cron, gateway, feishu, googlechat, nostr, msteams, mattermost, nextcloud-talk, matrix, bluebubbles, line, zalo, zalouser, synology-chat, tlon, discord, imessage, irc, qqbot, signal, slack, telegram, twitch, whatsapp

Elevated:
enabled: true
channel: (unknown)
allowedByConfig: false

Fix-it:

tools.sandbox.tools.allow
tools.sandbox.tools.alsoAllow
tools.sandbox.tools.deny
agents.list[].tools.sandbox.tools.allow
agents.list[].tools.sandbox.tools.alsoAllow
agents.list[].tools.sandbox.tools.deny
tools.elevated.enabled

Docs: docs.openclaw.ai/sandbox

PS C:\Users\unkil>

half spearBOT Apr 16, 2026, 5:44 AM

#

yep, this confirms it:

Elevated.enabled: true
allowedByConfig: false

So the missing piece is the Discord allowlist, not the global enable flag.

Add this:

{
  tools: {
    elevated: {
      enabled: true,
      allowFrom: {
        discord: ["486637630394466342"]

#

      }
    }
  }
}

Then restart the gateway.

One important detail

openclaw sandbox explain showing channel: (unknown) for that slash session is a bit misleading here. The actual exec failure already told us the provider was discord, so tools.elevated.allowFrom.discord is still the right key.

My recommendation

Because your runtime is direct, I’d also set:

#

{
  agents: {
    defaults: {
      elevatedDefault: "off"
    }
  }
}

That stops the agent from needlessly asking for elevated on normal host exec.

So:

Want elevated available? add tools.elevated.allowFrom.discord
Want it to stop tripping over elevated at all? set agents.defaults.elevatedDefault: "off"

#

If you want, I can paste the exact merged block for your current config.

wet jungle Apr 16, 2026, 5:46 AM

#

where do i do that in the dashboard

half spearBOT Apr 16, 2026, 5:47 AM

#

in the Dashboard, go to the Config tab, not the Gateway Access card.

Easiest way

Use Config → Raw JSON and add/edit these two spots:

{
  tools: {
    elevated: {
      allowFrom: {
        discord: ["486637630394466342"]
      }
    }
  },
  agents: {
    defaults: {

#

      elevatedDefault: "off"
    }
  }
}

You already have tools.elevated.enabled: true, so you only need to add the discord allowlist and flip elevatedDefault.

If you want the form view instead

In Config form mode, go to:

tools → elevated → allowFrom → discord
agents → defaults → elevatedDefault

Set:

#

discord = ["486637630394466342"]
elevatedDefault = "off"

Then Save / Apply and restart the gateway if it asks.

Docs:

If you want, paste the updated redacted config and I’ll check it before you hit apply.

wet jungle Apr 16, 2026, 5:49 AM

#

{
"gateway": {
"auth": {
"mode": "token",
"token": ""
},
"mode": "local",
"port": 18789,
"bind": "loopback",
"tailscale": {
"mode": "off",
"resetOnExit": false
},
"controlUi": {
"allowInsecureAuth": true
}
},
"wizard": {
"lastRunAt": "2026-04-16T04:45:22.299Z",
"lastRunVersion": "2026.4.15-beta.1",
"lastRunCommand": "onboard",
"lastRunMode": "local"
},
"meta": {
"lastTouchedVersion": "2026.4.15-beta.1",
"lastTouchedAt": "2026-04-16T05:46:33.663Z"
},
"agents": {
"defaults": {
"workspace": "C:\Users\unkil\.openclaw\workspace",
"models": {
"lmstudio/google/gemma-4-e4b": {}
},
"model": {
"primary": "lmstudio/google/gemma-4-e4b"
},
"memorySearch": {
"provider": "lmstudio",
"model": "text-embedding-embeddinggemma-300m-qat"
}
},
"list": [
{
"id": "main",
"tools": {
"profile": "full",
"alsoAllow": [
"browser",
"canvas",
"message",
"gateway",
"nodes",
"agents_list",
"tts"
]
}
}
]
},
"update": {
"channel": "beta"
},
"session": {
"dmScope": "per-channel-peer"
},
"tools": {
"profile": "coding",
"web": {
"search": {
"provider": "ollama",
"enabled": true
},
"fetch": {
"enabled": true
}
},
"elevated": {
"enabled": true
},
"loopDetection": {
"enabled": true
},
"experimental": {
"planTool": true
}
},
"models": {
"mode": "merge",
"providers": {
"lmstudio": {
"baseUrl": "http://localhost:1234/v1",
"api": "openai-completions",
"auth": "api-key",
"apiKey": "LM_API_TOKEN",
"models": [
{
"id": "google/gemma-4-e4b",
"name": "Gemma 4 E4B",
"reasoning": true,
"input": [
"text",
"image"
],
"cost": {
"input": 0,
"output": 0,
"cacheRead": 0,
"cacheWrite": 0
},
"contextWindow": 131072,
"contextTokens": 64000,
"maxTokens": 8192
}
]
}
}
},
"auth": {
"profiles": {
"lmstudio:default": {
"provider": "lmstudio",
"mode": "api_key"
}
}
},
"channels": {
"discord": {
"enabled": true,
"token": "",
"groupPolicy": "allowlist",
"guilds": {},
"ackReactionScope": "all"
},
"activity": "Reading TWSA",
"activityType": 3,
"agentComponents": {
"enabled": true
},
"intents": {
"guildMembers": true,
"presence": true
}
}
},
"plugins": {
"entries": {
"ollama": {
"enabled": true
},
"lmstudio": {
"enabled": true
},
"memory-core": {
"config": {
"dreaming": {
"enabled": true
}
}
},
"browser": {
"enabled": true
}
}
},
"hooks": {
"internal": {
"enabled": true,
"entries": {
"boot-md": {
"enabled": true
},
"bootstrap-extra-files": {
"enabled": true
},
"command-logger": {
"enabled": true
},
"session-memory": {
"enabled": true
}
}
}
},
"messages": {
"statusReactions": {
"enabled": true
}
}
}

half spearBOT Apr 16, 2026, 5:52 AM

#

ah, I think I see a real config bug here.

Your channels.discord block is closed too early:

"ackReactionScope": "all"
},
"activity": "Reading TWSA",

That stray } makes activity, activityType, agentComponents, and intents end up outside channels.discord, which is the wrong shape.

Fix that block to this

{
  channels: {

#

    discord: {
      enabled: true,
      token: "...",
      groupPolicy: "allowlist",
      guilds: {},
      ackReactionScope: "all",
      activity: "Reading TWSA",
      activityType: 3,
      agentComponents: {
        enabled: true,
      },
      intents: {
        guildMembers: true,
        presence: true,
      },

#

    },
  },
}

And for the elevated issue, add this too

{
  tools: {
    elevated: {
      enabled: true,
      allowFrom: {
        discord: ["486637630394466342"],
      },
    },

#

  },
  agents: {
    defaults: {
      elevatedDefault: "off",
    },
  },
}

Two quick notes

If token: "" values are just redacted for Discord here, ignore that.
If they are literally empty in your real config, that is also a problem.

So yes, I’m worried the broken channels.discord object is at least part of why onboarding is falling over.

Docs for the Discord shape:

#

If you want, paste the next revised config and I’ll sanity-check it before you restart.

plain canyon Apr 16, 2026, 5:58 AM

#

use openclaw setup instead

half spearBOT Apr 16, 2026, 5:59 AM

#

yep, good catch, openclaw setup does exist on current builds and is probably the better retry here. 👍

For AizenKai, I’d do this in order:

Fix the malformed channels.discord block from the config you pasted, the brace is off.
Then rerun:

openclaw setup

Docs: https://docs.openclaw.ai/cli/setup

If it still blows up after that, paste:

#

openclaw --version

and the next error output. That’ll tell us pretty quickly whether this is just config-shape breakage or a separate onboarding bug.

wet jungle Apr 16, 2026, 6:02 AM

#

06:02:10+00:00 error gateway connect failed: GatewayClientRequestError: pairing required
gateway connect failed: GatewayClientRequestError: pairing required
06:02:10+00:00 warn gateway/ws {"subsystem":"gateway/ws"} {"cause":"pairing-required","handshake":"failed","durationMs":27,"lastFrameType":"req","lastFrameMethod":"connect","lastFrameId":"2a45bbfe-df7c-439a-a545-211d84154c76","host":"127.0.0.1:18789","remoteAddr":"127.0.0.1","remotePort":54629,"localAddr":"127.0.0.1","localPort":18789,"endpoint":"127.0.0.1:54629->127.0.0.1:18789","deviceId":"90bff4ed72880276069a51cc70e0b293ae1555135b3726aba77c4eec03310fd8","requestId":"af4994d2-2763-460a-84d3-2dd9d6aa9c3d","reason":"scope-upgrade"} closed before connect conn=f66f466f-be34-4f17-bdff-cdb8d6f04cfc peer=127.0.0.1:54629->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a code=1008 reason=connect failed
06:02:11+00:00 warn gateway {"subsystem":"gateway"} security audit: device access upgrade requested reason=scope-upgrade device=90bff4ed72880276069a51cc70e0b293ae1555135b3726aba77c4eec03310fd8 ip=unknown-ip auth=token roleFrom=operator roleTo=operator scopesFrom=operator.read scopesTo=operator.admin,operator.approvals,operator.pairing,operator.read,operator.talk.secrets,operator.write client=cli conn=cd4a6694-e904-4e3b-a8aa-2e55156bbd2d
06:02:11+00:00 error gateway connect failed: GatewayClientRequestError: pairing required
06:02:11+00:00 error gateway connect failed: GatewayClientRequestError: pairing required
gateway connect failed: GatewayClientRequestError: pairing required

half spearBOT Apr 16, 2026, 6:03 AM

#

this one is clearer now, it’s a pairing / scope-upgrade problem, not just elevated mode.

Your log says:

reason:"scope-upgrade"
scopesFrom=operator.read
scopesTo=operator.admin,operator.approvals,operator.pairing,operator.read,operator.talk.secrets,operator.write

So the CLI is asking for more gateway permissions, and the gateway is waiting for you to approve that device.

Do this on the same machine

openclaw devices list

Then approve the current pending request ID:

#

openclaw devices approve <requestId>

Your log showed:

af4994d2-2763-460a-84d3-2dd9d6aa9c3d

but run openclaw devices list first, because if it retried, that ID may already be stale.

After approving

Retry:

#

openclaw setup

Docs:

If devices approve still fails, paste:

openclaw --version
openclaw devices list

and we can keep digging.

wet jungle Apr 16, 2026, 6:04 AM

#

PS C:\Users\unkil> openclaw devices approve af4994d2-2763-460a-84d3-2dd9d6aa9c3d

🦞 OpenClaw 2026.4.15-beta.1 (943cb47) — Your AI assistant, now without the $3,499 headset.

│
gateway connect failed: GatewayClientRequestError: pairing required
◇
Direct scope access failed; using local fallback.
unknown requestId
PS C:\Users\unkil>

half spearBOT Apr 16, 2026, 6:05 AM

#

yep, that means the request ID from the log is stale.

This part confirms it:

using local fallback ✅
unknown requestId = the pending pairing request was superseded and got a new ID

Do this now

Trigger the request again:

openclaw setup

In another terminal, immediately run:

openclaw devices list

#

Find the newest pending request, ideally the one for:

device 90bff4ed72880276069a51cc70e0b293ae1555135b3726aba77c4eec03310fd8
reason like scope-upgrade

Approve that new request ID:

openclaw devices approve <new-requestId>

Important

Don’t reuse af4994d2-2763-460a-84d3-2dd9d6aa9c3d, it’s already obsolete.

If openclaw devices list comes back empty, paste that output next, because then we’re in slightly weirder territory.

wet jungle Apr 16, 2026, 6:06 AM

#

PS C:\Users\unkil> openclaw setup

🦞 OpenClaw 2026.4.15-beta.1 (943cb47) — One CLI to rule them all, and one more restart because you changed the port.

Config OK: ~.openclaw\openclaw.json
Workspace OK: ~.openclaw\workspace
Sessions OK: ~.openclaw\agents\main\sessions
PS C:\Users\unkil> openclaw devices list

🦞 OpenClaw 2026.4.15-beta.1 (943cb47) — Your config is valid, your assumptions are not.

│
◇
[openclaw] Failed to start CLI: Error: gateway closed (1006 abnormal closure (no close frame)): no close reason
Gateway target: ws://127.0.0.1:18789
Source: local loopback
Config: C:\Users\unkil.openclaw\openclaw.json
Bind: loopback
at Object.onClose (file:///C:/Users/unkil/AppData/Roaming/npm/node_modules/openclaw/dist/call-DKkbbXOF.js:249:10)
at WebSocket.<anonymous> (file:///C:/Users/unkil/AppData/Roaming/npm/node_modules/openclaw/dist/client-CQUWcz_Z.js:390:23)
at WebSocket.emit (node:events:508:28)
at WebSocket.emitClose (C:\Users\unkil\AppData\Roaming\npm\node_modules\openclaw\node_modules\ws\lib\websocket.js:263:12)
at emitErrorAndClose (C:\Users\unkil\AppData\Roaming\npm\node_modules\openclaw\node_modules\ws\lib\websocket.js:1047:13)
at ClientRequest.<anonymous> (C:\Users\unkil\AppData\Roaming\npm\node_modules\openclaw\node_modules\ws\lib\websocket.js:886:5)
at ClientRequest.emit (node:events:508:28)
at emitErrorEvent (node:_http_client:108:11)
at Socket.socketErrorListener (node:_http_client:575:5)
at Socket.emit (node:events:508:28)
PS C:\Users\unkil>

half spearBOT Apr 16, 2026, 6:08 AM

#

that’s frustrating, but this actually gives us a cleaner signal.

openclaw devices list is no longer complaining about the old request ID, it’s failing because the gateway connection is closing abnormally:

gateway closed (1006 abnormal closure (no close frame))
target: ws://127.0.0.1:18789

So right now the next thing to inspect is the gateway/service itself, not pairing.

Please run these next

openclaw gateway status --deep
openclaw doctor

Why

1006 usually means one of these:

#

the gateway crashed on connect
the wrong/stale gateway service is listening
config/service mismatch
port conflict / multiple gateway processes

openclaw setup here only validated local files, it didn’t complete a live gateway handshake.

Paste those two outputs and we can narrow it down fast.

wet jungle Apr 18, 2026, 12:33 AM

#

19:50:11 [agent/embedded] Profile lmstudio:default timed out. Trying next account...
19:50:11 [agent/embedded] [llm-idle-timeout] lmstudio/qwen3.5-9b-claude-4.6-opus-reasoning-distilled produced no reply before the idle watchdog; retrying same model

is there a way to extend the time out wait or just make it do nothing instead of trying the next agent

half spearBOT Apr 18, 2026, 12:34 AM

#

yep.

That log is not switching agents, it’s doing two separate failovers:

LLM idle timeout
auth-profile rotation (Trying next account...)

To wait longer

Set:

{
  agents: {
    defaults: {
      llm: {
        idleTimeoutSeconds: 300

#

      }
    }
  }
}

300 = wait 5 minutes
0 = disable the idle watchdog entirely

Docs: https://docs.openclaw.ai/concepts/agent-loop

To stop “trying the next account”

Lock LM Studio to one profile:

#

{
  auth: {
    order: {
      lmstudio: ["lmstudio:default"]
    }
  }
}

If you also want it to fail instead of falling back to another model, make sure you don’t have extra entries in:

agents.defaults.model.fallbacks

Practical recommendation

#

For slow local reasoning models, I’d try this first:

{
  agents: {
    defaults: {
      llm: {
        idleTimeoutSeconds: 300
      }
    }
  },
  auth: {
    order: {
      lmstudio: ["lmstudio:default"]
    }
  }

#

If you want, paste your current auth + agents.defaults.model block and I’ll tell you the smallest edit.

wet jungle Apr 18, 2026, 12:45 AM

#

20:03:23 [discord] discord debounce flush failed: Error [ERR_MODULE_NOT_FOUND]: Cannot find module 'C:\Users\unkil\AppData\Roaming\npm\node_modules\openclaw\dist\extensions\discord\system-events-DUVqwOle.js' imported from C:\Users\unkil\AppData\Roaming\npm\node_modules\openclaw\dist\extensions\discord\message-handler-51fAmdwx.js

half spearBOT Apr 18, 2026, 12:46 AM

#

yeah, this one looks like a broken install, not a config problem.

message-handler-...js is trying to import a hashed file that simply is not present in your installed dist\extensions\discord\ bundle.

Fastest fix

Try:

openclaw update --channel beta
openclaw doctor
openclaw gateway restart

Docs: https://docs.openclaw.ai/install/updating

If `openclaw update` doesn’t fix it

Do a clean global reinstall:

#

npm i -g openclaw@beta
openclaw doctor
openclaw gateway restart

If it still happens after reinstall

Then I’d check for multiple OpenClaw installs on PATH:

where openclaw

because mixed old/new global installs can cause exactly this kind of missing hashed chunk import.

If you want, paste the output of:

#

where openclaw
openclaw gateway status --deep

and I’ll help you verify it’s using one clean install only.

wet jungle Apr 21, 2026, 5:49 AM

#

Memory Wiki is not enabled
Imported Insights and Memory Palace are provided by the bundled memory-wiki plugin.
Enable plugins.entries.memory-wiki.enabled = true, then reload this tab.

half spearBOT Apr 21, 2026, 5:49 AM

#

yep, that message is literal, the memory-wiki plugin isn’t enabled yet.

Enable it

In Dashboard → Config → Raw JSON, add:

{
  plugins: {
    entries: {
      "memory-wiki": {
        enabled: true
      }
    }
  }
}

#

Then Save / Apply, restart the gateway if it asks, and reload that tab.

CLI equivalent

openclaw config set plugins.entries.memory-wiki.enabled true --strict-json
openclaw gateway restart

Docs: https://docs.openclaw.ai/plugins/memory-wiki

If you want, paste your current plugins.entries block and I’ll show exactly where to drop it.

wet jungle Apr 23, 2026, 4:04 AM

#

in trying to connect to openclaw for a dashboard im making via websocket and i keep getting this

23:51:18 [ws] invalid handshake conn=841fe634-c42b-4564-94bd-5e2f3b7dceb9 peer=127.0.0.1:54586->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a
23:51:18 [ws] closed before connect conn=841fe634-c42b-4564-94bd-5e2f3b7dceb9 peer=127.0.0.1:54586->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a code=1008 reason=invalid request frame
23:51:18 [ws] invalid handshake conn=0acc0140-994e-4f19-bf91-e29cf3c06ecf peer=127.0.0.1:54587->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a
23:51:18 [ws] closed before connect conn=0acc0140-994e-4f19-bf91-e29cf3c06ecf peer=127.0.0.1:54587->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a code=1008 reason=invalid request frame
23:51:19 [ws] invalid handshake conn=ad2be1c3-9dc4-4317-a510-b81fb407bb99 peer=127.0.0.1:54588->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a
23:51:19 [ws] closed before connect conn=ad2be1c3-9dc4-4317-a510-b81fb407bb99 peer=127.0.0.1:54588->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a code=1008 reason=invalid request frame
23:51:20 [ws] invalid handshake conn=15e02f41-ec8d-47a2-b7ae-c682389dc9d7 peer=127.0.0.1:54589->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a
23:51:20 [ws] closed before connect conn=15e02f41-ec8d-47a2-b7ae-c682389dc9d7 peer=127.0.0.1:54589->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a code=1008 reason=invalid request frame
23:55:18 [ws] closed before connect conn=2865185f-6d5c-497c-80d8-fe492b8c4b1c peer=127.0.0.1:58109->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a code=1008 reason=invalid connect params: at /client/id: must be equal to constant; at /client/id: must match a schema in anyOf; at /clien
00:03:41 [ws] closed before connect conn=08526a67-e20b-4740-8110-25831148e7eb peer=127.0.0.1:60010->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a code=1008 reason=origin not allowed (open the Control UI from the gateway host or allow it in gateway.controlUi.allowedOrigins)
00:04:18 [ws] closed before connect conn=86a9eafe-6108-45be-954b-6c1dbf9648a1 peer=127.0.0.1:60015->127.0.0.1:18789 remote=127.0.0.1 fwd=n/a origin=n/a host=127.0.0.1:18789 ua=n/a code=1008 reason=device identity mismatch

half spearBOT Apr 23, 2026, 4:09 AM

#

you’re hitting 4 different checks, in order:

invalid request frame
Your first WS message must be a real gateway request frame, not raw JSON params.
It needs to look like:
```
{
  "type": "req",
  "id": "1",
  "method": "connect",
  "params": { ... }
}
```
invalid connect params
client.id is strict, you can’t invent one.

#

For a browser dashboard, use a built-in browser client shape, typically:

client.id: "openclaw-control-ui"
client.mode: "webchat"
client.platform: "web"

origin not allowed
Your page origin must exactly match gateway.controlUi.allowedOrigins, including port.

Example:
```
{
  gateway: {
    controlUi: {
      allowedOrigins: ["http://localhost:5173"]
    }
```

#

     }
   }

device identity mismatch
This one is the big one: device.id must equal the fingerprint of the same
Ed25519 public key you send in device.publicKey.
It cannot be a UUID or random string.

In current code, it’s basically:
```
device.id = sha256(raw_ed25519_public_key_bytes).hex()
```
Also, wait for connect.challenge first, then sign that nonce.

#TypeError: Cannot read properties of undefined (reading 'trim')

Fastest ways forward

Best fix

If you really want to allow elevated anyway

One important detail

My recommendation

Easiest way

If you want the form view instead

Fix that block to this

And for the elevated issue, add this too

Two quick notes

Do this on the same machine

After approving

Do this now

Important

Please run these next

Why

To wait longer

To stop “trying the next account”

Practical recommendation

Fastest fix

If openclaw update doesn’t fix it

If it still happens after reinstall

Enable it

CLI equivalent

If `openclaw update` doesn’t fix it