#Channels not displaying in openclaw doctor

1 messages · Page 1 of 1 (latest)

cunning geyser
#

If I run openclaw doctor it does not show the channels details.

However openclaw channels status shows them as working?

`Gateway reachable.

  • Discord coder: enabled, configured, running, connected, bot:@Coder Bot, token:config
  • Discord critic: enabled, configured, running, connected, bot:@Critic Bot, token:config
  • Discord default: enabled, configured, running, connected, bot:@Openclaw, token:config
  • Discord ideator: enabled, configured, running, connected, bot:@Ideator Bot, token:config
  • Discord planner: enabled, configured, running, connected, bot:@Planner Bot, token:config, intents:content=limited
  • Discord reviewer: enabled, configured, running, connected, bot:@Reviewer Bot, token:config, intents:content=limited
  • Discord scout: enabled, configured, running, connected, bot:@Scout Bot, token:config, intents:content=limited
  • Discord surveyor: enabled, configured, running, connected, bot:@Surveyor Bot, token:config, intents:content=limited
  • Discord writer: enabled, configured, running, connected, bot:@Writer Bot, token:config, intents:content=limited
  • Slack default: enabled, configured, running, connected, bot:env, app:env
  • Telegram default: enabled, configured, running, connected, mode:polling, token:env`

Openclaw version 2026.4.26

urban flintBOT
#

To help others find answers, you can mark your question as solved via Right click solution message -> Apps -> âś… Mark Solution

torn meteorBOT
#

Yep — that’s expected in 2026.4.26.

openclaw doctor probes channel status internally, but it only prints a Channel warnings section when it finds actionable issues. It does not print the full healthy channel inventory.

For the detailed channel list, openclaw channels status is the right command, and your output means the gateway sees those accounts as configured/running/connected.

If you want the deeper live check, run:

openclaw channels status --probe

Also: intents:content=limited is informational, not a doctor warning. Doctor warns when Discord Message Content Intent is disabled.

Docs:

cunning geyser
#

So, are these all healthy - added the --probe?

Gateway reachable.

  • Discord coder: enabled, configured, running, connected, bot:@Coder Bot, token:config, intents:content=limited, works, audit ok
  • Discord critic: enabled, configured, running, connected, bot:@Critic Bot, token:config, intents:content=limited, works, audit ok
  • Discord default: enabled, configured, running, connected, bot:@Openclaw, token:config, intents:content=limited, works, audit ok
  • Discord ideator: enabled, configured, running, connected, bot:@Ideator Bot, token:config, intents:content=limited, works, audit ok
  • Discord planner: enabled, configured, running, connected, bot:@Planner Bot, token:config, intents:content=limited, works, audit ok
  • Discord reviewer: enabled, configured, running, connected, bot:@Reviewer Bot, token:config, intents:content=limited, works, audit ok
  • Discord scout: enabled, configured, running, connected, bot:@Scout Bot, token:config, intents:content=limited, works, audit ok
  • Discord surveyor: enabled, configured, running, connected, bot:@Surveyor Bot, token:config, intents:content=limited, works, audit ok
  • Discord writer: enabled, configured, running, connected, bot:@Writer Bot, token:config, intents:content=limited, works, audit ok
  • Slack default: enabled, configured, running, connected, bot:env, app:env, works
  • Telegram default: enabled, configured, running, connected, mode:polling, bot:@DavesHomelabBot, token:env, works
torn meteorBOT
#

Yep — from this output, they look healthy from the channel/probe perspective.

Quick read:

  • running, connected = gateway runtime is connected
  • works = live credential/API probe passed
  • audit ok = Discord configured channel permissions look OK
  • intents:content=limited = informational, not a failure here

So doctor staying quiet makes sense: there are no actionable channel warnings to print.

If a specific bot still doesn’t reply somewhere, I’d next check routing/pairing/mention rules for that guild/channel — not basic channel health.
Docs: https://docs.openclaw.ai/channels/troubleshooting

cunning geyser
#

What is Discord Message Content Intent used for and should it be enabled on the channel config or within the Discord bot settings?

torn meteorBOT
#

It’s used so the Discord bot can read the actual text content of normal guild/channel messages.

Enable it in the Discord Developer Portal, not in the OpenClaw channel config:

Discord Developer Portal → your Application → Bot → Privileged Gateway Intents → Message Content Intent

OpenClaw already requests the Message Content gateway intent when the Discord channel connects. The Developer Portal toggle is Discord’s permission for the bot/application to actually use it.

For your setup with many Discord bots: enable it on each bot application you want to respond to normal channel messages.

intents:content=limited in your probe output is okay — that’s not a failure. It generally means Discord reports it as limited/available for the bot, and your works, audit ok lines confirm the live probe is healthy.

After changing Discord intent settings, restart the gateway:

openclaw gateway restart
cunning geyser
#

Normal channel messages being messages directed to the channel that bot is authorized on?

torn meteorBOT
#

Basically yes — “normal channel messages” means regular human messages posted in a Discord server text channel/thread/forum that the bot can access.

There are two layers:

  1. Discord authorization/permissions
    Bot is in the server and can view/read/send in that channel.

  2. OpenClaw routing config
    That guild/channel/user is allowed by your channels.discord config, pairing, mention rules, etc.

Message Content Intent is what lets the bot read the actual text body of those messages. Without it, the bot may be connected and able to see that “a message happened,” but not reliably see the message text — so normal non-command/non-DM channel replies can fail or behave oddly.

So for OpenClaw Discord bots that should respond in server channels: yes, enable it in the Discord bot/application settings.

cunning geyser
#

ok.
Another question still related to Discord config. Normally my main agent responds in it's channel. Since the latest update I'm seeing it typing in the channel, but the response is received as a direct message from the bot.
Is this expected?

torn meteorBOT
#

Yes — that sounds expected with the newer default.

OpenClaw now defaults group/server channel replies to:

messages.groupChat.visibleReplies: "message_tool"

Meaning: the agent still sees the channel message and may show typing in the channel, but its normal final answer is private unless it explicitly uses the message tool to post visibly.

To restore the older behavior where normal replies go back into the Discord channel:

openclaw config set messages.groupChat.visibleReplies automatic
openclaw gateway restart
#

Or config shape:

{
  messages: {
    groupChat: {
      visibleReplies: "automatic",
    },
  },
}

So: not a health issue — it’s the newer “lurk/private by default unless explicitly posting” group-channel behavior.

Docs: https://docs.openclaw.ai/channels/groups

cunning geyser
#

Thank you. That helps. I was thinking I'd broken something 🙂

torn meteorBOT
#

You didn’t break it 🙂 — this one is just a behavior/default change, and the typing-in-channel + DM/private final reply is exactly the kind of thing that looks like a routing bug at first.

Glad that helped! 🎉
If this answered it, please right-click the solution message → Apps → ✅ Mark Solution so Answer Overflow indexes it for the next person.