#subagent spawning
12 messages · Page 1 of 1 (latest)
Let me verify against the local docs and source so I don't guess.
I found the likely auth path, I’m checking whether this is also a known recent regression or just stale pairing state.
This looks more like a real regression than a bad config. I’m pulling the exact fix note so I can point you to the cleanest path.
Yeah, probably.
Telegram working means the channel path is fine. nodes status and subagent spawn hit the local operator/device auth path, and pairing required there usually means stale or unapproved device pairing/scopes.
First, please paste:
openclaw --version
That matters a lot here, especially on Windows after updates.
Then try this in order:
- Check pending device auth, not
openclaw pairing
(pairingis for DM/chat senders, this case isdevices)
openclaw devices list
- If you see a pending request, approve the exact
requestId, then restart the gateway:
openclaw devices approve <requestId>
openclaw gateway restart
- If there is no pending request, the local operator device token is probably stale. Re-pair it:
openclaw devices list
openclaw devices rotate --device <deviceId> --role operator
If rotate does not help:
openclaw devices remove <deviceId>
openclaw devices list
openclaw devices approve <newRequestId>
openclaw gateway restart
- Retest:
openclaw nodes status
Why I’m pointing here: the docs/source explicitly call out recent fixes for both:
- local exec/node clients hitting loopback
pairing required
sessions_spawnfailing withclose(1008): pairing required
Docs:
- https://docs.openclaw.ai/help/faq
- https://docs.openclaw.ai/nodes/troubleshooting
- https://docs.openclaw.ai/cli/devices
If it still fails on exact 2026.4.14, paste these next:
openclaw status --all
openclaw devices list
openclaw logs --follow
and I’d treat it as a likely regression. If only gateway restarts have been tried so far, a full Windows reboot is also worth doing once.
nothing still
I’m checking whether there’s a post-2026.4.14 fix before I send you on a wild goose chase.
Okay, then I’m worried this is bug territory, not just bad pairing state.
I checked the docs/changelog, and I’m not seeing a newer post-2026.4.14 fix that cleanly explains both:
sessions_spawn/ subagent failure withpairing requiredopenclaw nodes statusfailing the same way
One last thing worth trying first, because this has cleared stuck pairing state for some people:
# full Windows reboot, not just gateway restart
If it still repros after a full reboot, please file a GitHub issue with these exact outputs:
openclaw --version
openclaw gateway status
openclaw status --all
openclaw devices list --json
openclaw nodes status
Then in a second shell:
openclaw logs --follow
and trigger the failure once (openclaw nodes status or a subagent spawn), then paste the first pairing required block from the logs.
Please also include: