Bug type

Crash (process/app exits or hangs)

Beta release blocker

No

Summary

Version: 2026.5.12 · Windows 10 · Node v24.14.0

Problem: After updating to 5.12, my Feishu agent takes 4-7 minutes to respond per message. Before the update it was fast. Another agent (DeepSeek) on the same gateway responds normally.

What I found: Verbose logs show phase=channels.feishu.start-account fires on every incoming message — the Feishu channel re-initializes instead of staying connected.

I noticed Telegram got a fix for this in the release notes:

"Telegram: cache successful startup bot identity by account and token fingerprint for up to 24 hours, so restarts can skip redundant getMe probes"

Could the same caching be applied to the Feishu plugin? It seems like the Feishu probe request hangs/times out repeatedly, causing the delay.

What I tried (nothing helped):

• Switched API providers (multiple tested, same delay)
• Cleaned session files (276MB → cleared, no improvement)
• Fixed a wrong probe endpoint in source (/bot/v1/openclaw_bot/ping → /bot/v3/info/)
• Restarted gateway in normal mode

Expected: Feishu channel should cache identity like Telegram does, not re-probe on every message.

Steps to reproduce

Version: 2026.5.12 · Windows 10 · Node v24.14.0

Problem: After updating to 5.12, my Feishu agent takes 4-7 minutes to respond per message. Before the update it was fast. Another agent (DeepSeek) on the same gateway responds normally.

What I found: Verbose logs show phase=channels.feishu.start-account fires on every incoming message — the Feishu channel re-initializes instead of staying connected.

I noticed Telegram got a fix for this in the release notes:

"Telegram: cache successful startup bot identity by account and token fingerprint for up to 24 hours, so restarts can skip redundant getMe probes"

Could the same caching be applied to the Feishu plugin? It seems like the Feishu probe request hangs/times out repeatedly, causing the delay.

What I tried (nothing helped):

• Switched API providers (multiple tested, same delay)
• Cleaned session files (276MB → cleared, no improvement)
• Fixed a wrong probe endpoint in source (/bot/v1/openclaw_bot/ping → /bot/v3/info/)
• Restarted gateway in normal mode

Expected: Feishu channel should cache identity like Telegram does, not re-probe on every message.

Expected behavior

Feishu channel should cache identity like Telegram does, not re-probe on every message.

Actual behavior

start-account re-probes on every message

OpenClaw version

2026.5.12

Operating system

windows10

Install method

npm

Model

Provider / routing chain

Additional provider/model setup details

No response

Logs, screenshots, and evidence

Impact and severity

No response

Additional information

No response