Symptoms: /codex-feishu start fails or daemon exits immediately.
Steps:
- Run
/codex-feishu doctorto identify the issue - Check that Node.js >= 20 is installed:
node --version - Check that Codex CLI is available:
codex --version - Confirm the installed Codex also supports
codex app-server - Verify config exists:
ls -la ~/.codex-feishu/config.envWindows:Get-Item ~/.codex-feishu/config.env - Check logs for startup errors:
/codex-feishu logs
Common causes:
- Missing or invalid config.env -- run
/codex-feishu setup - Node.js not found or wrong version -- install Node.js >= 20
- Codex is too old or from an incompatible install path -- update Codex or point
CODEX_FEISHU_CODEX_EXECUTABLEat a compatible binary - Port or resource conflict -- check if another instance is running with
/codex-feishu status
Symptoms: Bot is online but doesn't respond to messages.
Steps:
- Verify Feishu app credentials:
/codex-feishu doctor - Confirm the app has completed both publish cycles
- Confirm Long Connection is enabled
- Confirm
im.message.receive_v1andcard.action.triggerare saved - Check logs for incoming message events:
/codex-feishu logs 200
Symptoms: Codex starts working but the permission card does not arrive or approval does nothing.
Steps:
- Confirm
card.action.triggercallback is configured and published - Confirm the bridge is connected through Feishu long connection
- Check logs for permission forwarding and callback handling
- Trigger a deterministic permission test with
测试授权链路or/permtest
Symptoms: Feishu shows You've hit your usage limit, purchase more credits, or asks you to retry at a later time.
This is returned by Codex itself, not by Feishu. Open https://chatgpt.com/codex/settings/usage to check quota, upgrade or purchase credits if needed, then retry after the displayed reset time.
Symptoms: The thread list card appears, but button clicks do nothing.
Steps:
- Confirm
card.action.triggeris configured and the second publish is already approved - Confirm the bridge is running before saving Feishu callback settings
- Check
/codex-feishu logs 200for callback handling errors - As a fallback, use text commands such as
线程列表and切换线程 2
Symptoms: The daemon process consumes increasing memory over time.
Steps:
- Check current memory usage:
/codex-feishu status - Restart the daemon to reset memory:
/codex-feishu stop /codex-feishu start - If the issue persists, check how many concurrent sessions are active
- Review logs for error loops that may cause memory leaks
Symptoms: You switch to a desktop thread in Feishu and expect continuation, but the thread is still working.
Expected behavior:
- Feishu returns
当前线程忙碌中 - The current desktop thread output is mirrored to Feishu
- After the desktop thread finishes, the next Feishu message can continue the thread normally
If this does not happen, inspect /codex-feishu logs 200 and confirm you switched to the intended thread.
Symptoms: Status shows "running" but the process doesn't exist, or start refuses because it thinks a daemon is already running.
The daemon management scripts handle stale PID files automatically. If you still encounter issues:
- Run
/codex-feishu stop-- it will clean up the stale PID file - If stop also fails, manually remove the PID file:
Windows:
rm ~/.codex-feishu/runtime/bridge.pidRemove-Item ~/.codex-feishu/runtime/bridge.pid -Force
- Run
/codex-feishu startto launch a fresh instance