1 messages · Page 1 of 1 (latest)
clientSecret also accepts SecretRef input, not just a plaintext string. 官方文档说qqbot 支持SecretRef,实测 exec provider 不行呀
我看看
滨海大厦的朋友?
不是哦嘿嘿
我记得有腾讯云的兄弟在
@inner cypress Hello, how are you?
我base北京
hello~
I am senior full stack & AI developer from Mexico.
I am interested in openclaw and contributed a lot.
If you don't mind, would like to join your team and work together.
I will be good help to manage this channel and integrating QQ channel with openclaw.
@inner cypress what do you think?
我会说中文,但不太好。用英语聊天可以吗?
哈哈
Absolutely, no problem at all. For any issue you find, you can contribute by submitting a PR.
You can ping me and I can have a quickly review
By the way, the QQ channel has been integrated in openclaw
but there are also still many things we can work on, such as robustness and security
Yes, I agree. OpenClaw has been facing many issues recently. I see that they are fixing many of them, and similar issues can occur with QQ channels too.
Checked #78890. It is valid on current upstream/main at 238e72d74d80, and npm openclaw@latest is still 2026.5.6.
Source proof:
extensions/qqbot/src/engine/gateway/gateway-connection.ts:265handlesGatewayOp.RECONNECTby callingcleanup()and scheduling reconnect.extensions/qqbot/src/engine/gateway/gateway-connection.ts:289then always routes the WebSocket close event intohandleClose(code).extensions/qqbot/src/engine/gateway/reconnect.ts:133groups4009 SESSION_TIMEOUTwith invalid-session codes and returnsclearSession: true.extensions/qqbot/src/engine/gateway/gateway-connection.ts:314only sendsRESUMEwhensessionIdandlastSeqare still present.
I also ran a direct source probe for handleClose(4009, false), and it returns:
{"shouldReconnect":true,"clearSession":true,"refreshToken":true,"fatal":false,"reason":"session timed out"}
That conflicts with the official QQ WebSocket docs, which say 4009 is an expired connection and can retry RESUME: https://bot.q.qq.com/wiki/develop/api-v2/dev-prepare/error-trace/websocket.html
There is already an open same-author fix PR: https://github.com/openclaw/openclaw/pull/78898. It touches the active file and is directionally right for the RECONNECT(7) path, but I would not call it merge-ready yet: it has no live after-fix QQBot proof, no focused regression test, and no changelog entry.
https://github.com/openclaw/openclaw/issues/78895#issuecomment-4397845646
@inner cypress Could you add your opinions?
@inner cypress Another PR is waiting for review.
https://github.com/openclaw/openclaw/pull/73567