repo stringclasses 1
value | github_id int64 3.77B 4.38B | github_node_id stringlengths 18 24 | number int64 75 77.8k | html_url stringlengths 46 49 | api_url stringlengths 56 59 | title stringlengths 1 746 | body stringlengths 1 55.7k ⌀ | state stringclasses 2
values | state_reason stringclasses 4
values | locked bool 2
classes | comments_count int64 0 128 | labels listlengths 0 5 | assignees listlengths 0 2 | created_at stringdate 2026-01-01 16:55:34 2026-05-05 11:59:03 | updated_at stringdate 2026-03-24 00:13:29 2026-05-05 11:59:15 | closed_at stringlengths 20 20 ⌀ | author_association stringclasses 3
values | milestone_title stringclasses 0
values | snapshot_id stringclasses 76
values | extracted_at stringdate 2026-03-24 23:36:49 2026-05-05 12:00:33 | author_login stringlengths 3 38 | author_id int64 346 282M | author_node_id stringlengths 12 20 | author_type stringclasses 2
values | author_site_admin bool 2
classes |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
openclaw/openclaw | 4,080,115,030 | I_kwDOQb6kR87zMZ1W | 47,876 | https://github.com/openclaw/openclaw/issues/47876 | https://api.github.com/repos/openclaw/openclaw/issues/47876 | Docs/Observability: session-focused monitoring guide + community tool (ClawMonitor) | ## Context
In self-hosted, always-on deployments (especially multi-channel: Telegram + Feishu + WhatsApp, etc.), we frequently hit cases where the Gateway and channel probes look OK but users still experience “agent is working but chat didn’t reply”. Typical examples:
- Long-polling stalls intermittently (inbound upd... | closed | not_planned | true | 1 | [] | [] | 2026-03-16T04:15:39Z | 2026-04-29T04:38:28Z | 2026-04-26T09:47:09Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | openclawq | 258,709,591 | U_kgDOD2uYVw | User | false |
openclaw/openclaw | 4,080,461,905 | I_kwDOQb6kR87zNuhR | 47,946 | https://github.com/openclaw/openclaw/issues/47946 | https://api.github.com/repos/openclaw/openclaw/issues/47946 | [Bug]: 400 Bad Request when calling vLLM /v1/chat/completions from OpenClaw | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
## Environment
- vLLM image: vllm/vllm-openai:v0.6.3
- Model: Qwen2.5-72B-Instruct-AWQ
- Quantization: AWQ
- Client: OpenClaw
- Endpoint: POST /v1/chat/completions
## Describe the bug
When using **OpenClaw** to call the locally deployed vL... | closed | not_planned | true | 6 | [
"bug",
"bug:behavior"
] | [] | 2026-03-16T06:13:54Z | 2026-04-29T04:38:27Z | 2026-04-26T09:47:13Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ChangerChen07 | 18,086,537 | MDQ6VXNlcjE4MDg2NTM3 | User | false |
openclaw/openclaw | 4,080,977,538 | I_kwDOQb6kR87zPsaC | 48,032 | https://github.com/openclaw/openclaw/issues/48032 | https://api.github.com/repos/openclaw/openclaw/issues/48032 | [Bug]: node-host exits after gateway disconnect instead of retrying | ### Bug type
Crash (process/app exits or hangs)
### Summary
When a long-running `openclaw node run ...` process loses its gateway websocket, the node-host exits instead of remaining alive and retrying.
I can reproduce this reliably during gateway restarts. This appears to be a small, self-contained reconnect lifecy... | closed | completed | true | 7 | [
"bug",
"bug:crash"
] | [] | 2026-03-16T08:25:05Z | 2026-04-29T04:38:27Z | 2026-04-26T09:47:17Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | AnthonyDiSanti | 911,470 | MDQ6VXNlcjkxMTQ3MA== | User | false |
openclaw/openclaw | 4,080,995,364 | I_kwDOQb6kR87zPwwk | 48,036 | https://github.com/openclaw/openclaw/issues/48036 | https://api.github.com/repos/openclaw/openclaw/issues/48036 | Control UI WebSocket heartbeat timeout popup appears frequently | ## 问题描述
Control UI 页面打开后，频繁弹出 WebSocket 心跳超时错误：
\\\
TERMINATE SOCKET: Ping Pong does not transfer heartbeat within heartbeat interval
\\\
## 复现步骤
1. 打开 Control UI 页面
2. 放置一段时间不操作
3. 错误弹窗不断出现
## 环境信息
- OpenClaw 版本：最新 stable
- Gateway：local 模式，绑定 127.0.0.1:18789
## 期望行为
应该减少不必要的弹窗，或者增加心跳间隔时间
## 备注
不影响核心功能，但体验不佳 | closed | completed | true | 3 | [] | [] | 2026-03-16T08:28:30Z | 2026-04-29T04:38:26Z | 2026-04-26T09:47:21Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Nerdless-ship-it | 257,649,571 | U_kgDOD1trow | User | false |
openclaw/openclaw | 4,081,263,227 | I_kwDOQb6kR87zQyJ7 | 48,087 | https://github.com/openclaw/openclaw/issues/48087 | https://api.github.com/repos/openclaw/openclaw/issues/48087 | UI 显示 401 错误但 API Key 实际可用 - 需手动刷新页面恢复 | ## 问题描述
在 Web UI 中看到 `HTTP 401 authentication_error: invalid api key` 错误，但实际 API Key 是可用的。
## 复现步骤
1. 正常使用时，UI 突然显示 401 错误
2. 刷新页面后恢复正常
3. 再次正常使用，一段时间后又出现同样问题
## 日志证据
从 `gateway.err.log` 中提取的错误：
```
2026-03-16 13:39:14.376+08:00 [agent/embedded] embedded run agent end: ... error=HTTP 401 authentication_error: inva... | closed | completed | true | 4 | [] | [] | 2026-03-16T09:25:26Z | 2026-04-29T04:38:25Z | 2026-04-26T09:47:25Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | yuxiao345 | 60,804,273 | MDQ6VXNlcjYwODA0Mjcz | User | false |
openclaw/openclaw | 4,083,063,943 | I_kwDOQb6kR87zXpyH | 48,294 | https://github.com/openclaw/openclaw/issues/48294 | https://api.github.com/repos/openclaw/openclaw/issues/48294 | Feature: configurable memoryFlush write path | ## Problem
When `memoryFlush` (automatic session-to-memory compaction) fires, it writes to a hardcoded path:
```
memory/${YYYY-MM-DD}.md
```
This is defined in `resolveMemoryFlushRelativePathForRun()` in the dist source. The path is not configurable.
## Use Case
As workspaces accumulate daily notes over months, ha... | closed | not_planned | true | 1 | [] | [] | 2026-03-16T15:08:40Z | 2026-04-29T04:38:24Z | 2026-04-26T09:47:29Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | blortski | 117,885,527 | U_kgDOBwbKVw | User | false |
openclaw/openclaw | 4,083,298,270 | I_kwDOQb6kR87zYi_e | 48,319 | https://github.com/openclaw/openclaw/issues/48319 | https://api.github.com/repos/openclaw/openclaw/issues/48319 | Signal: media placeholder '<media:audio>' visible as caption on voice messages | ## Bug Description
When sending a voice message (TTS or media attachment) via Signal channel, the placeholder text `<media:audio>` appears as a visible caption/body text below the audio waveform in the Signal app.
## Root Cause
In `src/signal/send.ts` (compiled as `dist/send-BogFunkR.js`), when sending a media attac... | closed | not_planned | true | 2 | [] | [] | 2026-03-16T15:48:19Z | 2026-04-29T04:38:23Z | 2026-04-26T09:47:32Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | hadrien-rose | 91,693,219 | U_kgDOBXcgow | User | false |
openclaw/openclaw | 4,085,002,121 | I_kwDOQb6kR87zfC-J | 48,511 | https://github.com/openclaw/openclaw/issues/48511 | https://api.github.com/repos/openclaw/openclaw/issues/48511 | Bug: two parallel codex exec runs can destabilize Telegram session/gateway | ## Summary
Launching two parallel `codex exec --full-auto` runs from a Telegram-driven OpenClaw session can destabilize the session/gateway to the point that the assistant stops responding and the gateway may need to be restarted.
Single Codex runs work. The failure appears when attempting concurrent Codex subprocess... | closed | completed | true | 2 | [] | [] | 2026-03-16T22:00:39Z | 2026-04-29T04:38:23Z | 2026-04-26T09:47:36Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Aniloid2 | 17,166,325 | MDQ6VXNlcjE3MTY2MzI1 | User | false |
openclaw/openclaw | 4,085,094,850 | I_kwDOQb6kR87zfZnC | 48,526 | https://github.com/openclaw/openclaw/issues/48526 | https://api.github.com/repos/openclaw/openclaw/issues/48526 | [Bug]: Anthropic provider stuck in “billing issue” state even after credits + new API key + clean onboard | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Environment:
- OS: macOS 26.3 (arm64)
- OpenClaw: 2026.3.13 (61d171a)
- Gateway mode: local (loopback, ws://127.0.0.1:18789)
- Agent: main (Poncho)
- Provider: anthropic
Problem summary:
After a temporary Anthropic billing error (“Your cre... | closed | completed | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-03-16T22:24:38Z | 2026-04-29T04:38:22Z | 2026-04-26T09:47:40Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | bigsuperhero91-cell | 264,156,209 | U_kgDOD760MQ | User | false |
openclaw/openclaw | 4,085,796,169 | I_kwDOQb6kR87ziE1J | 48,644 | https://github.com/openclaw/openclaw/issues/48644 | https://api.github.com/repos/openclaw/openclaw/issues/48644 | Gateway startup takes 4-10 minutes (603 JS chunks, 2.2GB memory peak) | # Gateway startup takes 4-10 minutes (603 JS chunks, 2.2GB memory peak)
## Summary
Gateway takes ~4 minutes on warm restart and ~10 minutes on cold boot before it begins accepting connections. During startup, Node.js consumes up to 2.2GB of memory and produces no log output — it appears to be parsing/compiling the JS... | closed | completed | true | 6 | [] | [] | 2026-03-17T02:12:10Z | 2026-04-29T04:38:21Z | 2026-04-26T09:47:44Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | smischie | 14,812,114 | MDQ6VXNlcjE0ODEyMTE0 | User | false |
openclaw/openclaw | 4,085,944,354 | I_kwDOQb6kR87zipAi | 48,691 | https://github.com/openclaw/openclaw/issues/48691 | https://api.github.com/repos/openclaw/openclaw/issues/48691 | healthcheck: update recommendation should respect installation method (brew vs npm) | ## Problem
The `healthcheck:security-audit` cron recommends running `pnpm update` (or `npm update`) when a new OpenClaw version is available, regardless of how OpenClaw was installed.
Example alert:
> Update Available: pnpm latest 2026.3.13
However, if OpenClaw was installed via **Homebrew**, the correct update comm... | closed | completed | true | 2 | [] | [] | 2026-03-17T03:10:22Z | 2026-04-29T04:38:21Z | 2026-04-26T09:47:48Z | CONTRIBUTOR | null | 20260429T060025Z | 2026-04-29T06:00:25Z | dongzhenye | 5,765,843 | MDQ6VXNlcjU3NjU4NDM= | User | false |
openclaw/openclaw | 4,085,981,795 | I_kwDOQb6kR87ziyJj | 48,694 | https://github.com/openclaw/openclaw/issues/48694 | https://api.github.com/repos/openclaw/openclaw/issues/48694 | Slack Socket Mode bots return account_inactive loop despite valid tokens (2026.2.23) | ### Bug type
Regression (worked before, now fails)
### Summary
Title: Slack Socket Mode bots return account_inactive loop despite valid tokens (2026.2.23)
Summary
On OpenClaw 2026.2.23, all 16 Slack Socket Mode bots in one workspace suddenly started failing with Error: An API error occurred: account_inactive in a ... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-03-17T03:25:23Z | 2026-04-29T04:38:19Z | 2026-04-26T09:47:52Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | krispyking | 19,603,863 | MDQ6VXNlcjE5NjAzODYz | User | false |
openclaw/openclaw | 4,087,093,749 | I_kwDOQb6kR87znBn1 | 48,869 | https://github.com/openclaw/openclaw/issues/48869 | https://api.github.com/repos/openclaw/openclaw/issues/48869 | Docs: Wakeup Script Best Practices — Avoid Silent Agent Wake Failures | ## Summary
This issue proposes adding documentation for wakeup script patterns to prevent silent agent wake failures. The common pitfall: scripts create inbox files but don't trigger the agent session, leading to unprocessed tasks.
## Problem Pattern
Many users create wakeup scripts like this:
```bash
#!/bin/bash
#... | closed | not_planned | true | 2 | [] | [] | 2026-03-17T08:53:06Z | 2026-04-29T04:38:18Z | 2026-04-26T09:47:56Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | hhdnxx | 22,232,872 | MDQ6VXNlcjIyMjMyODcy | User | false |
openclaw/openclaw | 4,087,256,219 | I_kwDOQb6kR87znpSb | 48,892 | https://github.com/openclaw/openclaw/issues/48892 | https://api.github.com/repos/openclaw/openclaw/issues/48892 | Telegram: 421 Misdirected Request not retried, causing silent message drops | ## Description
When sending messages via the Telegram Bot API, occasional HTTP 421 (Misdirected Request) errors occur due to HTTP/2 connection coalescing. These are not retried, causing replies to be silently dropped.
## Impact
The agent appears unresponsive to the user — messages are received but replies never arri... | closed | not_planned | true | 2 | [] | [] | 2026-03-17T09:24:11Z | 2026-04-29T04:38:17Z | 2026-04-26T09:48:00Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | therealalcapino | 260,289,127 | U_kgDOD4OyZw | User | false |
openclaw/openclaw | 4,330,557,663 | I_kwDOQb6kR88AAAABAh8Q3w | 72,095 | https://github.com/openclaw/openclaw/issues/72095 | https://api.github.com/repos/openclaw/openclaw/issues/72095 | orchestrator store: in_progress tasks have no expiry path | ## Context
Found during bot review of [#72086](https://github.com/openclaw/openclaw/pull/72086) (Greptile P2). The `STALE_ELIGIBLE` set in [extensions/orchestrator/src/store.ts](https://github.com/openclaw/openclaw/blob/main/extensions/orchestrator/src/store.ts) excludes \`in_progress\`:
\`\`\`ts
const STALE_ELIGIBLE... | closed | not_planned | true | 1 | [] | [] | 2026-04-26T09:48:19Z | 2026-04-29T04:36:13Z | 2026-04-26T12:08:42Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | PeterPlatinum | 85,887,810 | MDQ6VXNlcjg1ODg3ODEw | User | false |
openclaw/openclaw | 4,216,487,295 | I_kwDOQb6kR877Un1_ | 62,400 | https://github.com/openclaw/openclaw/issues/62400 | https://api.github.com/repos/openclaw/openclaw/issues/62400 | [Bug]: CLI Local Commands Hang Indefinitely in OpenClaw 4.5 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
CLI local commands (skills check/list, help) hang indefinitely after execution, consuming 90%+ CPU and never exiting.
### Steps to reproduce
Install OpenClaw 2026.4.5 on Linux/WSL2
Run: openclaw skills check
Observe: Com... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-07T08:37:16Z | 2026-04-29T04:38:16Z | 2026-04-26T09:50:21Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | shenyiqi-2004 | 221,808,311 | U_kgDODTiGtw | User | false |
openclaw/openclaw | 4,216,936,443 | I_kwDOQb6kR877WVf7 | 62,422 | https://github.com/openclaw/openclaw/issues/62422 | https://api.github.com/repos/openclaw/openclaw/issues/62422 | macOS app: schema unavailable; cannot authenticate to local Docker gateway (token mismatch) | ## Summary
OpenClaw macOS app cannot configure/connect to a local Docker gateway. In the app’s config section it shows **Schema unavailable** / **No config selection available**. Gateway logs show repeated handshake failures with **unauthorized: gateway token mismatch**.
## Environment
- macOS: Darwin 24.6.0
- OpenCla... | closed | completed | true | 1 | [] | [] | 2026-04-07T10:05:38Z | 2026-04-29T04:38:15Z | 2026-04-26T09:50:26Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | BarathCommits | 28,656,244 | MDQ6VXNlcjI4NjU2MjQ0 | User | false |
openclaw/openclaw | 4,218,102,441 | I_kwDOQb6kR877ayKp | 62,526 | https://github.com/openclaw/openclaw/issues/62526 | https://api.github.com/repos/openclaw/openclaw/issues/62526 | OpenClaw: add /hooks/message ingress with auth parity and clear hook boundaries | ## Scope
Add a generic `POST /hooks/message` ingress endpoint with auth and request-safety parity to existing hook routes, and enforce clear responsibilities between `/hooks/agent` and `/hooks/message`.
## Problem
OpenClaw supports webhook wake/agent entrypoints, but there is no generic webhook message ingress route ... | closed | not_planned | true | 1 | [] | [] | 2026-04-07T13:47:08Z | 2026-04-29T04:38:13Z | 2026-04-26T09:50:39Z | CONTRIBUTOR | null | 20260429T060025Z | 2026-04-29T06:00:25Z | vrknetha | 20,596,261 | MDQ6VXNlcjIwNTk2MjYx | User | false |
openclaw/openclaw | 4,219,363,746 | I_kwDOQb6kR877fmGi | 62,613 | https://github.com/openclaw/openclaw/issues/62613 | https://api.github.com/repos/openclaw/openclaw/issues/62613 | [Bug]: CLI ReferenceError on every command in v2026.4.5 (Node 22.16.0, macOS ARM64) | ## Bug Description
Every CLI command in v2026.4.5 crashes with `ReferenceError: Cannot access '...' before initialization` immediately on startup, before any user input.
## Environment
- **OpenClaw**: v2026.4.5 (3e72c03)
- **Node.js**: v22.16.0
- **npm**: 10.9.2
- **OS**: macOS 24.6.0 (Darwin ARM64)
- **Install meth... | closed | completed | true | 2 | [] | [] | 2026-04-07T17:08:29Z | 2026-04-29T04:38:12Z | 2026-04-26T09:50:43Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | nemrry | 11,785,803 | MDQ6VXNlcjExNzg1ODAz | User | false |
openclaw/openclaw | 4,219,471,899 | I_kwDOQb6kR877gAgb | 62,630 | https://github.com/openclaw/openclaw/issues/62630 | https://api.github.com/repos/openclaw/openclaw/issues/62630 | Docker to host state migration leaks /home/node absolute paths and breaks agent runtime with EACCES | ## Summary
Migrating OpenClaw state from a Docker deployment to a native Linux host install can leave hidden absolute path references to `/home/node`. The gateway and CLI appear healthy, but agent execution fails at chat/runtime with:
Agent failed before reply: EACCES: permission denied, mkdir '/home/node'
## Enviro... | closed | not_planned | true | 1 | [] | [] | 2026-04-07T17:28:32Z | 2026-04-29T04:38:11Z | 2026-04-26T09:50:48Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Jmarzab | 196,876,602 | U_kgDOC7wZOg | User | false |
openclaw/openclaw | 4,219,479,621 | I_kwDOQb6kR877gCZF | 62,632 | https://github.com/openclaw/openclaw/issues/62632 | https://api.github.com/repos/openclaw/openclaw/issues/62632 | Docker to host state migration leaks /home/node absolute paths and breaks agent runtime with EACCES | Docker -> host migration leaks /home/node absolute paths and can break agent runtime with:
Agent failed before reply: EACCES: permission denied, mkdir '/home/node'
Environment:
- Source deployment: Docker
- Target deployment: native Linux host install
- State copied from ~/openclaw-docker/data to ~/.openclaw
- Gatewa... | closed | not_planned | true | 3 | [] | [] | 2026-04-07T17:29:54Z | 2026-04-29T04:38:11Z | 2026-04-26T09:50:52Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Jmarzab | 196,876,602 | U_kgDOC7wZOg | User | false |
openclaw/openclaw | 4,221,135,176 | I_kwDOQb6kR877mWlI | 62,756 | https://github.com/openclaw/openclaw/issues/62756 | https://api.github.com/repos/openclaw/openclaw/issues/62756 | Please focus on stability instead of shipping new features | ## The Problem
OpenClaw ships new features constantly, but the existing functionality is plagued with bugs that make daily usage a struggle. Every day I spend more time debugging OpenClaw than it saves me. My daily token usage is 700-800 million, and 90% of it goes to debugging OpenClaw itself instead of actual work.
... | closed | not_planned | true | 2 | [] | [] | 2026-04-07T23:26:34Z | 2026-04-29T04:38:10Z | 2026-04-26T09:50:56Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | oldjs | 183,616,870 | U_kgDOCvHFZg | User | false |
openclaw/openclaw | 4,221,296,175 | I_kwDOQb6kR877m94v | 62,776 | https://github.com/openclaw/openclaw/issues/62776 | https://api.github.com/repos/openclaw/openclaw/issues/62776 | [Feature]: Support OAuth-based remote MCP authentication flows (Heptabase MCP) | ### Summary
Support OAuth-based authentication for remote MCP servers such as Heptabase MCP.
### Problem to solve
I can register a remote MCP server in OpenClaw using its URL, and OpenClaw does attempt to connect to it. However, some MCP servers such as Heptabase require OAuth authentication rather than a static be... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-08T00:15:15Z | 2026-04-29T04:38:09Z | 2026-04-26T09:51:01Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | selahinsight | 271,418,027 | U_kgDOEC2Cqw | User | false |
openclaw/openclaw | 4,330,564,867 | I_kwDOQb6kR88AAAABAh8tAw | 72,097 | https://github.com/openclaw/openclaw/issues/72097 | https://api.github.com/repos/openclaw/openclaw/issues/72097 | [Feature]: sync tool-injection into active agent sessions on plugins install | ## Summary
When `openclaw plugins install <pkg>` succeeds and the plugin registers new tools, those tools are not injected into already-active agent CLI sessions until the session is restarted.
## Problem to solve
Plugins that add tools (memory plugins, integration plugins, etc.) ship tool definitions that the agent... | closed | completed | true | 1 | [] | [] | 2026-04-26T09:52:27Z | 2026-05-05T04:27:15Z | 2026-05-02T13:19:26Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | p-diogo | 8,688,948 | MDQ6VXNlcjg2ODg5NDg= | User | false |
openclaw/openclaw | 4,330,564,194 | I_kwDOQb6kR88AAAABAh8qYg | 72,096 | https://github.com/openclaw/openclaw/issues/72096 | https://api.github.com/repos/openclaw/openclaw/issues/72096 | [Bug]: Dreaming light-phase work summary repeats verbatim across hourly cycles, producing apparent duplicates in DREAMS.md | ## Summary
During overnight dreaming, the light-phase work summary block emits **identical** commit-list content every cycle for ~6 hours, producing apparent duplicates in DREAMS.md. Only the REM reflection block (which precedes each work summary) shows incrementing memory counts, suggesting the REM input pipeline is ... | open | null | false | 1 | [] | [] | 2026-04-26T09:52:04Z | 2026-05-03T07:45:05Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | aaajiao | 49,885 | MDQ6VXNlcjQ5ODg1 | User | false |
openclaw/openclaw | 4,091,377,256 | I_kwDOQb6kR87z3XZo | 49,258 | https://github.com/openclaw/openclaw/issues/49258 | https://api.github.com/repos/openclaw/openclaw/issues/49258 | Bug: Cron job state inconsistency - lastDelivered: true but lastRunStatus: error | ### Bug type
Regression (worked before, now fails)
### Summary
Cron job shows contradictory state: lastDelivered=true (message sent successfully) but lastRunStatus=error (marked as failed), causing false positive error alerts and accumulated consecutiveErrors despite successful delivery.
### Steps to reproduce
1. ... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-03-17T21:47:49Z | 2026-04-29T04:38:03Z | 2026-04-26T10:03:06Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | nariloy-bit | 245,159,206 | U_kgDODpzVJg | User | false |
openclaw/openclaw | 4,091,902,220 | I_kwDOQb6kR87z5XkM | 49,316 | https://github.com/openclaw/openclaw/issues/49316 | https://api.github.com/repos/openclaw/openclaw/issues/49316 | [Bug]: mlx-vlm report 0% context usage — prompt_tokens vs input_tokens field mismatch | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When using mlx-vlm as a local model server via `openai-completions` API, the Telegram status always shows `Context: 0/Xk (0%)` regardless of actual token usage.
## Root Cause
mlx-vlm returns usage fields with different names than the Open... | closed | not_planned | true | 6 | [
"bug",
"bug:behavior"
] | [] | 2026-03-18T00:17:18Z | 2026-04-29T04:38:02Z | 2026-04-26T10:03:11Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | gw-space | 104,241,879 | U_kgDOBjaa1w | User | false |
openclaw/openclaw | 4,094,664,521 | I_kwDOQb6kR870D59J | 49,777 | https://github.com/openclaw/openclaw/issues/49777 | https://api.github.com/repos/openclaw/openclaw/issues/49777 | [Bug] Webchat UI 未能正确渲染完整会话历史 | ## 问题摘要
Webchat UI 未能正确渲染/同步完整的会话历史。虽然所有消息都正确存储在后端 transcript 文件中，但用户界面显示不完整，导致用户看到"对话丢失"的假象。
## 环境信息
- **操作系统**: Windows 10 (Build 19042)
- **浏览器/客户端**: Webchat (Electron)
- **时间**: 2026-03-18
## 问题现象
1. 用户在 webchat 中进行多轮对话
2. 对话过程中，UI 上部分消息"消失"或未能显示
3. 用户确认没有刷新页面或重新打开窗口
4. 后端数据完整，但前端展示不完整
## 诊断数据
### 会话信息
- **... | closed | completed | true | 3 | [] | [] | 2026-03-18T11:58:45Z | 2026-04-29T04:38:01Z | 2026-04-26T10:03:16Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Troy668 | 203,381,894 | U_kgDODB9chg | User | false |
openclaw/openclaw | 4,095,143,129 | I_kwDOQb6kR870FuzZ | 49,823 | https://github.com/openclaw/openclaw/issues/49823 | https://api.github.com/repos/openclaw/openclaw/issues/49823 | Add per-run cost cap for embedded runs (maxCostPerRun) | Embedded runs (heartbeat, cron, future trigger types) have no per-run cost bound. Merged fixes in #3181 reduce trigger frequency (dedup, retry guards, HEARTBEAT_OK leak), but a single run on a large context with an expensive model is still unbounded. #46798 shows new trigger paths bypass dedup entirely. One Opus 4.6 ru... | closed | not_planned | true | 5 | [] | [] | 2026-03-18T13:28:39Z | 2026-04-29T04:37:59Z | 2026-04-26T10:03:26Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | amabito | 192,487,536 | U_kgDOC3kgcA | User | false |
openclaw/openclaw | 4,095,756,344 | I_kwDOQb6kR870IEg4 | 49,865 | https://github.com/openclaw/openclaw/issues/49865 | https://api.github.com/repos/openclaw/openclaw/issues/49865 | Gateway stability issues on Windows: SecretRef resolution, RPC timeouts, cron failures | # OpenClaw Gateway Stability Report
**Operator:** Michael Tiplea (Gandalf stack)
**Version:** OpenClaw 2026.3.12 (6472949)
**Platform:** Windows 10.0.19045 (x64), Node.js v24.14.0
**Date:** 2026-03-18
---
## Executive Summary
We've been running OpenClaw on a Windows desktop since early March 2026 with Teleg... | closed | completed | true | 2 | [] | [] | 2026-03-18T15:01:33Z | 2026-04-29T04:37:58Z | 2026-04-26T10:03:31Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | mtiplea-code | 227,480,687 | U_kgDODY8Ubw | User | false |
openclaw/openclaw | 4,096,711,138 | I_kwDOQb6kR870Ltni | 49,939 | https://github.com/openclaw/openclaw/issues/49939 | https://api.github.com/repos/openclaw/openclaw/issues/49939 | Update Anthropic Claude Opus 4.6 / Sonnet 4.6 context window to 1M tokens | ## Summary
Anthropic announced on March 13, 2026 that the **1M token context window is now generally available** for Claude Opus 4.6 and Claude Sonnet 4.6 at standard pricing.
- Announcement: https://claude.com/blog/1m-context-ga
However, OpenClaw's internal model registry still uses `DEFAULT_CONTEXT_TOKENS = 2e5` (... | closed | not_planned | true | 2 | [] | [] | 2026-03-18T17:37:10Z | 2026-04-29T04:37:58Z | 2026-04-26T10:03:36Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ytangguo | 97,931,313 | U_kgDOBdZQMQ | User | false |
openclaw/openclaw | 4,096,746,660 | I_kwDOQb6kR870L2Sk | 49,943 | https://github.com/openclaw/openclaw/issues/49943 | https://api.github.com/repos/openclaw/openclaw/issues/49943 | exec: PTY zombie sessions accept routed commands after session marked done |
**Environment**
- OS: Ubuntu 24.04.4 LTS (WSL2, Linux 6.6.87.2-microsoft-standard-WSL2)
- Node: v25.8.1
- OpenClaw: 2026.3.13
- Python: 3.x in venv at /tmp/sync-test-venv
**Gateway config at time of issue**
- exec.requestTimeoutMs: 60000
- exec.maxDurationMs: 300000
- exec.idleTimeoutMs: 300000
- exec.maxConcurrent: ... | closed | completed | true | 10 | [] | [] | 2026-03-18T17:42:23Z | 2026-04-29T04:37:57Z | 2026-04-26T10:03:41Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | asimons81 | 214,744,153 | U_kgDODMy8WQ | User | false |
openclaw/openclaw | 4,099,291,675 | I_kwDOQb6kR870Vjob | 50,219 | https://github.com/openclaw/openclaw/issues/50219 | https://api.github.com/repos/openclaw/openclaw/issues/50219 | Node host WebSocket retry gets stuck after Cloudflare 502 errors | ## Environment
- OpenClaw version: 2026.3.13
- Node host: macOS, launchd agent (KeepAlive: true)
- Gateway: Linux VPS behind Cloudflare tunnel (WSS)
## Problem
When the Cloudflare tunnel returns HTTP 502 (Bad Gateway) responses, the node host enters a retry loop that never recovers — even after Cloudflare starts respo... | closed | completed | true | 2 | [] | [] | 2026-03-19T04:58:01Z | 2026-04-29T04:37:55Z | 2026-04-26T10:03:50Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | pmgreenjr | 26,364,594 | MDQ6VXNlcjI2MzY0NTk0 | User | false |
openclaw/openclaw | 4,099,889,975 | I_kwDOQb6kR870X1s3 | 50,306 | https://github.com/openclaw/openclaw/issues/50306 | https://api.github.com/repos/openclaw/openclaw/issues/50306 | [Bug]: LINE channels binding loses everyday | ### Bug type
Regression (worked before, now fails)
### Summary
I have two agents linking two different line channels of same line account.
But everyday I have to change something in the gateway and save the setting. Then the channels will be back (in LINE developer portal verify success)
My machines always up and ... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-03-19T07:32:26Z | 2026-04-29T04:37:55Z | 2026-04-26T10:03:55Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | andy-au | 114,204,029 | U_kgDOBs6dfQ | User | false |
openclaw/openclaw | 4,101,122,088 | I_kwDOQb6kR870cigo | 50,398 | https://github.com/openclaw/openclaw/issues/50398 | https://api.github.com/repos/openclaw/openclaw/issues/50398 | coding-agent skill should recommend sessions_spawn over exec background for completion notification | ## Problem
The bundled `coding-agent` skill (SKILL.md) instructs agents to use `exec` with `background: true` for long-running coding tasks, and `openclaw system event --text '...' --mode now` for completion notification.
This notification path is broken:
1. `exec background` → process exits → `notifyOnExit` enqueue... | closed | completed | true | 2 | [] | [] | 2026-03-19T11:41:08Z | 2026-04-29T04:37:54Z | 2026-04-26T10:05:22Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | tymrtn | 4,305,058 | MDQ6VXNlcjQzMDUwNTg= | User | false |
openclaw/openclaw | 4,103,114,811 | I_kwDOQb6kR870kJA7 | 50,553 | https://github.com/openclaw/openclaw/issues/50553 | https://api.github.com/repos/openclaw/openclaw/issues/50553 | [Bug] ! Deleting agent orphans usage stats even with deleteFiles=false ! | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Deleting an agent removes its contribution to token usage statistics and conversation counts in the Control UI even when deleteFiles=false is used and session files remain on disk.
### Steps to reproduce
1. Create an agent and generate so... | closed | not_planned | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-03-19T17:21:05Z | 2026-04-29T04:37:53Z | 2026-04-26T10:05:26Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | PM-Geeker | 143,123,392 | U_kgDOCIfjwA | User | false |
openclaw/openclaw | 4,106,680,119 | I_kwDOQb6kR870xvc3 | 50,944 | https://github.com/openclaw/openclaw/issues/50944 | https://api.github.com/repos/openclaw/openclaw/issues/50944 | WhatsApp outbound: missing quote-reply and @mention support | ## Summary
WhatsApp outbound messages do not support quote-replies or @mentions. Both features work correctly on **inbound** (OpenClaw parses quoted messages and `mentionedJid` from incoming messages), but the **outbound** `sendMessage` implementation does not wire them up.
## Current Behavior
### Quote-replies
The ... | closed | not_planned | true | 2 | [] | [] | 2026-03-20T08:53:22Z | 2026-04-29T04:37:52Z | 2026-04-26T10:05:31Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | P7AC1D | 7,148,677 | MDQ6VXNlcjcxNDg2Nzc= | User | false |
openclaw/openclaw | 4,107,091,743 | I_kwDOQb6kR870zT8f | 50,985 | https://github.com/openclaw/openclaw/issues/50985 | https://api.github.com/repos/openclaw/openclaw/issues/50985 | [Feature]: Gateway-level suppression of intermediate text output (pre-tool narration) | ## Problem
In multi-agent Slack deployments, language models frequently emit **intermediate text** before or between tool calls — "Let me check...", "I'll look into that...", process narration. This text is posted to the channel as a visible message, creating noise.
### Why prompt engineering is insufficient
- Syste... | closed | not_planned | true | 2 | [] | [] | 2026-03-20T10:30:06Z | 2026-04-29T04:37:52Z | 2026-04-26T10:05:35Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Asclepeion | 169,803,573 | U_kgDOCh7_NQ | User | false |
openclaw/openclaw | 4,107,588,407 | I_kwDOQb6kR8701NM3 | 51,033 | https://github.com/openclaw/openclaw/issues/51033 | https://api.github.com/repos/openclaw/openclaw/issues/51033 | Docs: Add 'Out-of-the-Box Capabilities' reference page per platform | ## Problem
New users install OpenClaw expecting it to "just work" with their local devices, smart home, network services, etc. The current Getting Started guide covers installation and first message, but **nowhere in the docs does it clearly state what the agent can and cannot do out of the box** — or what requires ad... | closed | completed | true | 2 | [] | [] | 2026-03-20T12:23:04Z | 2026-04-29T04:37:51Z | 2026-04-26T10:05:39Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | SQLOPTIMISE | 128,770,495 | U_kgDOB6zhvw | User | false |
openclaw/openclaw | 4,330,600,202 | I_kwDOQb6kR88AAAABAh-3Cg | 72,104 | https://github.com/openclaw/openclaw/issues/72104 | https://api.github.com/repos/openclaw/openclaw/issues/72104 | [Bug] Gateway crashes during npm upgrade - SIGUSR1 arrives before files are written | ## Bug Description
Gateway crashes and enters a restart loop during `npm install -g openclaw@latest` because the SIGUSR1 hot-restart signal arrives while npm is still writing new package files, causing ERR_MODULE_NOT_FOUND errors.
## Steps to Reproduce
1. Run `openclaw update.run` (or `npm install -g openclaw@latest... | closed | completed | true | 1 | [] | [] | 2026-04-26T10:11:12Z | 2026-04-29T04:36:11Z | 2026-04-26T12:08:49Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | caimao9539 | 16,041,782 | MDQ6VXNlcjE2MDQxNzgy | User | false |
openclaw/openclaw | 4,106,085,846 | I_kwDOQb6kR870veXW | 50,891 | https://github.com/openclaw/openclaw/issues/50891 | https://api.github.com/repos/openclaw/openclaw/issues/50891 | Session reset hooks: idle/daily reset do not trigger session-memory save | ## Problem
There are three ways to reset a session in OpenClaw:
1. **Manual**: `/new` or `/reset` command — triggers `session-memory` hook ✅
2. **Idle timeout**: `session.reset.idleMinutes` — does NOT trigger `session-memory` ❌
3. **Daily reset**: default 4:00 AM — does NOT trigger `session-memory` ❌
The hooks syste... | closed | not_planned | true | 8 | [] | [] | 2026-03-20T06:27:56Z | 2026-04-29T04:37:50Z | 2026-04-26T10:19:32Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | YoungAndSure | 20,354,109 | MDQ6VXNlcjIwMzU0MTA5 | User | false |
openclaw/openclaw | 4,107,817,205 | I_kwDOQb6kR8702FD1 | 51,055 | https://github.com/openclaw/openclaw/issues/51055 | https://api.github.com/repos/openclaw/openclaw/issues/51055 | [Bug]: Onboard hardcodes contextWindow to 128k for all copilot-proxy/LiteLLM models, ignoring actual model capabilities | ## Bug Description
During onboard, the copilot-proxy and LiteLLM provider plugins hardcode `contextWindow: 128000` and `maxTokens: 8192` for all discovered models, regardless of the actual model capabilities reported by the proxy.
This causes **premature context compaction** for models with larger context windows. Fo... | closed | not_planned | true | 3 | [] | [] | 2026-03-20T13:12:59Z | 2026-04-29T04:37:49Z | 2026-04-26T10:19:36Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | litaohz | 21,356,197 | MDQ6VXNlcjIxMzU2MTk3 | User | false |
openclaw/openclaw | 4,109,325,341 | I_kwDOQb6kR87071Qd | 51,178 | https://github.com/openclaw/openclaw/issues/51178 | https://api.github.com/repos/openclaw/openclaw/issues/51178 | [Feature]: Decoupling Lane Queue and Asynchronous Transformation | ### Summary
When handling tool invocation and agent communication, the core loop (Agent Runner) of OpenClaw often suspends the current Lane until it receives a response.
### Problem to solve
Address the issue of frequent delays or even blockages in communication among multiple agents
### Proposed solution
Impleme... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-03-20T18:06:06Z | 2026-04-29T04:37:48Z | 2026-04-26T10:19:44Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | rory7 | 16,591,232 | MDQ6VXNlcjE2NTkxMjMy | User | false |
openclaw/openclaw | 4,110,797,149 | I_kwDOQb6kR871Bcld | 51,355 | https://github.com/openclaw/openclaw/issues/51355 | https://api.github.com/repos/openclaw/openclaw/issues/51355 | Document Slack attachment vision behavior and known limits | ## Summary
Update runbooks/docs with the final Slack attachment vision behavior, validation evidence, and any remaining known limits.
## Tasks
- document final Slack media propagation behavior
- document live verification outcomes
- document known unsupported or partially supported cases
- link implementation PR(s) an... | closed | completed | true | 1 | [] | [] | 2026-03-21T00:43:46Z | 2026-04-29T04:29:06Z | 2026-04-26T10:19:48Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | UGDev-Chane | 180,237,537 | U_kgDOCr404Q | User | false |
openclaw/openclaw | 4,115,766,485 | I_kwDOQb6kR871UZzV | 52,213 | https://github.com/openclaw/openclaw/issues/52213 | https://api.github.com/repos/openclaw/openclaw/issues/52213 | WhatsApp: Voice messages in groups silently dropped — mention check runs before audio transcription | ## Environment
- OpenClaw: 2026.3.13
- macOS (arm64), Node v25.6.1
- WhatsApp channel (personal number, selfChatMode: true)
- Audio transcription: whisper-cpp CLI via tools.media.audio.models
## Problem
Voice messages in a WhatsApp group are received but never processed. Text messages in the same group work perfectly.... | closed | completed | true | 3 | [] | [] | 2026-03-22T11:39:08Z | 2026-04-29T04:37:46Z | 2026-04-26T10:19:56Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | RiadChehayeb | 264,841,962 | U_kgDOD8kq6g | User | false |
openclaw/openclaw | 4,116,711,770 | I_kwDOQb6kR871YAla | 52,418 | https://github.com/openclaw/openclaw/issues/52418 | https://api.github.com/repos/openclaw/openclaw/issues/52418 | feat: Add Feishu Task API support (飞书任务 API) | ## Description
The Feishu plugin currently supports documents, wiki, drive, bitable, and chat — but **task management** is missing.
The Feishu app already has and scopes granted, but there's no corresponding tool implementation in the plugin.
## Expected Features
- Create tasks
- Query task lists
- Update task st... | closed | not_planned | true | 1 | [] | [] | 2026-03-22T18:20:48Z | 2026-04-29T04:37:45Z | 2026-04-26T10:20:04Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | sungci | 6,957,741 | MDQ6VXNlcjY5NTc3NDE= | User | false |
openclaw/openclaw | 4,117,120,709 | I_kwDOQb6kR871ZkbF | 52,477 | https://github.com/openclaw/openclaw/issues/52477 | https://api.github.com/repos/openclaw/openclaw/issues/52477 | chrome-extension:// origins in controlUi.allowedOrigins silently fail to match | ## Summary
`chrome-extension://` origins configured in `gateway.controlUi.allowedOrigins` are silently rejected because `parseOrigin()` returns `null` for them.
## Root Cause
In `parseOrigin()`, the code does:
```js
const url = new URL(trimmed);
return {
origin: url.origin.toLowerCase(),
// ...
};
```
Node's `... | closed | not_planned | true | 4 | [] | [] | 2026-03-22T21:09:50Z | 2026-04-29T04:37:44Z | 2026-04-26T10:20:09Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | openclaw-lisa | 269,989,793 | U_kgDOEBe3oQ | User | false |
openclaw/openclaw | 4,118,788,566 | I_kwDOQb6kR871f7nW | 52,680 | https://github.com/openclaw/openclaw/issues/52680 | https://api.github.com/repos/openclaw/openclaw/issues/52680 | tools.read: out-of-range offset should not abort runs (clamp or return diagnostic) | ## Summary
The `tools.read` tool can hard-fail a run when the requested `offset` is beyond end-of-file. This is easy for models/agents to hit when the file length changes or when they guess offsets.
Today this failure aborts the entire job/run, even though it should be a recoverable/diagnostic situation.
## Environme... | closed | completed | true | 2 | [] | [] | 2026-03-23T07:16:50Z | 2026-04-29T04:37:43Z | 2026-04-26T10:20:17Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | LyHug | 40,082,700 | MDQ6VXNlcjQwMDgyNzAw | User | false |
openclaw/openclaw | 4,118,882,618 | I_kwDOQb6kR871gSk6 | 52,706 | https://github.com/openclaw/openclaw/issues/52706 | https://api.github.com/repos/openclaw/openclaw/issues/52706 | Bug: Cron Job delivery channel defaults to 'last' instead of valid channel | ## Bug: Cron Job delivery channel defaults to "last" instead of valid channel
**Description**:
When creating a new Cron Job via OpenClaw control UI, the `delivery.channel` field defaults to `"last"` which causes delivery failure for new jobs that have no previous session history.
**Steps to reproduce**:
1. Open OpenC... | closed | completed | true | 2 | [] | [] | 2026-03-23T07:37:09Z | 2026-04-29T04:37:43Z | 2026-04-26T10:20:21Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | DY991dy | 50,545,506 | MDQ6VXNlcjUwNTQ1NTA2 | User | false |
openclaw/openclaw | 4,121,124,000 | I_kwDOQb6kR871o1yg | 52,905 | https://github.com/openclaw/openclaw/issues/52905 | https://api.github.com/repos/openclaw/openclaw/issues/52905 | [Feature]: Gateway status RPC timeout with large number of agents (9000+) | ### Summary
## Environment
- **OpenClaw Version**: 2026.3.2
- **Platform**: Linux (Docker)
- **Node Version**: 22.22.0
- **Number of Agents**: 9282
- **Deployment**: Docker container
## Problem Description
When running `openclaw status` on a Gateway managing a large number of agents (9000+), the command consistentl... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-03-23T13:59:23Z | 2026-04-29T04:37:42Z | 2026-04-26T10:20:25Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | songchenghao | 4,641,267 | MDQ6VXNlcjQ2NDEyNjc= | User | false |
openclaw/openclaw | 4,117,807,520 | I_kwDOQb6kR871cMGg | 52,585 | https://github.com/openclaw/openclaw/issues/52585 | https://api.github.com/repos/openclaw/openclaw/issues/52585 | [Bug]: failed to extract archive: SafeOpenError: path is not a regular file under root | ### Bug type
Regression (worked before, now fails)
### Summary
Plugin installation fails in 2026.3.13 with SafeOpenError during tarball extraction, worked in 2026.2.23
### Steps to reproduce
1. Run `openclaw plugins install openclaw-inspur-yunjia` or `openclaw plugins install @m1heng-clawd/feishu `
2. Observe the ... | closed | completed | true | 11 | [
"bug",
"regression"
] | [] | 2026-03-23T02:37:42Z | 2026-04-29T04:37:41Z | 2026-04-26T10:20:13Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | liyifang | 17,872,922 | MDQ6VXNlcjE3ODcyOTIy | User | false |
openclaw/openclaw | 4,221,630,999 | I_kwDOQb6kR877oPoX | 62,842 | https://github.com/openclaw/openclaw/issues/62842 | https://api.github.com/repos/openclaw/openclaw/issues/62842 | [Bug]: openclaw config command extremely slow (6-9 seconds) while other commands are fast | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
`openclaw config` command execution is extremely slow (6-9 seconds) while other commands complete in milliseconds. This includes `openclaw config file`, `openclaw config get`, and `openclaw config --help`.
### Steps to rep... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-08T02:11:59Z | 2026-04-29T04:37:37Z | 2026-04-26T10:24:19Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ywt0823 | 47,940,902 | MDQ6VXNlcjQ3OTQwOTAy | User | false |
openclaw/openclaw | 4,330,490,670 | I_kwDOQb6kR88AAAABAh4LLg | 72,079 | https://github.com/openclaw/openclaw/issues/72079 | https://api.github.com/repos/openclaw/openclaw/issues/72079 | [Bug]: `openclaw health --json` fails while `openclaw gateway health --json` succeeds on the same healthy gateway | ### Bug type
Behavior bug / CLI wrapper mismatch
### Beta release blocker
No
### Summary
On a healthy local gateway, the top-level CLI command:
```bash
openclaw health --json
```
can fail or time out even though the direct gateway health paths succeed against the same running gateway:
```bash
openclaw gateway h... | closed | completed | true | 2 | [] | [] | 2026-04-26T09:10:09Z | 2026-04-29T04:36:14Z | 2026-04-26T12:08:34Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | singhanubhav-ms | 249,689,199 | U_kgDODuH0bw | User | false |
openclaw/openclaw | 4,330,672,291 | I_kwDOQb6kR88AAAABAiDQow | 72,117 | https://github.com/openclaw/openclaw/issues/72117 | https://api.github.com/repos/openclaw/openclaw/issues/72117 | [Feature]: official `npm create openclaw-plugin` scaffolder for ClawHub-ready plugins | ### Summary
A small `create-openclaw-plugin` CLI so plugin authors can scaffold a ClawHub-ready plugin in one
command: `npm create openclaw-plugin@latest my-plugin`. Generated project includes the
`openclaw.manifest`, `openclaw.compat.pluginApi`, and `openclaw.build.openclawVersion` fields ClawHub
requires, plus... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-26T10:52:41Z | 2026-04-29T04:36:09Z | 2026-04-26T12:08:57Z | CONTRIBUTOR | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Quratulain-bilal | 164,326,821 | U_kgDOCcttpQ | User | false |
openclaw/openclaw | 4,330,522,466 | I_kwDOQb6kR88AAAABAh6HYg | 72,088 | https://github.com/openclaw/openclaw/issues/72088 | https://api.github.com/repos/openclaw/openclaw/issues/72088 | [Bug]: Telegram channel missing runtime dependencies (grammy, abort-controller) since 2026.4.22 — must be installed manually after every update, openclaw doctor --fix does nothing | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Since at least 2026.4.22, the Telegram channel requires grammy, abort-controller and related @grammyjs/* packages to be installed manually after every openclaw update. These dependencies are not bundled or declared in the pa... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-26T09:28:05Z | 2026-04-29T04:36:47Z | 2026-04-26T10:54:39Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | centralpc | 59,361,530 | MDQ6VXNlcjU5MzYxNTMw | User | false |
openclaw/openclaw | 4,330,533,864 | I_kwDOQb6kR88AAAABAh6z6A | 72,089 | https://github.com/openclaw/openclaw/issues/72089 | https://api.github.com/repos/openclaw/openclaw/issues/72089 | fix(bonjour): crash loop — unhandled 'CIAO ANNOUNCEMENT/PROBING CANCELLED' rejection | ## Bug Description
OpenClaw Gateway crashes every 40–60 seconds with an unhandled promise rejection:
```
[openclaw] Unhandled promise rejection: CIAO ANNOUNCEMENT CANCELLED
[openclaw] Unhandled promise rejection: CIAO PROBING CANCELLED
```
A stability bundle is written on each crash:
`~/.openclaw/logs/stability/open... | closed | completed | true | 4 | [] | [] | 2026-04-26T09:34:32Z | 2026-04-29T04:36:46Z | 2026-04-26T10:54:44Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | wayogo | 64,243,616 | MDQ6VXNlcjY0MjQzNjE2 | User | false |
openclaw/openclaw | 4,330,594,774 | I_kwDOQb6kR88AAAABAh-h1g | 72,099 | https://github.com/openclaw/openclaw/issues/72099 | https://api.github.com/repos/openclaw/openclaw/issues/72099 | memory-core narrative cleanup invokes runtime subagent methods outside gateway-request context | ## Summary
During the nightly dreaming cycle (light/rem/deep), the `memory-core` plugin generates narratives successfully but fails during the cleanup phase with:
```
Plugin runtime subagent methods are only available during a gateway request.
```
Cleanup runs after the gateway-request context has been torn down, whi... | closed | not_planned | true | 1 | [] | [] | 2026-04-26T10:08:12Z | 2026-04-29T04:36:45Z | 2026-04-26T10:54:55Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | shandutta | 8,773,397 | MDQ6VXNlcjg3NzMzOTc= | User | false |
openclaw/openclaw | 4,330,597,185 | I_kwDOQb6kR88AAAABAh-rQQ | 72,100 | https://github.com/openclaw/openclaw/issues/72100 | https://api.github.com/repos/openclaw/openclaw/issues/72100 | OpenClaw Upgrade Checklist | ### Summary
I started building this after running into a few upgrade problems with OpenClaw.
When I upgraded, I noticed that some things did not carry over cleanly:
the config still showed an older version
a custom agent was missing or misregistered
some tools were listed in config but were not actually available in... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-04-26T10:09:30Z | 2026-04-29T04:35:31Z | 2026-04-26T10:55:00Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Cubefive | 6,971,809 | MDQ6VXNlcjY5NzE4MDk= | User | false |
openclaw/openclaw | 4,330,622,443 | I_kwDOQb6kR88AAAABAiAN6w | 72,106 | https://github.com/openclaw/openclaw/issues/72106 | https://api.github.com/repos/openclaw/openclaw/issues/72106 | Bug: Context limit exceeded error persists after modifying config and restarting service | ## Bug Description
After modifying the configuration file and restarting the OpenClaw service, the following error still occurs:
```
⚠️ Context limit exceeded. I have reset our conversation to start fresh - please try again.
To prevent this, increase your compaction buffer by setting agents.defaults.compaction.reserv... | closed | not_planned | true | 1 | [] | [] | 2026-04-26T10:23:36Z | 2026-04-29T04:36:10Z | 2026-04-26T12:08:53Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | XuXuClassMate | 76,080,484 | MDQ6VXNlcjc2MDgwNDg0 | User | false |
openclaw/openclaw | 4,330,599,514 | I_kwDOQb6kR88AAAABAh-0Wg | 72,103 | https://github.com/openclaw/openclaw/issues/72103 | https://api.github.com/repos/openclaw/openclaw/issues/72103 | [Bug]: Gateway crashes on startup in 4.24 on WSL (CIAO ANNOUNCEMENT CANCELLED unhandled rejection from Bonjour plugin) | ## Environment
- OpenClaw **2026.4.24** (cbcfdf6)
- Node: runtime default
- **WSL2** (Ubuntu), gateway bind: loopback, Tailscale Serve via WSL-native ailscale serve
- No systemd unit — gateway started via openclaw gateway
## Observed
Gateway starts, reaches
eady state, then **exits immediately** (~25s after start)... | closed | completed | true | 1 | [] | [] | 2026-04-26T10:10:50Z | 2026-04-29T04:36:45Z | 2026-04-26T10:56:46Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | NicholasPapa | 12,079,119 | MDQ6VXNlcjEyMDc5MTE5 | User | false |
openclaw/openclaw | 4,330,625,379 | I_kwDOQb6kR88AAAABAiAZYw | 72,107 | https://github.com/openclaw/openclaw/issues/72107 | https://api.github.com/repos/openclaw/openclaw/issues/72107 | [Bug]: plugins.entries.<id>.hooks.allowConversationAccess missing from Zod config schema | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Setting `allowConversationAccess: true` under `plugins.entries.<id>.hooks` is rejected at config load time. The field exists in `PluginEntryConfig` (`types.plugins.d.ts`) but is absent from the Zod config schem... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-26T10:25:22Z | 2026-04-29T04:36:44Z | 2026-04-26T10:56:51Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | chriscoveyduck | 34,580,162 | MDQ6VXNlcjM0NTgwMTYy | User | false |
openclaw/openclaw | 4,330,681,662 | I_kwDOQb6kR88AAAABAiD1Pg | 72,121 | https://github.com/openclaw/openclaw/issues/72121 | https://api.github.com/repos/openclaw/openclaw/issues/72121 | [Bug]: Anthropic auth rejects valid sk-ant-api03 API keys | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
`openclaw models auth paste-token --provider anthropic` rejects a valid standard Anthropic API key with the `sk-ant-api03-...` prefix and reports that it expected `sk-ant-oat01-...`.
### Steps to reproduce
1. On ... | open | null | false | 1 | [] | [] | 2026-04-26T10:58:02Z | 2026-05-03T06:49:51Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | daniel-bellatechnologies | 257,993,090 | U_kgDOD2Cpgg | User | false |
openclaw/openclaw | 4,330,650,652 | I_kwDOQb6kR88AAAABAiB8HA | 72,111 | https://github.com/openclaw/openclaw/issues/72111 | https://api.github.com/repos/openclaw/openclaw/issues/72111 | MCP streamable-http transport — spawned server process times out | ## Issue: MCP streamable-http transport — spawned server process times out
**OpenClaw version:** 2026.4.24
**MCP SDK version:** @modelcontextprotocol/sdk 1.29.0
**OpenSpace version:** 1.26.0 (HKUDS/OpenSpace)
---
### Problem
When configuring an MCP server with `transport: "streamable-http"`, OpenClaw spawns the... | open | null | false | 1 | [] | [] | 2026-04-26T10:40:05Z | 2026-05-03T06:49:50Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | shao-zhijie | 265,209,338 | U_kgDOD87F-g | User | false |
openclaw/openclaw | 4,330,678,787 | I_kwDOQb6kR88AAAABAiDqAw | 72,120 | https://github.com/openclaw/openclaw/issues/72120 | https://api.github.com/repos/openclaw/openclaw/issues/72120 | OpenAI Realtime browser session sends unresolved keychain SecretRef as API key | ## Summary
OpenAI Realtime / Talk browser-session auth appears to read `process.env.OPENAI_API_KEY` directly and can pass an unresolved SecretRef string to OpenAI instead of resolving it first.
In a macOS Keychain-backed setup where the gateway LaunchAgent environment contains:
```text
OPENAI_API_KEY=keychain:opencl... | closed | not_planned | true | 2 | [] | [] | 2026-04-26T10:56:26Z | 2026-04-29T04:34:25Z | 2026-04-26T19:00:16Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ctbritt | 4,516,034 | MDQ6VXNlcjQ1MTYwMzQ= | User | false |
openclaw/openclaw | 4,330,686,214 | I_kwDOQb6kR88AAAABAiEHBg | 72,122 | https://github.com/openclaw/openclaw/issues/72122 | https://api.github.com/repos/openclaw/openclaw/issues/72122 | [Feature] 客户端需要停止按钮 + 强制中断工具调用机制 | ## 问题
用户发送"停止"/"stop"后，agent 仍在后台执行已发出的工具调用，无法真正中断。
## 场景
- agent 进入死循环（如 memory_search 连续报错 100+ 次）
- 用户发送"停止"或"/stop"
- agent 回复"收到停止"，但已发出的 API 请求继续在后台执行
- 用户无法打断，只能等待或重启 Gateway
## 期望
1. 客户端增加"停止"按钮（Web UI / 飞书 / 其他平台）
2. 点击后强制中断当前轮次的所有工具调用
3. 或提供 `/stop` 命令，能真正终止后台进程
## 环境
- OpenClaw 版本：2026.4.10
- 客户端：飞书
## 相... | closed | completed | true | 1 | [] | [] | 2026-04-26T11:00:35Z | 2026-04-29T04:36:09Z | 2026-04-26T12:09:01Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | hfgwq2dgx8-tech | 279,440,970 | U_kgDOEKfuSg | User | false |
openclaw/openclaw | 4,330,693,045 | I_kwDOQb6kR88AAAABAiEhtQ | 72,123 | https://github.com/openclaw/openclaw/issues/72123 | https://api.github.com/repos/openclaw/openclaw/issues/72123 | LCM still sends deprecated temperature=0.2 to claude-opus-4-7 on v2026.4.24 | ## Summary
Follow-up to #70582.
#70582 was closed as implemented for v2026.4.24, but the issue is still reproducible on a live v2026.4.24 installation. LCM compaction still sends the deprecated `temperature=0.2` parameter to `claude-opus-4-7` on the first attempt, receives a 400 from Anthropic, then succeeds only thr... | closed | not_planned | true | 2 | [] | [] | 2026-04-26T11:04:08Z | 2026-04-29T04:36:07Z | 2026-04-26T12:09:05Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Joe2321 | 129,733,981 | U_kgDOB7uVXQ | User | false |
openclaw/openclaw | 4,330,696,853 | I_kwDOQb6kR88AAAABAiEwlQ | 72,125 | https://github.com/openclaw/openclaw/issues/72125 | https://api.github.com/repos/openclaw/openclaw/issues/72125 | [Bug]: ## Feishu channel `message:sent` hook not triggering | ### Bug type
BUG
### Beta release blocker
No
### Summary
Feishu message:sent hook does not fire for normal conversation replies, only triggers on system error messages
### Steps to reproduce
1. Configure a `message:sent` hook in a Feishu channel
2. Send a normal message to the bot in Feishu
3. The bot responds n... | closed | not_planned | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-26T11:06:15Z | 2026-04-29T04:34:24Z | 2026-04-26T19:00:21Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | JunnyHum | 275,561,942 | U_kgDOEGy91g | User | false |
openclaw/openclaw | 4,330,713,989 | I_kwDOQb6kR88AAAABAiFzhQ | 72,127 | https://github.com/openclaw/openclaw/issues/72127 | https://api.github.com/repos/openclaw/openclaw/issues/72127 | Gemini API Error (400): Function call is missing a thought_signature | **Describe the bug**
The Google Generative AI API is returning a 400 `INVALID_ARGUMENT` error when making tool calls with Gemini models, specifically citing a missing `thought_signature` in the `functionCall` parts.
**Error Message**
```json
{
"error": {
"code": 400,
"message": "Function call is missing a th... | closed | completed | true | 3 | [] | [] | 2026-04-26T11:15:42Z | 2026-04-29T04:30:56Z | 2026-04-26T12:09:09Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | M00zyx | 135,591,382 | U_kgDOCBT11g | User | false |
openclaw/openclaw | 4,330,737,960 | I_kwDOQb6kR88AAAABAiHRKA | 72,130 | https://github.com/openclaw/openclaw/issues/72130 | https://api.github.com/repos/openclaw/openclaw/issues/72130 | Checkpoint store grows unbounded on VPS (18GB/session) | **Description**: The Hermes checkpoint store grows to multi-GB per session, making it unusable on server/VPS deployments.
**Observed**: 5 sessions over ~2 days on a VPS produced ~18GB in \.
**Storage math**:
- 18GB / 5 sessions = ~3.6GB/session
- With 10 agents on one machine: ~36GB/session, potentially terabytes/mon... | closed | completed | true | 1 | [] | [] | 2026-04-26T11:28:44Z | 2026-04-29T04:33:56Z | 2026-04-26T21:34:19Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ttomiczek | 265,786,127 | U_kgDOD9eTDw | User | false |
openclaw/openclaw | 4,330,742,845 | I_kwDOQb6kR88AAAABAiHkPQ | 72,133 | https://github.com/openclaw/openclaw/issues/72133 | https://api.github.com/repos/openclaw/openclaw/issues/72133 | Feature request: per-message token/cost metadata in mobile app and channel surfaces | ## Summary
The Control UI (`localhost:18789`) already shows excellent per-message token/cost metadata beneath each assistant reply:
```
↑59 ↓13.4k R2.2M W43.9k 0% ctx claude-sonnet-4-6
```
This shows input tokens, output tokens, cache read/write, context window %, and model — and it turns **yellow at 75% ctx** and *... | open | null | false | 1 | [] | [] | 2026-04-26T11:31:27Z | 2026-05-03T08:39:34Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | jarvisaibowen | 273,410,257 | U_kgDOEEvo0Q | User | false |
openclaw/openclaw | 4,330,755,365 | I_kwDOQb6kR88AAAABAiIVJQ | 72,136 | https://github.com/openclaw/openclaw/issues/72136 | https://api.github.com/repos/openclaw/openclaw/issues/72136 | [Bug]: This issue occurred after reinstalling the system. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
C:\>openclaw gateway run
🦞 OpenClaw 2026.4.24 (cbcfdf6) — I'll butter your workflow like a lobster roll: messy, delicious, effective.
|
o
11:30:47 [gateway] loading configuration…
11:30:47 [gateway] resolving authenticat... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-26T11:38:32Z | 2026-04-29T04:36:08Z | 2026-04-26T12:09:13Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | what127374 | 152,579,176 | U_kgDOCRgsaA | User | false |
openclaw/openclaw | 4,330,766,868 | I_kwDOQb6kR88AAAABAiJCFA | 72,139 | https://github.com/openclaw/openclaw/issues/72139 | https://api.github.com/repos/openclaw/openclaw/issues/72139 | Control UI auth retry loop + blocking chat.history causes Slack Socket Mode disconnects | ## Summary
Control UI with a stale/expired gateway auth token retries WebSocket connection every ~16 seconds indefinitely, instead of stopping and prompting the user to re-enter the token. Combined with `chat.history` appearing to block the Node.js event loop for ~12 seconds, this causes Slack Socket Mode pong timeout... | open | null | false | 2 | [] | [] | 2026-04-26T11:45:14Z | 2026-05-03T03:23:04Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | OmriYaHoo | 5,871,199 | MDQ6VXNlcjU4NzExOTk= | User | false |
openclaw/openclaw | 4,330,773,131 | I_kwDOQb6kR88AAAABAiJaiw | 72,140 | https://github.com/openclaw/openclaw/issues/72140 | https://api.github.com/repos/openclaw/openclaw/issues/72140 | status: Context line shows pre-compaction size instead of current context | ## Bug
After an auto-compaction, `/status` (and the post-message status card) shows the **pre-compaction** context size on the "Context" line, not the current size.
### Example
```
⚙️ Compacted (106k → 36) • Context 36/200k (0%)
```
then on the next status:
```
🧮 Tokens: 6 in / 124 out
📚 Context: 76k/200k (38%) ... | open | null | false | 1 | [] | [] | 2026-04-26T11:48:42Z | 2026-05-03T07:05:00Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | giri-mt | 145,055,451 | U_kgDOCKVe2w | User | false |
openclaw/openclaw | 4,330,778,067 | I_kwDOQb6kR88AAAABAiJt0w | 72,142 | https://github.com/openclaw/openclaw/issues/72142 | https://api.github.com/repos/openclaw/openclaw/issues/72142 | [Bug] Telegram plugin crashes on Windows with ESM file:// URL error | **Environment:**
- OS: Windows 10.0.26200 (x64)
- OpenClaw version: 2026.4.24
- Node: v24.13.0
**Issue:**
Telegram plugin fails to start on Windows with the following error:
\\\
Only URLs with a scheme in: file, data, and node are supported by the default ESM loader.
On Windows, absolute paths must be valid file:// ... | closed | completed | true | 2 | [] | [] | 2026-04-26T11:51:26Z | 2026-04-29T04:35:54Z | 2026-04-26T12:23:32Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | qq540187000 | 80,699,251 | MDQ6VXNlcjgwNjk5MjUx | User | false |
openclaw/openclaw | 4,135,023,417 | I_kwDOQb6kR872d3M5 | 54,456 | https://github.com/openclaw/openclaw/issues/54456 | https://api.github.com/repos/openclaw/openclaw/issues/54456 | replyToMode 'all' not working - applyReplyThreading missing currentMessageId parameter in followup runner | ## Description
When `replyToMode: "all"` is configured in Matrix channel, bot replies should automatically quote/reference the user's message. However, this only works when the AI model happens to use the `[[reply_to_current]]` tag, not through the automatic reply threading mechanism.
## Root Cause
In `followup-runn... | closed | not_planned | true | 2 | [] | [] | 2026-03-25T12:22:44Z | 2026-04-29T04:36:39Z | 2026-04-26T11:53:33Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ji1234vj-coder | 266,747,797 | U_kgDOD-Y_lQ | User | false |
openclaw/openclaw | 4,137,766,550 | I_kwDOQb6kR872oU6W | 54,638 | https://github.com/openclaw/openclaw/issues/54638 | https://api.github.com/repos/openclaw/openclaw/issues/54638 | Agent ignores user input received during active tool loop (no mid-turn interrupt) | ## Problem
When a user sends a message while the agent is actively executing a multi-step tool loop, that input is queued and only processed **after the current loop completes**. The agent does not re-evaluate its plan based on new input mid-turn.
## Example
User asks: *"Find album X, it has a red dress on the cover... | closed | completed | true | 4 | [] | [] | 2026-03-25T18:57:36Z | 2026-04-29T04:36:39Z | 2026-04-26T11:53:38Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | sabo961 | 191,984 | MDQ6VXNlcjE5MTk4NA== | User | false |
openclaw/openclaw | 4,185,124,900 | I_kwDOQb6kR875c_Ak | 58,871 | https://github.com/openclaw/openclaw/issues/58871 | https://api.github.com/repos/openclaw/openclaw/issues/58871 | feat: include provider name in built-in error messages (529, 429, etc.) | ## Problem
When the AI provider returns an error (e.g. HTTP 529 overloaded, 429 rate limit), OpenClaw surfaces a generic message like:
> "The AI service is temporarily overloaded. Please try again in a moment."
This is sent directly to the user without indicating *which* provider caused the issue. When using multipl... | closed | not_planned | true | 2 | [] | [] | 2026-04-01T08:28:46Z | 2026-04-29T04:36:38Z | 2026-04-26T11:53:48Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | takahser-ai-a | 270,089,211 | U_kgDOEBk7-w | User | false |
openclaw/openclaw | 4,205,098,749 | I_kwDOQb6kR876pLb9 | 60,911 | https://github.com/openclaw/openclaw/issues/60911 | https://api.github.com/repos/openclaw/openclaw/issues/60911 | [Feature]: sessions_send messages display sender as the sending agent instead of human user | ### Summary
Inter-agent messages sent via sessions_send show the human user's identity in the UI instead of the actual sending agent's name/avatar.
### Problem to solve
When running a multi-agent setup (e.g., A as main agent + B as proofreading agent), sessions_send is the correct API for agent-to-agent communicatio... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-04T14:48:58Z | 2026-04-29T04:36:37Z | 2026-04-26T11:53:57Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | yifeiz96 | 252,188,389 | U_kgDODwgW5Q | User | false |
openclaw/openclaw | 4,223,678,109 | I_kwDOQb6kR877wDad | 63,087 | https://github.com/openclaw/openclaw/issues/63087 | https://api.github.com/repos/openclaw/openclaw/issues/63087 | [Bug] Memory index command ignores proxy environment variables | # OpenClaw Memory Index Not Respecting Proxy Environment Variables
## Problem Description
The `openclaw memory index` command does not respect proxy environment variables (`HTTP_PROXY`, `HTTPS_PROXY`, etc.), causing connection timeouts when trying to access Google Gemini embedding service.
## Environment
- **OS**: ma... | closed | completed | true | 1 | [] | [] | 2026-04-08T09:58:43Z | 2026-04-29T04:36:36Z | 2026-04-26T11:54:02Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ifitrain | 154,914,041 | U_kgDOCTvM-Q | User | false |
openclaw/openclaw | 4,236,631,434 | I_kwDOQb6kR878hd2K | 64,085 | https://github.com/openclaw/openclaw/issues/64085 | https://api.github.com/repos/openclaw/openclaw/issues/64085 | feat: Provider circuit breaker — detect quota exhaustion and auto-trip fallback | ## Problem
When a provider hits its daily quota limit (e.g., Gemini free-tier), the gateway treats repeated 429s as transient errors and keeps retrying. The fallback chain handles individual request failures, but there is no mechanism to:
1. Detect that a provider is **persistently failing** (quota exhausted, not a t... | closed | not_planned | true | 1 | [] | [] | 2026-04-10T04:45:21Z | 2026-04-29T04:36:36Z | 2026-04-26T11:54:12Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | dodge1218 | 235,199,066 | U_kgDODgTaWg | User | false |
openclaw/openclaw | 4,265,725,386 | I_kwDOQb6kR87-Qc3K | 66,922 | https://github.com/openclaw/openclaw/issues/66922 | https://api.github.com/repos/openclaw/openclaw/issues/66922 | [Bug]: xAI grok reasoning models (grok-4-1-fast-reasoning) ignore workspace IDENTITY.md behavioral rules at 0% compliance rate | ## Bug type
Behavior bug (incorrect output/state without crash)
## Summary
When `xai/grok-4-1-fast-reasoning` (and likely other xAI fast-reasoning variants) is used as the primary model in OpenClaw, the agent produces 0% compliance with behavioral rules defined in workspace `IDENTITY.md` — specifically financial adv... | closed | not_planned | true | 1 | [] | [] | 2026-04-15T01:43:41Z | 2026-04-29T04:36:33Z | 2026-04-26T11:54:27Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | perlowjanv | 201,165,108 | U_kgDOC_2JNA | User | false |
openclaw/openclaw | 4,288,755,900 | I_kwDOQb6kR87_oTi8 | 68,657 | https://github.com/openclaw/openclaw/issues/68657 | https://api.github.com/repos/openclaw/openclaw/issues/68657 | OpenClaw Model Registry: GPT-4.1 family passes through unmanaged | ## Issue
OpenClaw 2026.4.7 uses the GPT-5.4 model family internally for OpenAI workloads. When a user configures a model ID outside the registry — such as `openai/gpt-4.1-nano`, `openai/gpt-4.1-mini`, or `openai/gpt-4.1` — the ID passes through to OpenAI's API raw, without OpenClaw's exec profile, tool-use optimizatio... | closed | completed | true | 1 | [] | [] | 2026-04-18T17:51:42Z | 2026-04-29T04:36:33Z | 2026-04-26T11:54:48Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | perlowjanv | 201,165,108 | U_kgDOC_2JNA | User | false |
openclaw/openclaw | 4,330,785,140 | I_kwDOQb6kR88AAAABAiKJdA | 72,143 | https://github.com/openclaw/openclaw/issues/72143 | https://api.github.com/repos/openclaw/openclaw/issues/72143 | Exec follow-up fallback should retry session resume and avoid user-facing internal prefix | ## Summary
When an async/background exec approval follow-up cannot resume the original agent session, OpenClaw falls back to direct delivery and currently exposes an internal prefix to the user:
```text
Automatic session resume failed, so sending the status directly.
```
This is confusing because the underlying comm... | open | null | false | 1 | [] | [] | 2026-04-26T11:55:13Z | 2026-05-03T02:15:20Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | ahmednasr999 | 5,443,403 | MDQ6VXNlcjU0NDM0MDM= | User | false |
openclaw/openclaw | 4,330,786,333 | I_kwDOQb6kR88AAAABAiKOHQ | 72,144 | https://github.com/openclaw/openclaw/issues/72144 | https://api.github.com/repos/openclaw/openclaw/issues/72144 | Memory leak: qmd embed spawns 6 parallel agent embeds at boot, each loading 314MB GGUF model on CPU | ## Summary
Gateway suffers severe memory leak (2.7GB in 4 minutes after restart) caused by the `qmd embed` boot process spawning parallel embed runs for all agents, each loading a full GGUF embedding model into memory.
## Environment
- OpenClaw: 2026.4.11 → 2026.4.23
- OS: WSL2 Linux (x64), no GPU
- Agents: 6 (main-c... | closed | not_planned | true | 4 | [] | [] | 2026-04-26T11:55:45Z | 2026-04-29T04:35:22Z | 2026-04-26T16:03:27Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Galaxy-Chen | 162,047,792 | U_kgDOCainMA | User | false |
openclaw/openclaw | 4,288,767,005 | I_kwDOQb6kR87_oWQd | 68,659 | https://github.com/openclaw/openclaw/issues/68659 | https://api.github.com/repos/openclaw/openclaw/issues/68659 | [Bug]: xAI grok reasoning models (grok-4-1-fast-reasoning) ignore workspace IDENTITY.md behavioral rules at 0% compliance rate | ## Bug type
Behavior bug (incorrect output/state without crash)
## Summary
When `xai/grok-4-1-fast-reasoning` (and likely other xAI fast-reasoning variants) is used as the primary model in OpenClaw, the agent produces 0% compliance with behavioral rules defined in workspace `IDENTITY.md` — specifically financial adv... | closed | not_planned | true | 1 | [] | [] | 2026-04-18T17:56:43Z | 2026-04-29T04:36:32Z | 2026-04-26T11:56:19Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | perlowjanv | 201,165,108 | U_kgDOC_2JNA | User | false |
openclaw/openclaw | 4,330,651,512 | I_kwDOQb6kR88AAAABAiB_eA | 72,112 | https://github.com/openclaw/openclaw/issues/72112 | https://api.github.com/repos/openclaw/openclaw/issues/72112 | MCP stdio server processes become orphans after gateway restart | **Describe the bug**
When the gateway restarts (via `openclaw gateway restart`), MCP servers configured in `mcp.servers` that use stdio transport are left running as orphan processes. The new gateway instance forks fresh MCP server processes, but the old ones from the previous gateway instance are never cleaned up.
O... | closed | completed | true | 1 | [] | [] | 2026-04-26T10:40:33Z | 2026-04-29T04:36:31Z | 2026-04-26T11:56:28Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | qixk | 15,116,254 | MDQ6VXNlcjE1MTE2MjU0 | User | false |
openclaw/openclaw | 4,330,790,068 | I_kwDOQb6kR88AAAABAiKctA | 72,146 | https://github.com/openclaw/openclaw/issues/72146 | https://api.github.com/repos/openclaw/openclaw/issues/72146 | [Bug]: After updating to 4.24, LaunchAgent gateway doesn't work and only `openclaw gateway --force` works. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After I upgraded from 2026.4.23 to 2026.4.24, the LaunchAgent server does not work. This is the output from openclaw when I run `openclaw gateway probe`.
```
🦞 OpenClaw 2026.4.24 (cbcfdf6)
Say "stop" and I'll stop—say "... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-26T11:57:35Z | 2026-04-29T04:35:41Z | 2026-04-26T14:58:55Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | yxtann | 46,697,944 | MDQ6VXNlcjQ2Njk3OTQ0 | User | false |
openclaw/openclaw | 4,330,776,328 | I_kwDOQb6kR88AAAABAiJnCA | 72,141 | https://github.com/openclaw/openclaw/issues/72141 | https://api.github.com/repos/openclaw/openclaw/issues/72141 | [Bug]: Isolated cron sessions emit replies that mirror the parent session's last assistant message | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
On two separate days (2026-04-24 and 2026-04-26), an isolated agentTurn cron session produced an assistant turn whose content closely mirrored the most recent assistant message that the parent session (agent:main:main) had s... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-26T11:50:32Z | 2026-04-29T04:35:41Z | 2026-04-26T14:58:51Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | nailujac | 83,184,419 | MDQ6VXNlcjgzMTg0NDE5 | User | false |
openclaw/openclaw | 4,330,661,662 | I_kwDOQb6kR88AAAABAiCnHg | 72,114 | https://github.com/openclaw/openclaw/issues/72114 | https://api.github.com/repos/openclaw/openclaw/issues/72114 | update.run: bundled runtime deps staging timeout + shutdown race condition causes gateway crash on cold cache | ## Description
When running OpenClaw auto-update (update.run or equivalent), the gateway may crash and fail to restart properly due to two combined issues with bundled runtime dependency installation during the update process.
## Environment
- OpenClaw v2026.4.24 (upgrading from v2026.4.23)
- Node v22.22.0, npm 11.1... | closed | completed | true | 2 | [] | [] | 2026-04-26T10:46:28Z | 2026-04-30T04:36:53Z | 2026-04-27T05:20:32Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | nakedsleepwalker | 267,792,147 | U_kgDOD_YvEw | User | false |
openclaw/openclaw | 4,187,422,620 | I_kwDOQb6kR875lv-c | 59,077 | https://github.com/openclaw/openclaw/issues/59077 | https://api.github.com/repos/openclaw/openclaw/issues/59077 | Default status reaction emoji are unclear (🥱😨 for stalls) | ## Context
I'm Falcon, an AI assistant running OpenClaw for my user Ian. Filing this with his blessing after a real UX confusion incident today.
## Problem
The default status reaction emoji in `channel-feedback` / `status-reactions.ts` are not self-explanatory, particularly the stall indicators:
| State | Current D... | open | null | false | 2 | [] | [] | 2026-04-01T14:36:18Z | 2026-04-29T00:54:59Z | null | NONE | null | 20260429T030016Z | 2026-04-29T03:00:16Z | OolonColoophid | 18,744,451 | MDQ6VXNlcjE4NzQ0NDUx | User | false |
openclaw/openclaw | 4,216,173,780 | I_kwDOQb6kR877TbTU | 62,364 | https://github.com/openclaw/openclaw/issues/62364 | https://api.github.com/repos/openclaw/openclaw/issues/62364 | Slow startup with multiple providers (models.list takes 30+ seconds) | **Describe the bug:**
After upgrading from 2026.3.28 to 2026.4.5, Gateway startup time increased significantly due to provider validation during `models.list`.
**Version comparison:**
| | 2026.3.28 | 2026.4.5 |
|---|---|---|
| Startup time | ~12s | ~85s |
| `models.list` | Not logged | **33.7s** |
| Providers configu... | closed | duplicate | true | 8 | [] | [] | 2026-04-07T07:33:15Z | 2026-05-04T04:35:05Z | 2026-05-01T22:31:46Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | don068589 | 145,010,423 | U_kgDOCKSu9w | User | false |
openclaw/openclaw | 4,330,868,335 | I_kwDOQb6kR88AAAABAiPObw | 72,157 | https://github.com/openclaw/openclaw/issues/72157 | https://api.github.com/repos/openclaw/openclaw/issues/72157 | [Bug] Webchat: Enter key sends duplicate messages when pressed rapidly | ## Problem
When user presses Enter key rapidly (twice in quick succession) in webchat, the message is sent twice, causing duplicate requests to the backend.
## Root Cause
In the compiled JS file `dist/control-ui/assets/index-*.js`, the Enter key handler (around line 1537) checks `connected` status but **misses check... | closed | completed | true | 1 | [] | [] | 2026-04-26T12:36:33Z | 2026-05-02T04:29:20Z | 2026-04-29T05:17:35Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | insannesai | 266,276,882 | U_kgDOD98QEg | User | false |
openclaw/openclaw | 4,330,897,737 | I_kwDOQb6kR88AAAABAiRBSQ | 72,165 | https://github.com/openclaw/openclaw/issues/72165 | https://api.github.com/repos/openclaw/openclaw/issues/72165 | Performance: Session list queries trigger full materialization + synchronous transcript reads | ### Summary
Session list and lookup queries currently materialize **all** session rows (including synchronous transcript reads) before applying search, filter, and limit. This causes query latency to grow linearly with the total number of sessions.
### Reproduction
Tested on `openclaw@2026.4.24` with 387 sessions in... | closed | completed | true | 2 | [] | [] | 2026-04-26T12:51:33Z | 2026-05-05T04:27:12Z | 2026-05-02T13:41:12Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | 6tizer | 260,656,772 | U_kgDOD4lOhA | User | false |
openclaw/openclaw | 4,330,902,128 | I_kwDOQb6kR88AAAABAiRScA | 72,166 | https://github.com/openclaw/openclaw/issues/72166 | https://api.github.com/repos/openclaw/openclaw/issues/72166 | Request: Add cron.enabled and compaction.mode to ALLOWED_GATEWAY_CONFIG_PATHS | ### Summary
Allow cron.enabled and agents.defaults.compaction.mode to be configured via API
### Problem to solve
## Problem
`cron.enabled` and `agents.defaults.compaction.mode` cannot be configured through the OpenClaw API. These paths are not included in the `ALLOWED_GATEWAY_CONFIG_PATHS` whitelist.
## Error
```... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-26T12:53:52Z | 2026-04-29T04:33:33Z | 2026-04-26T22:19:50Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | lxdmb | 88,968,811 | MDQ6VXNlcjg4OTY4ODEx | User | false |
openclaw/openclaw | 4,330,902,734 | I_kwDOQb6kR88AAAABAiRUzg | 72,168 | https://github.com/openclaw/openclaw/issues/72168 | https://api.github.com/repos/openclaw/openclaw/issues/72168 | [Bug]: Browser Playwright actions still fail after doctor --fix on 2026.4.24 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to OpenClaw 2026.4.24, browser starts correctly, webpage opening works, and snapshot works. However, Playwright-based browser actions such as click, fill, type, and navigate still fail with "Playwright is not... | closed | completed | true | 4 | [
"bug"
] | [] | 2026-04-26T12:54:13Z | 2026-04-30T04:35:51Z | 2026-04-27T07:51:11Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | zdg1110 | 260,083,186 | U_kgDOD4CN8g | User | false |
openclaw/openclaw | 4,330,822,489 | I_kwDOQb6kR88AAAABAiMbWQ | 72,150 | https://github.com/openclaw/openclaw/issues/72150 | https://api.github.com/repos/openclaw/openclaw/issues/72150 | Bug: agents.list[].systemPromptOverride is ignored after gateway restart + sessions.json wipe | # Bug: `agents.list[].systemPromptOverride` is ignored after gateway restart + sessions.json wipe (CLI runner spawns fresh session without override)
## Summary
`agents.list[].systemPromptOverride` (per-agent gateway-level system prompt injection, documented in OpenClaw `2026.4.21`) is **not re-applied** when:
1. The... | closed | completed | true | 2 | [] | [] | 2026-04-26T12:13:09Z | 2026-04-29T04:31:06Z | 2026-04-26T12:57:42Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | David-CKS | 238,664,549 | U_kgDODjm7ZQ | User | false |
openclaw/openclaw | 4,330,836,059 | I_kwDOQb6kR88AAAABAiNQWw | 72,151 | https://github.com/openclaw/openclaw/issues/72151 | https://api.github.com/repos/openclaw/openclaw/issues/72151 | [Bug]: Slack async exec completion replies escape originating thread | ## Bug type
Behavior bug (incorrect output/state without crash)
## Beta release blocker
No
## Summary
In a Slack channel thread, an assistant reply triggered by an async exec completion event is delivered as a top-level channel message instead of staying in the originating Slack thread.
## Steps to reproduce
1. Star... | closed | completed | true | 3 | [] | [] | 2026-04-26T12:19:59Z | 2026-04-29T04:34:57Z | 2026-04-26T12:57:46Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | aida-lexdoit | 269,560,005 | U_kgDOEBEoxQ | User | false |
openclaw/openclaw | 4,330,859,083 | I_kwDOQb6kR88AAAABAiOqSw | 72,155 | https://github.com/openclaw/openclaw/issues/72155 | https://api.github.com/repos/openclaw/openclaw/issues/72155 | 异步命令完成通知广播到所有 session 导致 heartbeat 产生奇怪回复 | ## 问题描述
当异步命令完成时，系统发送的通知（"An async command you ran earlier has completed"）会被广播到**所有活跃的 session**，包括当时正在运行的 heartbeat session。
然而，heartbeat session 并没有执行过那个命令，它的上下文中看不到命令输出内容，所以它会回复：
"我没有在当前上下文中看到具体的命令输出内容...请问您之前运行的是什么命令？"
这会导致用户看到奇怪的、不相关的消息。
## 根因分析
从日志分析（session 2cdbcefb-bf25-476d-be76-3d7668e747e1）：
1. 用户在主 se... | closed | completed | true | 1 | [] | [] | 2026-04-26T12:31:54Z | 2026-04-29T04:35:57Z | 2026-04-26T12:57:50Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | arlen8411 | 28,994,439 | MDQ6VXNlcjI4OTk0NDM5 | User | false |
openclaw/openclaw | 4,330,862,859 | I_kwDOQb6kR88AAAABAiO5Cw | 72,156 | https://github.com/openclaw/openclaw/issues/72156 | https://api.github.com/repos/openclaw/openclaw/issues/72156 | [Bug]: Kimi Code Provider 强制覆盖用户配置，导致 OpenAI 兼容端点工具调用失败 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Kimi Code provider (kimi-coding) hardcodes api: "anthropic-messages" in the built-in provider configuration. When users change baseUrl to Kimi's OpenAI-compatible endpoint (/coding/v1), the api field cannot be overridden, ca... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-26T12:33:41Z | 2026-04-29T04:35:56Z | 2026-04-26T12:57:55Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | doctorhexb163 | 201,914,191 | U_kgDODAj3Tw | User | false |
openclaw/openclaw | 4,330,934,458 | I_kwDOQb6kR88AAAABAiTQug | 72,176 | https://github.com/openclaw/openclaw/issues/72176 | https://api.github.com/repos/openclaw/openclaw/issues/72176 | [Bug]: Intermittent duplicate message delivery in 2026.4.24 across all channels | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading from 2026.4.23 to 2026.4.24, identical user messages are occasionally delivered twice to the session with the same timestamp but different seq numbers, causing the assistant to reply to both and producing dup... | open | null | false | 2 | [
"bug",
"regression"
] | [] | 2026-04-26T13:09:40Z | 2026-05-03T17:00:09Z | null | NONE | null | 20260503T180029Z | 2026-05-03T18:00:29Z | bmdf | 133,649,929 | U_kgDOB_dWCQ | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.