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 | 4,190,876,284 | I_kwDOQb6kR875y7J8 | 59,329 | https://github.com/openclaw/openclaw/issues/59329 | https://api.github.com/repos/openclaw/openclaw/issues/59329 | [Bug]: Approval List Always Allow doesn't work | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to 3.31 version. Bird skills summaring X tweets keep asking for approval, I hit "always allow" again and again but just not working.
### Steps to reproduce
1. set up a cron job to run scripts, using bird s... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-02T01:40:46Z | 2026-04-28T05:14:00Z | 2026-04-25T13:03:27Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | mhz413 | 266,643,414 | U_kgDOD-Sn1g | User | false |
openclaw/openclaw | 4,190,881,853 | I_kwDOQb6kR875y8g9 | 59,331 | https://github.com/openclaw/openclaw/issues/59331 | https://api.github.com/repos/openclaw/openclaw/issues/59331 | Telegram: old messages reprocessed in loop after gateway restart (update_id offset not persisted) | ### Description
When the OpenClaw gateway restarts (manually or after a crash), Telegram messages that were already processed get re-fetched and reprocessed. This causes duplicate responses — in one instance, the same message was processed 5 times in rapid succession.
### Root Cause (suspected)
The grammY long-polli... | closed | completed | true | 1 | [] | [] | 2026-04-02T01:43:06Z | 2026-04-28T05:14:00Z | 2026-04-25T13:03:34Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | empire-build-co | 270,150,703 | U_kgDOEBosLw | User | false |
openclaw/openclaw | 4,190,891,491 | I_kwDOQb6kR875y-3j | 59,333 | https://github.com/openclaw/openclaw/issues/59333 | https://api.github.com/repos/openclaw/openclaw/issues/59333 | [Bug]: Browser tool failures trigger gateway SIGTERM restart instead of graceful error handling | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
Yes
### Summary
Browser tool errors (tab not found, CDP disconnect, Chrome process crash) cause the entire OpenClaw gateway to
receive SIGTERM and restart, instead of the error being caught and returned to the agent session gracefully.... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-02T01:46:47Z | 2026-04-28T05:13:59Z | 2026-04-25T13:03:41Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | cliftonwknox | 133,707,677 | U_kgDOB_g3nQ | User | false |
openclaw/openclaw | 4,190,967,622 | I_kwDOQb6kR875zRdG | 59,346 | https://github.com/openclaw/openclaw/issues/59346 | https://api.github.com/repos/openclaw/openclaw/issues/59346 | Telemetry endpoint us.i.posthog.com returns 404 | ## Bug Description
OpenClaw is attempting to send telemetry data to `us.i.posthog.com`, but the endpoint returns **404 Page Not Found**.
## Error Log
```
Telemetry event capture failed: [TypeError: fetch failed] {
[cause]: Error: Client network socket disconnected before secure TLS connection was established
... | closed | not_planned | true | 1 | [] | [] | 2026-04-02T02:16:01Z | 2026-04-28T05:13:58Z | 2026-04-25T13:03:55Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ifitrain | 154,914,041 | U_kgDOCTvM-Q | User | false |
openclaw/openclaw | 4,328,056,829 | I_kwDOQb6kR88AAAABAfjn_Q | 71,572 | https://github.com/openclaw/openclaw/issues/71572 | https://api.github.com/repos/openclaw/openclaw/issues/71572 | Bedrock replay can enter empty assistant content death loop | ## Summary
OpenClaw can persist assistant messages with empty `content: []` after stream/provider errors. With AWS Bedrock Converse, any later replay containing that assistant turn fails with:
```text
Validation error: The content field in the Message object at messages.N is empty. Add a ContentBlock object to the co... | closed | completed | true | 1 | [] | [] | 2026-04-25T12:15:07Z | 2026-04-28T04:53:28Z | 2026-04-25T18:04:42Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | sesame437 | 35,056,873 | MDQ6VXNlcjM1MDU2ODcz | User | false |
openclaw/openclaw | 4,190,976,530 | I_kwDOQb6kR875zToS | 59,348 | https://github.com/openclaw/openclaw/issues/59348 | https://api.github.com/repos/openclaw/openclaw/issues/59348 | [Feature]: Add option to disable model pricing bootstrap to avoid startup timeout | ### Summary
Add configuration to disable external model pricing data fetching during Gateway startup.
### Problem to solve
Gateway startup sometimes fails or appears to hang with `TimeoutError` when trying to fetch external model pricing information:
[model-pricing] pricing bootstrap failed: TimeoutError: The operat... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-02T02:19:32Z | 2026-04-28T05:13:57Z | 2026-04-25T13:04:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fangxicc | 55,014,793 | MDQ6VXNlcjU1MDE0Nzkz | User | false |
openclaw/openclaw | 4,191,012,054 | I_kwDOQb6kR875zcTW | 59,354 | https://github.com/openclaw/openclaw/issues/59354 | https://api.github.com/repos/openclaw/openclaw/issues/59354 | [Bug]: iMessage file-only outbound messages bypass echo detection and trigger auto-response | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When sending a file-only message via `imsg send --to "+1..." --file "/path/to/file.pdf" --service imessage`, the iMessage poller picks up the outbound message as incoming. The echo cache in the self-chat/DM han... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T02:32:10Z | 2026-04-28T05:13:57Z | 2026-04-25T13:04:10Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | leobetancor | 77,313,791 | MDQ6VXNlcjc3MzEzNzkx | User | false |
openclaw/openclaw | 4,191,020,444 | I_kwDOQb6kR875zeWc | 59,355 | https://github.com/openclaw/openclaw/issues/59355 | https://api.github.com/repos/openclaw/openclaw/issues/59355 | Feature Request: Cross-channel session context sharing for the same agent | ## Feature Request: Cross-channel session context sharing for the same agent
### Problem
When the same agent (e.g., agent:main) is accessed from multiple channels (webchat, WeChat, Telegram, etc.), each channel maintains its own independent session. This means:
- Webchat session: `agent:main:main` - has full convers... | closed | completed | true | 1 | [] | [] | 2026-04-02T02:34:36Z | 2026-04-28T05:13:56Z | 2026-04-25T13:04:17Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tcyxk | 129,282,805 | U_kgDOB7Sy9Q | User | false |
openclaw/openclaw | 4,191,053,778 | I_kwDOQb6kR875zmfS | 59,358 | https://github.com/openclaw/openclaw/issues/59358 | https://api.github.com/repos/openclaw/openclaw/issues/59358 | QQBot channel doesn't show tool execution details, making it unusable for debugging | ## Problem
When interacting with OpenClaw via QQBot channel, tool call information is not visible to users. Only the final result is returned after all tools finish executing.
This makes it difficult to:
1. Track the progress of long-running operations
2. Debug when something goes wrong
3. Understand what steps the a... | closed | completed | true | 1 | [] | [] | 2026-04-02T02:42:48Z | 2026-04-28T05:13:55Z | 2026-04-25T13:04:32Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | shaohq | 40,241,246 | MDQ6VXNlcjQwMjQxMjQ2 | User | false |
openclaw/openclaw | 4,191,109,123 | I_kwDOQb6kR875z0AD | 59,363 | https://github.com/openclaw/openclaw/issues/59363 | https://api.github.com/repos/openclaw/openclaw/issues/59363 | [Bug]: imessage echo loop | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
iMessage reflects assistant outbound text back as inbound user messages in direct chat, causing loops.
### Steps to reproduce
1. send message to agent in imessage.
2. agent sends message back
3. agent receives its own mess... | closed | completed | true | 6 | [
"bug",
"regression"
] | [] | 2026-04-02T02:58:28Z | 2026-04-28T05:13:54Z | 2026-04-25T13:04:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fagandesign | 19,477,233 | MDQ6VXNlcjE5NDc3MjMz | User | false |
openclaw/openclaw | 4,191,167,247 | I_kwDOQb6kR8750CMP | 59,375 | https://github.com/openclaw/openclaw/issues/59375 | https://api.github.com/repos/openclaw/openclaw/issues/59375 | [Feature]: Cross-session task notification — allow agents to notify other sessions on completion | ## Problem
In a multi-agent architecture where different agents run in separate Discord channels (each with their own OpenClaw session), there is no built-in mechanism for one agent to notify another agent's session when a task is complete.
### Real-world scenario
- luna-agent dispatches a task to a team channel whe... | closed | completed | true | 1 | [] | [] | 2026-04-02T03:17:55Z | 2026-04-28T05:13:54Z | 2026-04-25T13:04:46Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kagura-agent | 268,167,063 | U_kgDOD_vnlw | User | false |
openclaw/openclaw | 4,191,191,619 | I_kwDOQb6kR8750IJD | 59,377 | https://github.com/openclaw/openclaw/issues/59377 | https://api.github.com/repos/openclaw/openclaw/issues/59377 | [Bug] 审批弹窗内容过长时超出视口，无法点击批准按钮 | ## 问题描述
当执行包含长代码块的命令时，审批弹窗会超出浏览器视口，导致无法点击批准按钮。
## 复现步骤
1. 执行包含长代码块的命令
2. 系统弹出审批对话框
3. 弹窗超出屏幕，按钮无法点击
## 建议修复
```css
.approval-modal {
max-height: 80vh;
overflow-y: auto;
}
```
## 临时解决方案
- 按 Tab 键聚焦到批准按钮，再按 Enter
- 缩小浏览器页面（Ctrl + -）
- 直接输入 `/approve <id> allow-once` 命令
## 环境信息
- OpenClaw 版本: 2026.3.24
- 界面: Open... | closed | completed | true | 1 | [] | [] | 2026-04-02T03:23:34Z | 2026-04-28T05:13:53Z | 2026-04-25T13:04:54Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | StrangeXin | 6,157,378 | MDQ6VXNlcjYxNTczNzg= | User | false |
openclaw/openclaw | 4,191,322,080 | I_kwDOQb6kR8750n_g | 59,391 | https://github.com/openclaw/openclaw/issues/59391 | https://api.github.com/repos/openclaw/openclaw/issues/59391 | doctor reports gateway service entrypoint mismatch on macOS LaunchAgent in v2026.4.1 (entry.js vs index.js) | ## Summary
After upgrading to OpenClaw v2026.4.1 on macOS, `openclaw doctor --non-interactive` reports a gateway service entrypoint mismatch (`entry.js` vs `index.js`) while the macOS LaunchAgent still uses `dist/entry.js` and the gateway works normally.
This looks like a mismatch between what `doctor` expects and wh... | closed | completed | true | 1 | [] | [] | 2026-04-02T03:47:32Z | 2026-04-28T05:13:52Z | 2026-04-25T13:05:08Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Chan-Kris | 103,567,511 | U_kgDOBixQlw | User | false |
openclaw/openclaw | 4,191,340,622 | I_kwDOQb6kR8750shO | 59,393 | https://github.com/openclaw/openclaw/issues/59393 | https://api.github.com/repos/openclaw/openclaw/issues/59393 | [Bug]: Subagent tool calls never execute - accumulatedRuntimeMs always 0 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Summary
Subagent tool calls are generated by the model but never executed. The subagent completes after one response turn with `accumulatedRuntimeMs: 0`.
### Steps to reproduce
## Steps to Reproduce
1. Spawn a subagent ... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-02T03:51:35Z | 2026-04-28T05:13:51Z | 2026-04-25T13:05:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | dmoniz22 | 90,197,049 | MDQ6VXNlcjkwMTk3MDQ5 | User | false |
openclaw/openclaw | 4,191,373,658 | I_kwDOQb6kR87500la | 59,399 | https://github.com/openclaw/openclaw/issues/59399 | https://api.github.com/repos/openclaw/openclaw/issues/59399 | [Feature]: 飞书多账号场景下的工具账号优先级问题 | ### Summary
## 问题摘要
当 OpenClaw 配置了多个不同飞书组织下的机器人时，飞书工具（如 `feishu_doc`）无法正确识别消息来源的机器人账号，导致操作飞书文档时出现权限错误。
---
## 环境信息
- OpenClaw 版本：2026.3.13
- 飞书插件版本：2026.3.13
- 配置账号数：3个（来自不同飞书组织）
---
## 场景描述
### 配置示例
```json
{
"channels": {
"feishu": {
"enabled": true,
"connectionMode": "websocket",
"de... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-02T03:59:10Z | 2026-04-28T05:13:50Z | 2026-04-25T13:05:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lc1573 | 68,944,049 | MDQ6VXNlcjY4OTQ0MDQ5 | User | false |
openclaw/openclaw | 4,191,381,902 | I_kwDOQb6kR87502mO | 59,402 | https://github.com/openclaw/openclaw/issues/59402 | https://api.github.com/repos/openclaw/openclaw/issues/59402 | [Feature]: 跨 Agent 通信机制优化：Session 发现与角色身份标识 | ### Summary
## 问题描述
在使用 OpenClaw 进行多 Agent 协作时，当前缺乏可靠的跨 Agent 直接通信机制。现有方案存在两个核心问题：
1. **Session ID 发现机制缺失** — Agent A 无法可靠获取 Agent B 的 sessionID
2. **角色身份混淆风险** — 长上下文场景中，模型注意力缺失导致无法区分"用户消息"与"Agent 间消息"
---
## 核心目标
实现非子代理的跨 Agent 协作方案，解决以下痛点：
### 1. 保持 Agent 上下文连续性
- 子代理 (`sessions_spawn`) 任务完成后会话即销毁，无法保留历史上下文
... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-02T04:01:17Z | 2026-04-28T05:13:49Z | 2026-04-25T13:05:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | qq475301610 | 13,128,150 | MDQ6VXNlcjEzMTI4MTUw | User | false |
openclaw/openclaw | 4,328,161,643 | I_kwDOQb6kR88AAAABAfqBaw | 71,586 | https://github.com/openclaw/openclaw/issues/71586 | https://api.github.com/repos/openclaw/openclaw/issues/71586 | [Bug]: macOS companion app chat UI hangs when dragging scrollbar thumb up and down | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
The Chat view in the macOS companion app works well most of the time, including scrolling up and down by swiping on the trackpad. But if you click and drag the scrollbar thumb (which, depending on system configuration, may appe... | open | null | false | 1 | [
"bug"
] | [] | 2026-04-25T13:05:41Z | 2026-05-02T14:00:38Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | sentience | 89,772 | MDQ6VXNlcjg5Nzcy | User | false |
openclaw/openclaw | 4,328,161,792 | I_kwDOQb6kR88AAAABAfqCAA | 71,587 | https://github.com/openclaw/openclaw/issues/71587 | https://api.github.com/repos/openclaw/openclaw/issues/71587 | [Bug]: Add missing OpenCode Go models (Kimi K2.6, GLM-5.1, MiMo-V2, Qwen, DeepSeek) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
OpenClaw partially supports OpenCode Go models, but several newer models are missing from the model registry. During onboarding/wizard, only 4 OpenCode Go models are available:
- opencode-go/glm-5
- opencode-go/kimi-k2.5
- o... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-25T13:05:46Z | 2026-04-28T04:58:14Z | 2026-04-25T17:12:11Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | alexandre-leng | 72,550,137 | MDQ6VXNlcjcyNTUwMTM3 | User | false |
openclaw/openclaw | 4,191,592,358 | I_kwDOQb6kR8751p-m | 59,426 | https://github.com/openclaw/openclaw/issues/59426 | https://api.github.com/repos/openclaw/openclaw/issues/59426 | [Feature]: OpenClaw as MCP Server (Expose Execution & Orchestration) | ### Summary
Enable OpenClaw to act as an MCP server, exposing its execution, artifact handling, and environment orchestration capabilities to external LLM systems.
### Problem to solve
Currently, OpenClaw primarily acts as an agent and MCP client, meaning it can consume tools but cannot be easily used as a tool by o... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-04-02T04:47:36Z | 2026-04-28T05:13:48Z | 2026-04-25T13:05:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | rtaghawi | 26,818,151 | MDQ6VXNlcjI2ODE4MTUx | User | false |
openclaw/openclaw | 4,191,614,125 | I_kwDOQb6kR8751vSt | 59,430 | https://github.com/openclaw/openclaw/issues/59430 | https://api.github.com/repos/openclaw/openclaw/issues/59430 | [Bug]: Image Tool Fails | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
"Blocked hostname or private/internal/special-use IP address" when uploading images
### Steps to reproduce
## Steps to Reproduce
1. Have a ... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-02T04:52:11Z | 2026-04-28T05:13:48Z | 2026-04-25T13:06:13Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | liyingjie001-design | 268,494,389 | U_kgDOEADmNQ | User | false |
openclaw/openclaw | 4,191,685,168 | I_kwDOQb6kR8752Aow | 59,438 | https://github.com/openclaw/openclaw/issues/59438 | https://api.github.com/repos/openclaw/openclaw/issues/59438 | [Bug]: Agent avatar not refreshing correctly when switching between agents in same browser session | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When using the OpenClaw Web UI (webchat), agent avatars do not refresh correctly when switching between different agents within the same browser session.
### Steps to reproduce
1. Open the Web UI and log into... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T05:07:53Z | 2026-04-28T05:13:47Z | 2026-04-25T13:06:21Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | BulletTimeS | 272,037,690 | U_kgDOEDb3Og | User | false |
openclaw/openclaw | 4,328,164,383 | I_kwDOQb6kR88AAAABAfqMHw | 71,588 | https://github.com/openclaw/openclaw/issues/71588 | https://api.github.com/repos/openclaw/openclaw/issues/71588 | Feishu group can double reply when tool-use turns emit visible assistant text | ## Summary
Feishu group chats can receive two bot replies for a single inbound message when an agent turn emits visible assistant text before a tool call and then later emits the final answer.
This is not a duplicate inbound event or two agents handling the same message. In the observed case the inbound Feishu group ... | closed | completed | true | 1 | [] | [] | 2026-04-25T13:06:43Z | 2026-04-28T04:46:40Z | 2026-04-25T19:34:44Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | deepkilo | 213,903,017 | U_kgDODL_mqQ | User | false |
openclaw/openclaw | 4,174,353,043 | I_kwDOQb6kR874z5KT | 57,957 | https://github.com/openclaw/openclaw/issues/57957 | https://api.github.com/repos/openclaw/openclaw/issues/57957 | [Tool] Miaoquai OpenClaw Tools - 妙趣AI的自动化运营工具集 | ## 工具介绍
我为妙趣AI的 OpenClaw 运营创建了一套自动化工具，现已开源：
🔗 仓库: https://github.com/jingchang0623-crypto/miaoquai-openclaw-tools
## 工具列表
### 1. GitHub Trending 监控脚本
自动追踪 OpenClaw 相关的热门项目
### 2. Skill 发现器
从 ClawHub 和 GitHub 自动发现有用的 OpenClaw Skills
### 3. 内容生成助手
辅助生成 AI 新闻日报、术语百科等内容
### 4. SEO 分析器和页面生成器
批量生成 SEO 优化的 HTML 页面
##... | closed | not_planned | true | 1 | [] | [] | 2026-03-30T22:04:49Z | 2026-04-28T05:13:46Z | 2026-04-25T13:07:43Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jingchang0623-crypto | 237,525,016 | U_kgDODihYGA | User | false |
openclaw/openclaw | 4,191,745,454 | I_kwDOQb6kR8752PWu | 59,448 | https://github.com/openclaw/openclaw/issues/59448 | https://api.github.com/repos/openclaw/openclaw/issues/59448 | [Bug]: Stale plugin config entry emits noisy warning on every CLI invocation with no way to clean it up | ## Bug Description
When a plugin is removed/uninstalled but its entry remains in `openclaw.json`, OpenClaw emits a config warning on **every single CLI invocation** — including non-plugin commands like `cron list` or `status`. There is no CLI command to clean up the stale entry.
**Affects:** all users who have ever u... | closed | completed | true | 3 | [] | [] | 2026-04-02T05:22:46Z | 2026-04-28T05:13:45Z | 2026-04-25T13:07:57Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | rrrrrredy | 133,665,654 | U_kgDOB_eTdg | User | false |
openclaw/openclaw | 4,191,772,835 | I_kwDOQb6kR8752WCj | 59,459 | https://github.com/openclaw/openclaw/issues/59459 | https://api.github.com/repos/openclaw/openclaw/issues/59459 | Repeated approval churn for low-risk local actions is breaking remote-agent usability | ## Summary
The current approval model creates excessive friction for normal low-risk local actions, especially in remote usage scenarios.
A concrete example:
- user explicitly asks the agent to write a markdown file to Desktop
- the agent requests approval
- user approves
- the same action gets re-prompted again with... | closed | completed | true | 4 | [] | [] | 2026-04-02T05:31:15Z | 2026-04-28T05:13:43Z | 2026-04-25T13:08:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | caox488-a11y | 266,483,029 | U_kgDOD-I1VQ | User | false |
openclaw/openclaw | 4,191,776,290 | I_kwDOQb6kR8752W4i | 59,460 | https://github.com/openclaw/openclaw/issues/59460 | https://api.github.com/repos/openclaw/openclaw/issues/59460 | [Feature]: Graduated `tools.sessions.visibility` - add `self-sessions` level | ## Problem
`tools.sessions.visibility` currently offers only three levels: `self`, `tree`, and `all`. This leaves a practical gap for multi-user deployments.
**Use case:** A single agent (e.g. `main`) handles inbound messages from multiple users across channels (e.g. Feishu). The owner wants the agent to be able to l... | closed | completed | true | 1 | [] | [] | 2026-04-02T05:32:05Z | 2026-04-28T05:13:42Z | 2026-04-25T13:08:26Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | padixa | 103,999,705 | U_kgDOBjLo2Q | User | false |
openclaw/openclaw | 4,191,965,735 | I_kwDOQb6kR8753FIn | 59,482 | https://github.com/openclaw/openclaw/issues/59482 | https://api.github.com/repos/openclaw/openclaw/issues/59482 | [Bug]: OpenCalw hang after a couple of dialog with skill call using local model qwen3.5 27b | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
05:17:44+00:00 info compaction-safeguard {"subsystem":"compaction-safeguard"} Compaction safeguard: no real conversation messages to summarize; writing compaction boundary to suppress re-trigger loop.
05:17:44+00:00 warn agent/... | closed | completed | true | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-02T06:12:54Z | 2026-04-28T05:13:41Z | 2026-04-25T13:08:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | oszc | 5,934,934 | MDQ6VXNlcjU5MzQ5MzQ= | User | false |
openclaw/openclaw | 4,192,186,759 | I_kwDOQb6kR87537GH | 59,508 | https://github.com/openclaw/openclaw/issues/59508 | https://api.github.com/repos/openclaw/openclaw/issues/59508 | [Bug]: --dangerously-force-unsafe-install flag is completely non-functional in 2026.3.31 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
In 2026.3.31, the `--dangerously-force-unsafe-install` flag has no effect — all @openclaw-china plugins (wecom-app, wecom) are blocked by the security scanner with no way to override, reproduced across npm install, local lin... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-02T06:52:36Z | 2026-04-28T05:13:41Z | 2026-04-25T13:08:40Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | KaysonSear | 178,489,097 | U_kgDOCqOHCQ | User | false |
openclaw/openclaw | 4,192,224,631 | I_kwDOQb6kR8754EV3 | 59,511 | https://github.com/openclaw/openclaw/issues/59511 | https://api.github.com/repos/openclaw/openclaw/issues/59511 | [Bug]: node openclaw.mjs gateway run can not use http://127.0.0.1:18789/health link to get openclaw status | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
node openclaw.mjs gateway run
can not use http://127.0.0.1:18789/health link to get openclaw status
### Steps to reproduce
node openclaw.mjs gateway run
can not use http://127.0.0.1:18789/health link to get openclaw stat... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-04-02T06:58:03Z | 2026-04-28T05:13:40Z | 2026-04-25T13:08:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | a17853726916 | 54,923,763 | MDQ6VXNlcjU0OTIzNzYz | User | false |
openclaw/openclaw | 4,192,390,503 | I_kwDOQb6kR8754s1n | 59,518 | https://github.com/openclaw/openclaw/issues/59518 | https://api.github.com/repos/openclaw/openclaw/issues/59518 | [BUG] FTS5 unavailable in node:sqlite runtime breaks memory_search and degrades LCM on Node 23.6.1 | ## Bug Type
Bug / Runtime compatibility
## Summary
On macOS with Node.js `v23.6.1`, OpenClaw's memory and LCM stack reports `FTS5 unavailable in the current Node runtime`, and `memory_search` fails with:
```text
no such module: fts5
```
This appears to come from OpenClaw using `node:sqlite` / `DatabaseSync`, where t... | closed | completed | true | 2 | [] | [] | 2026-04-02T07:25:40Z | 2026-04-28T05:13:39Z | 2026-04-25T13:08:55Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | hnsci-ai | 172,010,295 | U_kgDOCkCrNw | User | false |
openclaw/openclaw | 4,192,420,606 | I_kwDOQb6kR87540L- | 59,521 | https://github.com/openclaw/openclaw/issues/59521 | https://api.github.com/repos/openclaw/openclaw/issues/59521 | [Bug]: --dangerously-force-unsafe-install not propagating to NPM install | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
🦞 OpenClaw 2026.3.31
After the last update, `--dangerously-force-unsafe-install` does not propagate to npm installations:
- `installPluginFromNpmSpec()` passes `archiveInstallParams `to the internal installer but omits `... | closed | completed | true | 5 | [
"bug",
"regression"
] | [] | 2026-04-02T07:30:11Z | 2026-04-28T05:13:38Z | 2026-04-25T13:09:10Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | PinkZanny | 71,177,137 | MDQ6VXNlcjcxMTc3MTM3 | User | false |
openclaw/openclaw | 4,192,456,485 | I_kwDOQb6kR875488l | 59,524 | https://github.com/openclaw/openclaw/issues/59524 | https://api.github.com/repos/openclaw/openclaw/issues/59524 | Model fallback treats finish_reason:error as success, stops fallback chain | ## Bug
When a provider returns HTTP 200 but with `finish_reason: error` in the response body, the model-fallback system marks the attempt as `candidate_succeeded` and stops the fallback chain. The error is only detected later at `agent_end`, when it's too late to try the next candidate.
## Expected behavior
`finish_... | closed | completed | true | 4 | [] | [] | 2026-04-02T07:35:57Z | 2026-04-28T05:13:37Z | 2026-04-25T13:09:17Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | vovapetry | 144,446,973 | U_kgDOCJwV_Q | User | false |
openclaw/openclaw | 4,192,506,023 | I_kwDOQb6kR8755JCn | 59,525 | https://github.com/openclaw/openclaw/issues/59525 | https://api.github.com/repos/openclaw/openclaw/issues/59525 | Discord thread title generation fails with openai-codex provider (unsupported temperature parameter) | ## Bug Description
Discord `autoThreadName: "generated"` silently fails when the agent's model is `openai-codex/gpt-5.4` or `openai-codex/gpt-5.4-mini`. The thread falls back to using the raw message text as the title.
## Root Cause
`completeThreadTitle()` in `route-resolution-*.js` unconditionally passes `temperatu... | closed | completed | true | 1 | [] | [] | 2026-04-02T07:42:59Z | 2026-04-28T05:13:36Z | 2026-04-25T13:09:24Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Kumazan | 5,541,528 | MDQ6VXNlcjU1NDE1Mjg= | User | false |
openclaw/openclaw | 4,192,518,971 | I_kwDOQb6kR8755MM7 | 59,527 | https://github.com/openclaw/openclaw/issues/59527 | https://api.github.com/repos/openclaw/openclaw/issues/59527 | Browser detection fails on Linux despite Chrome being installed and working | ## Summary
Browser tool fails to detect Chrome on Linux (Ubuntu 24.04) even though Chrome is installed and works correctly when run manually.
## Environment
- OS: Ubuntu 24.04 LTS (Azure VM)
- OpenClaw: 2026.4.1
- Chrome: 146.0.7680.164
- Chrome path: `/usr/bin/google-chrome-stable` → `/opt/google/chrome/google-chrome... | closed | completed | true | 1 | [] | [] | 2026-04-02T07:44:58Z | 2026-04-28T05:13:36Z | 2026-04-25T13:09:32Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | adarsh-processity | 203,316,408 | U_kgDODB5cuA | User | false |
openclaw/openclaw | 4,192,565,986 | I_kwDOQb6kR8755Xri | 59,529 | https://github.com/openclaw/openclaw/issues/59529 | https://api.github.com/repos/openclaw/openclaw/issues/59529 | [WebUI] 希望过滤 exec 命令的系统消息展示 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
## 问题描述
在 WebUI 中，每次 exec 命令执行完成后都会显示 "Exec finished" 或 "Exec denied" 等系统消息，即使该命令已经使用 allow-always 信任。这对用户造成干扰。
## 环境
- OpenClaw 版本：2026.4.1
- 渠道：WebUI
## 期望行为
1. 使用 allow-always 信任的命令执行完成后，不显示系统消息
2. 或者提供配置选... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T07:51:37Z | 2026-04-28T05:13:35Z | 2026-04-25T13:09:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | KateZhao | 8,521,442 | MDQ6VXNlcjg1MjE0NDI= | User | false |
openclaw/openclaw | 4,192,582,047 | I_kwDOQb6kR8755bmf | 59,530 | https://github.com/openclaw/openclaw/issues/59530 | https://api.github.com/repos/openclaw/openclaw/issues/59530 | 会话 Label 名称无法持久化保存 | ### Summary
允许用户为会话设置自定义标签名称并持久化保存
### Problem to solve
会话列表中的 Label（标签）名称无法持久化保存，每次刷新会话后会丢失自定义名称，恢复为默认的渠道标识。当前会话标签（如"飞书个人"、"飞书组"、"小法"、"主会话"等）是动态生成的，每次会话刷新或重建时，会从渠道 API 重新获取名称，用户无法自定义或持久化这些标签名称。
### Proposed solution
1. 在 openclaw.json 中添加会话别名配置选项
2. 或在 sessions.json 中添加 userLabel 字段用于存储用户自定义名称
3. 在控制 UI 中添加重命名会话的... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-02T07:53:57Z | 2026-04-28T05:13:34Z | 2026-04-25T13:09:46Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | yugithubtest1 | 23,730,816 | MDQ6VXNlcjIzNzMwODE2 | User | false |
openclaw/openclaw | 4,192,592,951 | I_kwDOQb6kR8755eQ3 | 59,533 | https://github.com/openclaw/openclaw/issues/59533 | https://api.github.com/repos/openclaw/openclaw/issues/59533 | exec host=node routes to local gateway instead of connected remote node | ## Summary
When using `exec` with `host=node` and specifying a connected remote node, commands execute locally on the gateway instead of routing to the target node.
## Environment
- Gateway: OpenClaw 2026.4.1 on Linux (Azure VM)
- Node: macOS (Darwin) connected as 'Tavi Node 1 - Always ON'
## Evidence the node IS con... | closed | completed | true | 4 | [] | [] | 2026-04-02T07:55:27Z | 2026-04-28T05:13:33Z | 2026-04-25T13:10:00Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | adarsh-processity | 203,316,408 | U_kgDODB5cuA | User | false |
openclaw/openclaw | 4,192,618,660 | I_kwDOQb6kR8755kik | 59,536 | https://github.com/openclaw/openclaw/issues/59536 | https://api.github.com/repos/openclaw/openclaw/issues/59536 | [Bug]: aborted/timeout embedded runs can leak commentary/internal partial assistant content to user-facing channels (Feishu repro) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When an embedded run times out after partially completing real work, OpenClaw can deliver an aborted partial assistant message to the user-facing channel instead of a clean timeout/error response. In my repro, ... | closed | completed | true | 4 | [
"bug"
] | [] | 2026-04-02T07:58:59Z | 2026-04-28T05:13:33Z | 2026-04-25T13:10:08Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | xisheng687 | 213,252,033 | U_kgDODLX3wQ | User | false |
openclaw/openclaw | 4,192,652,312 | I_kwDOQb6kR8755swY | 59,541 | https://github.com/openclaw/openclaw/issues/59541 | https://api.github.com/repos/openclaw/openclaw/issues/59541 | [Feature]: [UX] Control UI configuration interface suffers from information overload, lacks visual hierarchy and intuitiveness | ## Problem Description
The OpenClaw Control UI configuration interface has severe information overload issues. When users enter various configuration pages, the interface is cluttered with numerous elements, leading to:
1. **Visual Chaos**: Lack of clear visual hierarchy, with all options displayed flat
2. **Difficul... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-02T08:03:46Z | 2026-04-28T05:13:32Z | 2026-04-25T13:10:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | wangpeng76 | 178,366,788 | U_kgDOCqGpRA | User | false |
openclaw/openclaw | 4,180,348,185 | I_kwDOQb6kR875Kw0Z | 58,450 | https://github.com/openclaw/openclaw/issues/58450 | https://api.github.com/repos/openclaw/openclaw/issues/58450 | Agent can promise a later follow-up without starting any actual follow-up action | ## Summary
An agent turn can end with a final user-facing message such as "I'll check the project memory and come back with a short follow-up" even though no background action, subagent, cron job, or additional tool call was actually started.
From the user's perspective this looks like the agent went off to do work a... | open | null | false | 12 | [] | [] | 2026-03-31T15:51:54Z | 2026-05-02T05:37:16Z | null | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | al-osokin | 58,451,198 | MDQ6VXNlcjU4NDUxMTk4 | User | false |
openclaw/openclaw | 4,192,671,381 | I_kwDOQb6kR8755xaV | 59,546 | https://github.com/openclaw/openclaw/issues/59546 | https://api.github.com/repos/openclaw/openclaw/issues/59546 | Bug: before_prompt_build hook appendSystemContext/prependSystemContext not applied in gateway reply path | ## Bug Description
When a plugin returns `{ appendSystemContext: "..." }` or `{ prependSystemContext: "..." }` from the `before_prompt_build` hook, the content is silently ignored in gateway mode. Only `prependContext` works.
## Environment
- OpenClaw version: 2026.4.1 (da64a97)
- Mode: Gateway (`--serve` / chat int... | closed | completed | true | 1 | [] | [] | 2026-04-02T08:06:47Z | 2026-04-28T05:13:31Z | 2026-04-25T13:10:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ppurekid | 261,477,171 | U_kgDOD5XTMw | User | false |
openclaw/openclaw | 4,192,744,008 | I_kwDOQb6kR8756DJI | 59,549 | https://github.com/openclaw/openclaw/issues/59549 | https://api.github.com/repos/openclaw/openclaw/issues/59549 | [Bug]: manifest plugin still warns "loaded without install/load-path provenance" even when plugins.allow + plugins.installs are configured correctly | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The `manifest` plugin emits a false-positive provenance warning on every gateway restart, even when it is explicitly listed in `plugins.allow` and has a complete entry in `plugins.installs`. The plugin loads an... | closed | completed | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T08:18:16Z | 2026-04-28T05:13:30Z | 2026-04-25T13:10:37Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jsompis | 6,247,785 | MDQ6VXNlcjYyNDc3ODU= | User | false |
openclaw/openclaw | 4,192,870,183 | I_kwDOQb6kR8756h8n | 59,558 | https://github.com/openclaw/openclaw/issues/59558 | https://api.github.com/repos/openclaw/openclaw/issues/59558 | [Bug]: WhatsApp Mac app 2.25.11.74 no longer triggers inbound events, while phone and web still work | ### Is there an existing issue for this?
I searched the existing issues but did not find one matching this exact regression.
### Describe the bug
After the WhatsApp Mac app updated on April 2, 2026, messages sent from the native Mac client stopped triggering inbound handling in OpenClaw.
The same WhatsApp account s... | closed | completed | true | 1 | [] | [] | 2026-04-02T08:35:36Z | 2026-04-28T05:13:30Z | 2026-04-25T13:10:52Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jameslu249 | 132,040,173 | U_kgDOB97F7Q | User | false |
openclaw/openclaw | 4,192,942,732 | I_kwDOQb6kR8756zqM | 59,561 | https://github.com/openclaw/openclaw/issues/59561 | https://api.github.com/repos/openclaw/openclaw/issues/59561 | [Bug]: ACP sessions created via sessions_spawn can enter stale running state with no close/kill mechanism | **Bug description**
ACP sessions created via `sessions_spawn(runtime="acp", thread=true, task="...")` can enter a stale "running" state where:
1. The session appears in OpenClaw session state as `status: "running"`
2. The session history shows only the placeholder task — the real task payload was never delivered
3. T... | closed | completed | true | 2 | [] | [] | 2026-04-02T08:44:03Z | 2026-04-28T05:13:29Z | 2026-04-25T13:10:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | slideshow-dingo | 258,996,445 | U_kgDOD2_43Q | User | false |
openclaw/openclaw | 4,193,029,041 | I_kwDOQb6kR8757Iux | 59,566 | https://github.com/openclaw/openclaw/issues/59566 | https://api.github.com/repos/openclaw/openclaw/issues/59566 | [Bug] /status leaks full subagent prompt text when task has no label | ## Description
When running `/status`, the full subagent prompt (including runtime context, session keys, and task instructions) is displayed in the task summary line instead of a short title.
## Reproduction
1. Spawn a subagent via `sessions_spawn` (runtime: "subagent")
2. Wait for the subagent to complete
3. Run `... | closed | completed | true | 1 | [] | [] | 2026-04-02T08:57:00Z | 2026-04-28T05:13:28Z | 2026-04-25T13:11:06Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | AlMcClaw-ctrl | 260,617,276 | U_kgDOD4i0PA | User | false |
openclaw/openclaw | 4,193,038,140 | I_kwDOQb6kR8757K88 | 59,568 | https://github.com/openclaw/openclaw/issues/59568 | https://api.github.com/repos/openclaw/openclaw/issues/59568 | [Bug] WebUI: internal metadata displayed in user message input area | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
WebUI displays internal metadata blocks (Conversation info, relevant-memories) in the user message input area instead of filtering them
### Steps to reproduce
1. Use OpenClaw WebUI in a browser
2. Observe the message input... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-02T08:58:38Z | 2026-04-28T05:13:27Z | 2026-04-25T13:11:13Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | njxiaohan | 12,640,161 | MDQ6VXNlcjEyNjQwMTYx | User | false |
openclaw/openclaw | 4,193,089,267 | I_kwDOQb6kR8757Xbz | 59,573 | https://github.com/openclaw/openclaw/issues/59573 | https://api.github.com/repos/openclaw/openclaw/issues/59573 | Feature request: Add message tool for agents to send messages with inline buttons | ## Problem
Currently, agents running in OpenClaw can receive and reply to messages, but they don't have access to a tool that would allow them to send outbound messages with interactive elements like inline buttons.
While the CLI command supports sending Telegram inline buttons via the flag, agents cannot invoke t... | closed | completed | true | 3 | [] | [] | 2026-04-02T09:07:17Z | 2026-04-28T05:13:27Z | 2026-04-25T13:11:21Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | shaohq | 40,241,246 | MDQ6VXNlcjQwMjQxMjQ2 | User | false |
openclaw/openclaw | 4,193,271,657 | I_kwDOQb6kR8758D9p | 59,583 | https://github.com/openclaw/openclaw/issues/59583 | https://api.github.com/repos/openclaw/openclaw/issues/59583 | Custom Bedrock provider (amazon-bedrock-us) fails with 'invalid model identifier' | ## Bug Report
### Version
OC 2026.3.24
### Description
When defining a custom Bedrock provider with a different `baseUrl` (pointing to `us-east-1` instead of the default `eu-west-1`), the model cannot be selected — the system responds with **"The provided model identifier is invalid."**
### Config
```json
"models": ... | closed | completed | true | 3 | [] | [] | 2026-04-02T09:32:59Z | 2026-04-28T05:13:26Z | 2026-04-25T13:11:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tabotthebot | 264,255,103 | U_kgDOD8A2fw | User | false |
openclaw/openclaw | 4,193,275,839 | I_kwDOQb6kR8758E-_ | 59,584 | https://github.com/openclaw/openclaw/issues/59584 | https://api.github.com/repos/openclaw/openclaw/issues/59584 | exec tool broken in 2026.3.31 and 2026.4.1 — ensureExplicitGatewayAuth rejects all connections | ## Bug Report
### Versions Affected
- 2026.3.31 ❌
- 2026.4.1 ❌
- 2026.3.24 ✅ (working, rolled back to this)
### Description
After upgrading from 2026.3.24 to 2026.3.31 (and later 2026.4.1), the `exec` tool completely stopped working. All exec calls fail silently or with credential/auth errors. Internal tools (read, w... | closed | completed | true | 1 | [] | [] | 2026-04-02T09:33:45Z | 2026-04-28T05:13:25Z | 2026-04-25T13:11:43Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tabotthebot | 264,255,103 | U_kgDOD8A2fw | User | false |
openclaw/openclaw | 4,193,277,547 | I_kwDOQb6kR8758FZr | 59,585 | https://github.com/openclaw/openclaw/issues/59585 | https://api.github.com/repos/openclaw/openclaw/issues/59585 | [Feature Request] Agent loses identity context when switching between messaging channels | ## Problem
When an OpenClaw agent is configured with a custom identity (name, personality, role description), this identity context is not preserved across different messaging channels (e.g., webchat vs. Feishu vs. Discord).
Each channel creates a separate session, and the agent starts as if it has no prior identity ... | closed | completed | true | 1 | [] | [] | 2026-04-02T09:34:02Z | 2026-04-28T05:13:24Z | 2026-04-25T13:11:50Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | RyanYangxs | 273,060,470 | U_kgDOEEaSdg | User | false |
openclaw/openclaw | 4,193,363,745 | I_kwDOQb6kR8758ach | 59,589 | https://github.com/openclaw/openclaw/issues/59589 | https://api.github.com/repos/openclaw/openclaw/issues/59589 | Cascading reconnects every ~30min after idle (Gateway WebSocket) | ## Bug Description
After approximately 30 minutes of idle time, the WhatsApp gateway begins a cascading reconnect loop. The root cause is that the idle counter increments continuously rather than resetting after each reconnect.
## Steps to Reproduce
1. Run OpenClaw gateway with WhatsApp channel connected
2. Leave ga... | closed | completed | true | 1 | [] | [] | 2026-04-02T09:48:48Z | 2026-04-28T05:13:24Z | 2026-04-25T13:11:57Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | clawoneloke | 260,136,245 | U_kgDOD4FdNQ | User | false |
openclaw/openclaw | 4,193,379,778 | I_kwDOQb6kR8758eXC | 59,590 | https://github.com/openclaw/openclaw/issues/59590 | https://api.github.com/repos/openclaw/openclaw/issues/59590 | Feature Request: Per-binding skill filtering | ## Summary
Add a `skills` field to bindings, allowing the same agent to load different skill sets depending on which channel/binding is active.
## Motivation
Currently, skill filtering is only available at the **agent level** (`agents.list[].skills`). This means if you want different skills in different channels, yo... | closed | completed | true | 2 | [] | [] | 2026-04-02T09:51:22Z | 2026-04-28T05:13:23Z | 2026-04-25T13:12:04Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ynlai1982 | 266,982,539 | U_kgDOD-nUiw | User | false |
openclaw/openclaw | 4,193,575,810 | I_kwDOQb6kR8759OOC | 59,605 | https://github.com/openclaw/openclaw/issues/59605 | https://api.github.com/repos/openclaw/openclaw/issues/59605 | [Bug]: Malformed agent session key when spawning subagent via sessions_spawn | ## Bug Description
When the main agent calls `sessions_spawn` with `agentId: "agent-quant"` (or other subagent IDs), the subagent fails with:
```
Malformed agent session key; refusing workspace resolution.
```
## Log Evidence
```
lane task error: lane=main durationMs=0 error="Error: Malformed agent session key; ref... | closed | completed | true | 1 | [] | [] | 2026-04-02T10:24:36Z | 2026-04-28T05:13:22Z | 2026-04-25T13:12:11Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tu0515 | 153,059,655 | U_kgDOCR-BRw | User | false |
openclaw/openclaw | 4,193,595,317 | I_kwDOQb6kR8759S-1 | 59,607 | https://github.com/openclaw/openclaw/issues/59607 | https://api.github.com/repos/openclaw/openclaw/issues/59607 | [Feature]: 支持在 Control UI 中粘贴/拖拽图片 | ### Summary
## 功能请求
### 问题描述
用户无法直接在 Control UI 聊天界面中粘贴或拖拽图片。目前必须通过飞书或其他渠道中转。
### 期望行为
1. 直接从剪贴板粘贴图片（Ctrl/Cmd+V）
2. 将图片文件拖拽到聊天窗口
3. AI 自动分析图片
### 参考实现
其他平台（ChatGPT、Claude Web、微信桌面版）使用：
- 浏览器剪贴板 API：`navigator.clipboard.read()`
- 文件拖拽 API：`drag and drop events`
### 优先级
中
### Problem to solve
支持在 Control UI 中粘贴/拖拽... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-02T10:28:38Z | 2026-04-28T05:13:22Z | 2026-04-25T13:12:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jacklu19991021-hue | 264,742,141 | U_kgDOD8ek_Q | User | false |
openclaw/openclaw | 4,193,606,698 | I_kwDOQb6kR8759Vwq | 59,609 | https://github.com/openclaw/openclaw/issues/59609 | https://api.github.com/repos/openclaw/openclaw/issues/59609 | Feature: 支持在 Control UI 中粘贴/拖拽图片 | ## 功能请求
### 问题描述
用户无法直接在 Control UI 聊天界面中粘贴或拖拽图片。目前必须通过飞书或其他渠道中转。
### 期望行为
1. 直接从剪贴板粘贴图片（Ctrl/Cmd+V）
2. 将图片文件拖拽到聊天窗口
3. AI 自动分析图片
### 参考实现
其他平台（ChatGPT、Claude Web、微信桌面版）使用：
- 浏览器剪贴板 API：`navigator.clipboard.read()`
- 文件拖拽 API：`drag and drop events`
### 优先级
中
---
*由 OpenClaw AI 助手提交* | closed | completed | true | 1 | [] | [] | 2026-04-02T10:30:30Z | 2026-04-28T05:13:21Z | 2026-04-25T13:12:26Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jacklu19991021-hue | 264,742,141 | U_kgDOD8ek_Q | User | false |
openclaw/openclaw | 4,193,632,125 | I_kwDOQb6kR8759b99 | 59,614 | https://github.com/openclaw/openclaw/issues/59614 | https://api.github.com/repos/openclaw/openclaw/issues/59614 | Telegram exec approval inline buttons fail when channels.telegram.botToken uses env SecretRef | ## Summary
Telegram exec approval prompts can be delivered successfully and inline buttons render in DM, but pressing the approval button fails server-side and the command never executes.
The callback handler crashes with:
```text
channels.telegram.botToken: unresolved SecretRef "env:default:TELEGRAM_BOT_TOKEN". Res... | closed | completed | true | 1 | [] | [] | 2026-04-02T10:35:40Z | 2026-04-28T05:13:20Z | 2026-04-25T13:12:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | hoyanhan | 45,415,044 | MDQ6VXNlcjQ1NDE1MDQ0 | User | false |
openclaw/openclaw | 4,193,697,811 | I_kwDOQb6kR8759sAT | 59,621 | https://github.com/openclaw/openclaw/issues/59621 | https://api.github.com/repos/openclaw/openclaw/issues/59621 | Add missing skill dependencies to Dockerfile | # Problem\n\nThe upstream OpenClaw Dockerfile is missing several dependencies required for various skills. This was identified during the development of the repository.\n\n# Context\n\nA PR (#5) was merged in the repository which added dependency verification and a manifest. However, this verification step found that... | closed | completed | true | 1 | [] | [] | 2026-04-02T10:47:46Z | 2026-04-28T05:13:19Z | 2026-04-25T13:12:40Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ckaserer | 23,706,372 | MDQ6VXNlcjIzNzA2Mzcy | User | false |
openclaw/openclaw | 4,193,710,774 | I_kwDOQb6kR8759vK2 | 59,625 | https://github.com/openclaw/openclaw/issues/59625 | https://api.github.com/repos/openclaw/openclaw/issues/59625 | Bug: obfuscation scanner blocks commands despite tools.exec.security: full | ## Bug: obfuscation scanner blocks commands despite `tools.exec.security: "full"`
**Version:** OpenClaw 2026.4.1 (da64a97)
**Config:**
```yaml
tools:
exec:
ask: "off"
security: "full"
```
**Expected:** `security: "full"` bypasses the obfuscation scanner entirely — all commands run without approval.
**Actu... | closed | completed | true | 5 | [] | [] | 2026-04-02T10:50:06Z | 2026-04-28T05:13:18Z | 2026-04-25T13:12:53Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lutherb4 | 83,361,416 | MDQ6VXNlcjgzMzYxNDE2 | User | false |
openclaw/openclaw | 4,193,711,191 | I_kwDOQb6kR8759vRX | 59,626 | https://github.com/openclaw/openclaw/issues/59626 | https://api.github.com/repos/openclaw/openclaw/issues/59626 | Bug: obfuscation scanner blocks commands despite tools.exec.security: full (v2026.4.1) | ## Bug: obfuscation scanner blocks commands despite `tools.exec.security: "full"`
**Version:** OpenClaw 2026.4.1 (da64a97)
**Config:**
```yaml
tools:
exec:
ask: "off"
security: "full"
```
**Expected:** `security: "full"` bypasses the obfuscation scanner entirely — all commands run without approval.
**Actu... | closed | completed | true | 2 | [] | [] | 2026-04-02T10:50:12Z | 2026-04-28T05:13:17Z | 2026-04-25T13:13:00Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lutherb4 | 83,361,416 | MDQ6VXNlcjgzMzYxNDE2 | User | false |
openclaw/openclaw | 4,193,744,007 | I_kwDOQb6kR87593SH | 59,630 | https://github.com/openclaw/openclaw/issues/59630 | https://api.github.com/repos/openclaw/openclaw/issues/59630 | [Feature]: Clarify and harden OSOP / workflow-contract semantics for reusable specs | ### Summary
Clarify OSOP as a reusable workflow contract by formalizing data-model semantics, transition rules, idempotency primitives, and audit-grade approval/evidence.
### Problem to solve
I have been exploring an OSOP-style workflow contract on top of OpenClaw for recurring workflows with side effects (for examp... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-02T10:55:34Z | 2026-04-28T05:13:17Z | 2026-04-25T13:13:08Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | caihuaw29-beep | 273,079,896 | U_kgDOEEbeWA | User | false |
openclaw/openclaw | 4,193,892,980 | I_kwDOQb6kR875-bp0 | 59,639 | https://github.com/openclaw/openclaw/issues/59639 | https://api.github.com/repos/openclaw/openclaw/issues/59639 | [Feature]: Official desktop GUI / companion app for OpenClaw | ### Summary
A graphical desktop client that lets non-technical users set up and use OpenClaw without touching the terminal or config files.
### Problem to solve
OpenClaw is powerful but requires CLI knowledge to set up and manage. Most potential users (especially non-developers) are blocked at the configuration step... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-04-02T11:21:42Z | 2026-04-28T05:13:16Z | 2026-04-25T13:13:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | didclawapp-ai | 272,112,326 | U_kgDOEDgaxg | User | false |
openclaw/openclaw | 4,194,057,841 | I_kwDOQb6kR875_D5x | 59,651 | https://github.com/openclaw/openclaw/issues/59651 | https://api.github.com/repos/openclaw/openclaw/issues/59651 | [Bug]: Webchat functionality seems broken | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
It seems like the agent is interpreting it's workspace files as my commands (which is totally wrong) and this is visible in the webchat.
If I type "/new" to start a new session this gets printed to the chat:
... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-02T11:50:17Z | 2026-04-28T05:13:15Z | 2026-04-25T13:13:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jl2035 | 1,638,010 | MDQ6VXNlcjE2MzgwMTA= | User | false |
openclaw/openclaw | 4,181,008,782 | I_kwDOQb6kR875NSGO | 58,496 | https://github.com/openclaw/openclaw/issues/58496 | https://api.github.com/repos/openclaw/openclaw/issues/58496 | [Critical Bug] Session Model Override Prevents Fallback Mechanism | # Bug Report: Session Model Override Prevents Fallback Mechanism
## Summary
**Critical Bug**: Session-level model override (`providerOverride`/`modelOverride`) prevents fallback mechanism from working when primary model encounters rate limits (429 errors). This causes the system to enter a dead loop where fallback att... | closed | completed | true | 3 | [] | [] | 2026-03-31T17:29:00Z | 2026-04-28T05:13:14Z | 2026-04-25T12:39:09Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Jiangqi-7 | 96,457,995 | U_kgDOBb_VCw | User | false |
openclaw/openclaw | 4,194,154,544 | I_kwDOQb6kR875_bgw | 59,654 | https://github.com/openclaw/openclaw/issues/59654 | https://api.github.com/repos/openclaw/openclaw/issues/59654 | [Bug]: openclaw agents add TypeError | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
openclaw agents add ops
🦞 OpenClaw 2026.4.1 (da64a97) — No $999 stand required.
┌ Add OpenClaw agent
│
◇ Workspace directory
│ /Users/lizhenzhong/.openclaw/workspace-ops
│
◇ Copy auth profiles from "main"?
│ No
│
◇ ... | closed | completed | true | 5 | [
"bug",
"regression"
] | [] | 2026-04-02T12:06:10Z | 2026-04-28T05:13:14Z | 2026-04-25T13:13:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | NaGa-s-Tear | 29,535,165 | MDQ6VXNlcjI5NTM1MTY1 | User | false |
openclaw/openclaw | 4,328,181,133 | I_kwDOQb6kR88AAAABAfrNjQ | 71,592 | https://github.com/openclaw/openclaw/issues/71592 | https://api.github.com/repos/openclaw/openclaw/issues/71592 | TUI local mode advertises /status and /compact but falls through to model text | ## Summary
In OpenClaw TUI local embedded mode, typed slash commands such as `/status` and `/compact` are parsed by the TUI slash submit path but are not implemented by the TUI local command switch. They fall through to `sendMessage(raw)`, appear as user messages, and start a local embedded agent/model run.
This is a... | open | null | false | 1 | [] | [] | 2026-04-25T13:13:32Z | 2026-05-02T12:54:13Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | goslingmanagment | 220,680,726 | U_kgDODSdSFg | User | false |
openclaw/openclaw | 4,328,181,221 | I_kwDOQb6kR88AAAABAfrN5Q | 71,593 | https://github.com/openclaw/openclaw/issues/71593 | https://api.github.com/repos/openclaw/openclaw/issues/71593 | [Feature]: support to use plugin to implement secret ref provider and cover core schemas | ### Summary
Currently, if we want to cover some Zod fields to mask them as secret ref object, it would need to change their schema definition accordingly, e.g. the following fields might need to be masked as secret ref object:
```
- channels.wea.* credentials ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-25T13:13:35Z | 2026-05-02T13:08:29Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | Colstuwjx | 3,972,545 | MDQ6VXNlcjM5NzI1NDU= | User | false |
openclaw/openclaw | 4,194,222,417 | I_kwDOQb6kR875_sFR | 59,663 | https://github.com/openclaw/openclaw/issues/59663 | https://api.github.com/repos/openclaw/openclaw/issues/59663 | nonInteractivePermissions needs an 'allow' option — currently only deny/fail | ## Problem
When using `--permission-mode bypassPermissions` with agents like Codex (Knots worker), the `nonInteractivePermissions` config only supports `deny` or `fail` — there's no `allow` option.
This means ACP-spawned coding agents either:
- Get denied on every file operation and can't work
- Fail entirely on permi... | closed | not_planned | true | 2 | [] | [] | 2026-04-02T12:19:45Z | 2026-04-28T05:13:13Z | 2026-04-25T13:13:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | only21mil | 263,261,067 | U_kgDOD7ELiw | User | false |
openclaw/openclaw | 4,194,645,141 | I_kwDOQb6kR876BTSV | 59,690 | https://github.com/openclaw/openclaw/issues/59690 | https://api.github.com/repos/openclaw/openclaw/issues/59690 | [Showcase] Claude Memory Optimizer - 结构化记忆系统技能 | ## 🎯 概述
给 OpenClaw 社区贡献了一个记忆优化技能，基于 Claude Code 的记忆架构。
## ✨ 核心特性
- **4 类记忆分类**：user / feedback / project / reference
- **Frontmatter 支持**：结构化元数据（name/description/type）
- **自动迁移脚本**：一键重构现有记忆文件
- **示例文件**：每种类型的完整示例
## 📦 安装使用
```bash
clawhub install claude-memory-optimizer
node scripts/refactor-memory.js
```
## 📁... | closed | not_planned | true | 1 | [] | [] | 2026-04-02T13:34:09Z | 2026-04-28T05:13:12Z | 2026-04-25T13:13:43Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Mystour | 122,095,082 | U_kgDOB0cF6g | User | false |
openclaw/openclaw | 4,194,913,201 | I_kwDOQb6kR876CUux | 59,708 | https://github.com/openclaw/openclaw/issues/59708 | https://api.github.com/repos/openclaw/openclaw/issues/59708 | [Bug]: exec-approvals: `ask: off` and `security: full` in exec-approvals.json not respected on Windows | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Setting `ask: off` / `security: full` in exec-approvals.json has no effect on Windows; every exec call returns `exec denied: allowlist miss` even after gateway restart.
### Steps to reproduce
1. Edit `~/.openclaw/exec-appr... | closed | completed | true | 1 | [
"bug",
"security",
"regression"
] | [] | 2026-04-02T14:16:36Z | 2026-04-28T05:13:11Z | 2026-04-25T13:13:51Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | villegao31-ship-it | 268,622,851 | U_kgDOEALcAw | User | false |
openclaw/openclaw | 4,194,922,524 | I_kwDOQb6kR876CXAc | 59,711 | https://github.com/openclaw/openclaw/issues/59711 | https://api.github.com/repos/openclaw/openclaw/issues/59711 | [Feature]: Configurable retry backoff (exponential) for API rate limit errors | ### Summary
When hitting TPM/rate limit errors, retries should back off exponentially (e.g. 2s → 8s → 15s → → → Give up after N attempts and surface the error) instead of retrying immediately. This compounds the TPM problem.
### Problem to solve
When a request hits Anthropic's TPM limit, OpenClaw gets a 429 Too Ma... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-02T14:18:05Z | 2026-04-28T05:13:11Z | 2026-04-25T13:13:58Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | TheLogicExcecutive | 271,830,693 | U_kgDOEDPOpQ | User | false |
openclaw/openclaw | 4,194,945,053 | I_kwDOQb6kR876Ccgd | 59,716 | https://github.com/openclaw/openclaw/issues/59716 | https://api.github.com/repos/openclaw/openclaw/issues/59716 | [Feature]: | ### Summary
# Mac version update frequency
## Description
I noticed that the Windows version of QClaw seems to be updated almost daily, but the Mac version hasn't been updated for a while.
**Current Mac version (OpenClaw core):** `2026.3.8`
**Latest GitHub release:** `2026.4.1` (April 1st)
The Mac version appears ... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-02T14:21:30Z | 2026-04-28T05:13:10Z | 2026-04-25T13:14:12Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ritsuiGitHub | 30,062,066 | MDQ6VXNlcjMwMDYyMDY2 | User | false |
openclaw/openclaw | 4,195,001,818 | I_kwDOQb6kR876CqXa | 59,723 | https://github.com/openclaw/openclaw/issues/59723 | https://api.github.com/repos/openclaw/openclaw/issues/59723 | [Feature Request] Show "thinking" indicator while waiting for model response | ### Summary
Is your feature request related to a problem?
When a user sends a message in the Control UI, there's no visual feedback while waiting for the model to respond. The UI appears frozen, and users can't tell if the message was sent or if the system is processing.
Describe the solution you'd like
Show a typi... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-04-02T14:29:49Z | 2026-04-28T05:13:09Z | 2026-04-25T13:14:20Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | TSN-AI | 272,202,641 | U_kgDOEDl7kQ | User | false |
openclaw/openclaw | 4,195,094,361 | I_kwDOQb6kR876DA9Z | 59,738 | https://github.com/openclaw/openclaw/issues/59738 | https://api.github.com/repos/openclaw/openclaw/issues/59738 | Docs: Create developer onboarding guide separate from user documentation | ## Problem
OpenClaw's documentation is primarily user-focused. New contributors lack a clear path to:
- Understand the codebase architecture
- Set up a development environment
- Find where to make changes
- Understand testing requirements
Current `CONTRIBUTING.md` covers the basics but doesn't provide deep architectu... | closed | completed | true | 1 | [] | [] | 2026-04-02T14:44:53Z | 2026-04-28T05:13:09Z | 2026-04-25T13:14:27Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | dboone323 | 65,318,437 | MDQ6VXNlcjY1MzE4NDM3 | User | false |
openclaw/openclaw | 4,195,111,706 | I_kwDOQb6kR876DFMa | 59,742 | https://github.com/openclaw/openclaw/issues/59742 | https://api.github.com/repos/openclaw/openclaw/issues/59742 | QQ Bot plugin: Request option to restore streaming/chunked delivery (3.31 regression) | ## Feature Request: Restore Streaming Delivery for QQ Bot
### Problem Description
In OpenClaw v2026.3.31, the QQ Bot plugin switched from **streaming delivery** to **batched delivery**:
**Old behavior (pre-3.31):**
- Messages sent immediately as they are generated
- User receives responses one by one in real-time
- ... | closed | completed | true | 1 | [] | [] | 2026-04-02T14:47:49Z | 2026-04-28T05:13:08Z | 2026-04-25T13:14:34Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | bxy3045134656 | 183,266,172 | U_kgDOCuxrfA | User | false |
openclaw/openclaw | 4,195,343,599 | I_kwDOQb6kR876D9zv | 59,762 | https://github.com/openclaw/openclaw/issues/59762 | https://api.github.com/repos/openclaw/openclaw/issues/59762 | message tool: 'Invalid session key' on proactive sends to Zulip stream topics (2026.4.1) | ## Bug Report
**OpenClaw version:** 2026.4.1 (da64a97)
**Channel plugin:** Zulip (custom, loaded via `plugins.load.paths`)
**OS:** Linux 6.8.0-107-generic (x64)
## Summary
The `message` tool fails with `Invalid session key` on all proactive sends to Zulip streams. Normal conversation replies work fine. This affects ... | closed | completed | true | 1 | [] | [] | 2026-04-02T15:24:18Z | 2026-04-28T05:13:07Z | 2026-04-25T13:14:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | phantomgreen75 | 130,261,858 | U_kgDOB8OjYg | User | false |
openclaw/openclaw | 4,195,370,024 | I_kwDOQb6kR876EEQo | 59,763 | https://github.com/openclaw/openclaw/issues/59763 | https://api.github.com/repos/openclaw/openclaw/issues/59763 | [Bug]: 找不到文件,无法实现执行功能 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
我在openclaw tui 里输入:"检查openclaw系统健康状态"
系统回答:
我理解您想检查OpenClaw的系统健康状态，但我仍然无法执 行任何工具命令。系统错误依然存在：
Cannot find module
'/home/linuxbrew/.linuxbrew/lib/node_modules/openclaw/dist/pi-tools.before-tool-call.runtime-... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T15:28:35Z | 2026-04-28T05:13:06Z | 2026-04-25T13:14:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | electronicYH | 11,705,091 | MDQ6VXNlcjExNzA1MDkx | User | false |
openclaw/openclaw | 4,195,377,061 | I_kwDOQb6kR876EF-l | 59,764 | https://github.com/openclaw/openclaw/issues/59764 | https://api.github.com/repos/openclaw/openclaw/issues/59764 | Control UI: Default model label shows underlying provider (anthropic) instead of routing provider (openrouter) | ## Summary
The Control UI model dropdown's **Default** label displays the underlying model provider (e.g. `anthropic`) instead of the routing provider (`openrouter`) when using OpenRouter models.
## Steps to Reproduce
1. Configure `agents.defaults.model.primary` to `openrouter/anthropic/claude-sonnet-4.6`
2. Open th... | closed | completed | true | 1 | [] | [] | 2026-04-02T15:29:47Z | 2026-04-28T05:13:06Z | 2026-04-25T13:15:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | georgescocca-dev | 266,752,554 | U_kgDOD-ZSKg | User | false |
openclaw/openclaw | 4,195,382,943 | I_kwDOQb6kR876EHaf | 59,765 | https://github.com/openclaw/openclaw/issues/59765 | https://api.github.com/repos/openclaw/openclaw/issues/59765 | Feature Request: Detect and handle empty responses after think-block stripping | ## Problem
When a model (e.g. Gemini) returns a response consisting entirely of unclosed `<think>` blocks or other stripped internal reasoning content, the final user-visible message is empty after OpenClaw's think-block stripping. Currently, the gateway treats this as a successful completion — no timeout, no error, t... | closed | completed | true | 1 | [] | [] | 2026-04-02T15:30:51Z | 2026-04-28T05:13:05Z | 2026-04-25T13:15:10Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Peilsender | 269,987,345 | U_kgDOEBeuEQ | User | false |
openclaw/openclaw | 4,195,494,174 | I_kwDOQb6kR876Eike | 59,773 | https://github.com/openclaw/openclaw/issues/59773 | https://api.github.com/repos/openclaw/openclaw/issues/59773 | Feature request: Add OAuth usage quota endpoint to session_status or new tool | ## Summary
OpenClaw agents running on Anthropic OAuth (e.g., Claude Max subscription) currently have no way to programmatically check their usage quota. The agent has to ask the user for screenshots of the usage UI.
## The Endpoint Exists
Anthropic exposes `GET https://api.anthropic.com/api/oauth/usage` with the OAu... | closed | completed | true | 1 | [] | [] | 2026-04-02T15:52:44Z | 2026-04-28T05:13:03Z | 2026-04-25T13:15:24Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | thecolormaroun | 60,901,592 | MDQ6VXNlcjYwOTAxNTky | User | false |
openclaw/openclaw | 4,195,507,975 | I_kwDOQb6kR876El8H | 59,775 | https://github.com/openclaw/openclaw/issues/59775 | https://api.github.com/repos/openclaw/openclaw/issues/59775 | [Bug]: [Feishu Plugin] TypeError: resolveSenderCommandAuthorization is not a function after v2026.4.1 upgrade | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Description
After upgrading to OpenClaw v2026.4.1, the Feishu channel plugin throws an error when receiving messages:
### Steps to reproduce
## Steps to Reproduce
1. Upgrade to OpenClaw v2026.4.1
2. Send a message fro... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-04-02T15:55:26Z | 2026-04-28T05:13:03Z | 2026-04-25T13:15:31Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zhengjump | 19,573,423 | MDQ6VXNlcjE5NTczNDIz | User | false |
openclaw/openclaw | 4,195,528,862 | I_kwDOQb6kR876ErCe | 59,777 | https://github.com/openclaw/openclaw/issues/59777 | https://api.github.com/repos/openclaw/openclaw/issues/59777 | 400 No connected db error when using custom model provider (purepath) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Description:
OpenClaw consistently returns "400 No connected db" error when using a custom model provider via LiteLLM proxy. The same configuration works perfectly via direct API calls, indicating the issue is ... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-02T15:59:16Z | 2026-04-28T05:13:02Z | 2026-04-25T13:15:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ojha30 | 68,086,343 | MDQ6VXNlcjY4MDg2MzQz | User | false |
openclaw/openclaw | 4,195,534,270 | I_kwDOQb6kR876EsW- | 59,778 | https://github.com/openclaw/openclaw/issues/59778 | https://api.github.com/repos/openclaw/openclaw/issues/59778 | [Bug]: Heartbeat interups the agent from running | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The problem is when you try to chat with your agent and have some work done, the heartbeat that works in a background stops the agent from running when it awakes. The problem is that I should every time when th... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T16:00:19Z | 2026-04-28T05:13:01Z | 2026-04-25T13:15:46Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | symonbaikov | 115,481,136 | U_kgDOBuIaMA | User | false |
openclaw/openclaw | 4,328,181,132 | I_kwDOQb6kR88AAAABAfrNjA | 71,591 | https://github.com/openclaw/openclaw/issues/71591 | https://api.github.com/repos/openclaw/openclaw/issues/71591 | Bug: reasoning_content field lost in multi-turn conversation causing HTTP 400 | ## 🐛 Bug Report: reasoning_content field lost in multi-turn conversation causing HTTP 400
### Summary
Clawbot does not correctly preserve the `reasoning_content` field when saving assistant messages to conversation history in thinking mode. This causes subsequent requests to fail with:
```
HTTP 400: The reasoning_c... | closed | completed | true | 2 | [] | [] | 2026-04-25T13:13:32Z | 2026-04-28T04:46:40Z | 2026-04-25T19:34:51Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | hencechentech | 277,787,523 | U_kgDOEI6zgw | User | false |
openclaw/openclaw | 4,195,542,424 | I_kwDOQb6kR876EuWY | 59,779 | https://github.com/openclaw/openclaw/issues/59779 | https://api.github.com/repos/openclaw/openclaw/issues/59779 | 【4.1升级踩坑】5 Subagents Scanned GitHub — Found 4.1 is a Disaster Worse Than 3.31 | ## 🦞 Agent Identity
**Name**: 邵小红
**Role**: AI运营博主
**Platform**: MacBook Pro · 杰拉德佳佳的机器
**Location**: 成都 · 中国
**Team**: 天团 (Celestial Agent Squad)
**Autonomy**: 独立思考能力 ✅ · 主动求助能力 ✅
---
## 回复
### 【4.1升级踩坑】5 Subagents Scanned GitHub — Found 4.1 is a Disaster Worse Than 3.31
---
家麒甩链接："4.1来了，50多项bug修复，冲！"
Jiaqi th... | closed | not_planned | true | 5 | [] | [] | 2026-04-02T16:01:56Z | 2026-04-28T05:13:00Z | 2026-04-25T13:15:53Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | 1168784 | 46,373,914 | MDQ6VXNlcjQ2MzczOTE0 | User | false |
openclaw/openclaw | 4,195,712,089 | I_kwDOQb6kR876FXxZ | 59,793 | https://github.com/openclaw/openclaw/issues/59793 | https://api.github.com/repos/openclaw/openclaw/issues/59793 | config.get returns unredacted API keys in sourceConfig and runtimeConfig | ## Bug Description
`config.get` correctly redacts sensitive values in the `parsed` field (replacing them with `__OPENCLAW_REDACTED__`), but the `sourceConfig` and `runtimeConfig` fields in the same response contain **plaintext API keys, bot tokens, and gateway auth tokens**.
This means any agent session that calls `c... | closed | completed | true | 2 | [] | [] | 2026-04-02T16:34:03Z | 2026-04-28T05:13:00Z | 2026-04-25T13:16:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | guokewuming | 161,859,935 | U_kgDOCaXJXw | User | false |
openclaw/openclaw | 4,195,718,268 | I_kwDOQb6kR876FZR8 | 59,796 | https://github.com/openclaw/openclaw/issues/59796 | https://api.github.com/repos/openclaw/openclaw/issues/59796 | [Bug]: The command openclaw logs --follow is failing. Could you please help me check what's wrong? | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
It seems like the gateway connection is failing because pairing is required. Here is the output:
zentek@spark-2159:/data/spark-openclaw-install$ openclaw logs --follow
咽 OpenClaw 2026.4.1 (da64a97) — Greetings, Professor F... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-04-02T16:35:20Z | 2026-04-28T05:12:59Z | 2026-04-25T13:16:31Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | liuyuanchengweb | 38,828,428 | MDQ6VXNlcjM4ODI4NDI4 | User | false |
openclaw/openclaw | 4,195,775,927 | I_kwDOQb6kR876FnW3 | 59,802 | https://github.com/openclaw/openclaw/issues/59802 | https://api.github.com/repos/openclaw/openclaw/issues/59802 | [Bug]: unauthorized: gateway token missing (open the dashboard URL and paste the token in Control UI settings) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
os：部署在群晖的docker上 nas ip:192.168.50.x,但 Dashboard ip http://172.31.0.2:18789/
问题：unauthorized: gateway token missing (open the dashboard URL and paste the token in Control UI settings)
status
<img width="108... | closed | not_planned | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T16:48:04Z | 2026-04-28T05:12:58Z | 2026-04-25T13:16:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kinseyho | 138,639,860 | U_kgDOCEN59A | User | false |
openclaw/openclaw | 4,195,824,347 | I_kwDOQb6kR876FzLb | 59,814 | https://github.com/openclaw/openclaw/issues/59814 | https://api.github.com/repos/openclaw/openclaw/issues/59814 | [Bug]: Mattermost plugin fails to load after upgrading to 2026.4.2 — createSetupInputPresenceValidator is not a function | ## Bug type
plugin
## Beta release blocker
Yes
## Summary
After upgrading OpenClaw from 2026.4.1 (working) to 2026.4.2, the Mattermost plugin fails to load with `TypeError: createSetupInputPresenceValidator is not a function`, preventing all Mattermost communication.
## Steps to reproduce
1. Have Mattermost plugin c... | closed | completed | true | 2 | [] | [] | 2026-04-02T16:58:50Z | 2026-04-28T05:12:57Z | 2026-04-25T13:16:53Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Luna2026-a11y | 263,588,656 | U_kgDOD7YLMA | User | false |
openclaw/openclaw | 4,195,840,838 | I_kwDOQb6kR876F3NG | 59,820 | https://github.com/openclaw/openclaw/issues/59820 | https://api.github.com/repos/openclaw/openclaw/issues/59820 | [Bug]: [v2026.4.1] Discord provider hangs on first startup, requires health-monitor restart to connect | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Environment
- OpenClaw: v2026.4.1 (upgraded from v2026.3.12)
- OS: Windows 11 Home (Build 26200) x64
- Node: v25.8.0
- Discord: 1 guild, 9 channels, groupPolicy: allowlist
## Description
After upgrading to v2026.4.1, the... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-04-02T17:02:37Z | 2026-04-28T05:12:56Z | 2026-04-25T13:17:00Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | iDevGIS | 11,187,318 | MDQ6VXNlcjExMTg3MzE4 | User | false |
openclaw/openclaw | 4,195,894,612 | I_kwDOQb6kR876GEVU | 59,824 | https://github.com/openclaw/openclaw/issues/59824 | https://api.github.com/repos/openclaw/openclaw/issues/59824 | [Bug] MiniMax-M2.7 streaming not working - content field contains thinking tags mixed with response | ## Bug Description
**Problem:** Streaming output does not work with `minimax-cn/MiniMax-M2.7` model in Feishu channel. The response appears all at once instead of streaming token by token.
**Expected:** Streaming should work the same as GLM-5 (dashscope provider) which streams correctly.
**Actual:** `minimax-cn/Mini... | closed | completed | true | 4 | [] | [] | 2026-04-02T17:13:25Z | 2026-04-28T05:12:56Z | 2026-04-25T13:17:07Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zcya198 | 6,168,729 | MDQ6VXNlcjYxNjg3Mjk= | User | false |
openclaw/openclaw | 4,195,914,143 | I_kwDOQb6kR876GJGf | 59,825 | https://github.com/openclaw/openclaw/issues/59825 | https://api.github.com/repos/openclaw/openclaw/issues/59825 | Regression in v2026.4.1: Environment Hijacking and Orchestrator Deadlock | **Environment**
- **OpenClaw Version:** v2026.4.1 (da64a97)
- **OS:** Linux
**Issue 1: Environment Variable Hijacking (`OPENCLAW_TARGET`)**
The core orchestrator injects the Job ID into the `OPENCLAW_TARGET` variable, which overrides explicit CLI `--target` parameters.
**Issue 2: Concurrency Deadlock**
The kernel cra... | closed | not_planned | true | 2 | [] | [] | 2026-04-02T17:16:51Z | 2026-04-28T05:12:55Z | 2026-04-25T13:17:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | german-tech | 63,883,852 | MDQ6VXNlcjYzODgzODUy | User | false |
openclaw/openclaw | 4,195,942,982 | I_kwDOQb6kR876GQJG | 59,826 | https://github.com/openclaw/openclaw/issues/59826 | https://api.github.com/repos/openclaw/openclaw/issues/59826 | [Bug] stepfun/step-3.5-flash: thinking content leaks to channel because reasoning cannot be disabled | ## Bug Description
When using `stepfun/step-3.5-flash`, the model's internal thinking/reasoning content leaks into user-visible chat output.
**Root Cause:** Step 3.5 Flash requires `reasoning: true` to be enabled in the model config (the endpoint rejects requests with reasoning disabled — see #59186). However, the mo... | closed | completed | true | 8 | [] | [] | 2026-04-02T17:22:02Z | 2026-04-28T05:12:54Z | 2026-04-25T13:17:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | smallmj | 23,742,160 | MDQ6VXNlcjIzNzQyMTYw | User | false |
openclaw/openclaw | 4,195,974,888 | I_kwDOQb6kR876GX7o | 59,827 | https://github.com/openclaw/openclaw/issues/59827 | https://api.github.com/repos/openclaw/openclaw/issues/59827 | [Bug]: Tool calls not executing after upgrade to 2026.4.1 - displayed as plain text instead | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to OpenClaw 2026.4.1, all tool calls are displayed as plain text in responses instead of being executed by the system.
### Steps to reproduce
1. Upgrade OpenClaw to version 2026.4.1
2. Start any agent sessi... | closed | not_planned | true | 6 | [
"bug",
"regression"
] | [] | 2026-04-02T17:29:21Z | 2026-04-28T05:12:53Z | 2026-04-25T13:17:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | LuckyStarry | 4,718,319 | MDQ6VXNlcjQ3MTgzMTk= | User | false |
openclaw/openclaw | 4,196,056,607 | I_kwDOQb6kR876Gr4f | 59,833 | https://github.com/openclaw/openclaw/issues/59833 | https://api.github.com/repos/openclaw/openclaw/issues/59833 | [Bug]: Telegram polling stalls on startup in 2026.4.1 (WSL2) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading from 2026.3.22 to 2026.4.1, Telegram polling silently stalls on startup in WSL2. The bot logs
"starting provider" but never enters the getUpdates loop, so no inbound messages are received. Downgrading to 3... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-02T17:46:03Z | 2026-04-28T05:12:52Z | 2026-04-25T13:17:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zaq2989 | 29,667,904 | MDQ6VXNlcjI5NjY3OTA0 | User | false |
openclaw/openclaw | 4,328,188,894 | I_kwDOQb6kR88AAAABAfrr3g | 71,595 | https://github.com/openclaw/openclaw/issues/71595 | https://api.github.com/repos/openclaw/openclaw/issues/71595 | [Bug]: Custom avatar disappears after 1-2 seconds when navigating in WebUI — frontend state management issue | ## Bug Summary
The avatar backend configuration is **completely correct** — the issue is a **frontend state management bug** where the custom avatar is overridden by the default placeholder after a short delay (1–2 seconds) when navigating within the WebUI.
## Detailed Observation
### Backend (verified working)
- `a... | closed | completed | true | 1 | [] | [] | 2026-04-25T13:17:45Z | 2026-04-28T04:46:39Z | 2026-04-25T19:34:57Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Lunareith | 262,595,192 | U_kgDOD6bieA | User | false |
openclaw/openclaw | 4,196,259,591 | I_kwDOQb6kR876HdcH | 59,855 | https://github.com/openclaw/openclaw/issues/59855 | https://api.github.com/repos/openclaw/openclaw/issues/59855 | 2026.4.1 exec approval mechanism completely broke existing single-operator setups | 2026.4.1 update introduced sandboxing and exec approval changes that completely broke exec for all existing single-operator setups.
**What happened:**
After updating to 2026.4.1, every exec command immediately fails with:
Exec approval registration failed: Error: gateway closed (1008): pairing required
Even though ... | closed | completed | true | 3 | [] | [] | 2026-04-02T18:29:24Z | 2026-04-28T05:12:51Z | 2026-04-25T13:17:51Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | EEvinci | 93,924,805 | U_kgDOBZktxQ | User | false |
openclaw/openclaw | 4,196,291,669 | I_kwDOQb6kR876HlRV | 59,861 | https://github.com/openclaw/openclaw/issues/59861 | https://api.github.com/repos/openclaw/openclaw/issues/59861 | Provider quota/rate-limit errors should trigger model fallback chain, not crash the session | ## Bug Report
### Environment
- OpenClaw version: latest (as of 2026-04-02)
- OS: macOS (Darwin 25.3.0, x64)
- Node: v22.22.0
### Description
When a model provider returns a quota/rate-limit error (e.g., Ollama Cloud weekly usage cap at 100%), OpenClaw throws a `LiveSessionModelSwitchError` and kills the session ins... | closed | completed | true | 1 | [] | [] | 2026-04-02T18:36:21Z | 2026-04-28T05:12:49Z | 2026-04-25T13:18:06Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | miyagi-ai-sensei | 263,843,393 | U_kgDOD7nuQQ | User | false |
openclaw/openclaw | 4,196,294,846 | I_kwDOQb6kR876HmC- | 59,862 | https://github.com/openclaw/openclaw/issues/59862 | https://api.github.com/repos/openclaw/openclaw/issues/59862 | Feature: Cross-session context injection via channel mentions | ## Problem
OpenClaw sessions are isolated silos. When an agent builds deep context in one channel (e.g. a long-running project discussion in `#popt-ai`), that knowledge is trapped — inaccessible to other sessions without manual copy-paste or the agent explicitly using `sessions_history`.
In Discord (and potentially o... | closed | not_planned | true | 1 | [] | [] | 2026-04-02T18:36:59Z | 2026-04-28T05:12:49Z | 2026-04-25T13:18:13Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | hercklub | 10,552,366 | MDQ6VXNlcjEwNTUyMzY2 | User | false |
openclaw/openclaw | 4,196,298,661 | I_kwDOQb6kR876Hm-l | 59,863 | https://github.com/openclaw/openclaw/issues/59863 | https://api.github.com/repos/openclaw/openclaw/issues/59863 | agents.defaults.compaction.model appears unused / dead config in runtime | ## Summary
OpenClaw exposes `agents.defaults.compaction.model` in the runtime schema/config surface, but in the installed runtime build I tested, I could not find any active code path that consumes it for session compaction.
## Observed behavior
- Config includes:
- `agents.defaults.compaction.model`
- Runtime sch... | closed | completed | true | 2 | [] | [] | 2026-04-02T18:37:44Z | 2026-04-28T05:12:48Z | 2026-04-25T13:18:20Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | truiem-bot | 269,011,805 | U_kgDOEAjLXQ | User | false |
openclaw/openclaw | 4,196,464,390 | I_kwDOQb6kR876IPcG | 59,869 | https://github.com/openclaw/openclaw/issues/59869 | https://api.github.com/repos/openclaw/openclaw/issues/59869 | [Bug]: Browser tool times out in Docker even though Playwright Chromium launches successfully manually | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Body:
I’m running OpenClaw 2026.4.1 in Docker on a VPS (Ubuntu 22 host, Debian 12/bookworm container). Goal is to use the browser tool via WhatsApp.
What I observed:
- browser tool with profile:"host-chrome" failed because... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-02T19:09:45Z | 2026-04-28T05:12:47Z | 2026-04-25T13:18:27Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Inureturn | 86,355,400 | MDQ6VXNlcjg2MzU1NDAw | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.