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 | 4,075,756,690 | I_kwDOQb6kR87y7xyS | 46,114 | https://github.com/openclaw/openclaw/issues/46114 | https://api.github.com/repos/openclaw/openclaw/issues/46114 | CI: pnpm test:extensions is failing on main | ### Summary
As of March 14, 2026, the extension test lane appears to be red on `main`, not just on unrelated PRs.
### Evidence
- `CI / checks (node, extensions, pnpm test:extensions)` failed on `main` run `23087164772`
- the same lane also failed on unrelated PR `#46051` run `23087380217`
- local reproduction on a cle... | closed | completed | true | 1 | [] | [] | 2026-03-14T12:46:50Z | 2026-04-27T04:43:32Z | 2026-04-25T01:09:29Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | javery556 | 128,285,906 | U_kgDOB6V80g | User | false |
openclaw/openclaw | 4,201,476,956 | I_kwDOQb6kR876bXNc | 60,447 | https://github.com/openclaw/openclaw/issues/60447 | https://api.github.com/repos/openclaw/openclaw/issues/60447 | Bug: openclaw cron edit --tools fails with agentTurn payload (isolated sessions cannot get exec permission) | ## Bug: `openclaw cron edit --tools` fails with agentTurn payload
### Environment
- OpenClaw version: 2026.4.1 (da64a97)
- OS: Darwin 25.3.0 (arm64)
### Steps to Reproduce
```bash
# 1. Create a cron job with systemEvent payload
openclaw cron add --name test --cron "0 16 * * *" --session main --system-event "test"
#... | closed | duplicate | true | 1 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-04-03T17:26:32Z | 2026-04-27T04:43:30Z | 2026-04-25T01:09:46Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | 6tizer | 260,656,772 | U_kgDOD4lOhA | User | false |
openclaw/openclaw | 4,120,177,855 | I_kwDOQb6kR871lOy_ | 52,806 | https://github.com/openclaw/openclaw/issues/52806 | https://api.github.com/repos/openclaw/openclaw/issues/52806 | [Bug]: Cron: Isolated sessions not executing | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Cron jobs with sessionTarget: "isolated" are not executing. Manual trigger via cron run returns enqueued: true but no session is created and no run records appear in cron runs.
### Steps to reproduce
Create a cron job with sessionTarget: ... | closed | duplicate | true | 2 | [
"bug",
"close:duplicate",
"dedupe:child",
"bug:behavior"
] | [] | 2026-03-23T11:30:03Z | 2026-04-27T04:43:30Z | 2026-04-25T01:09:46Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | wfenger | 95,682,725 | U_kgDOBbQApQ | User | false |
openclaw/openclaw | 4,075,759,446 | I_kwDOQb6kR87y7ydW | 46,116 | https://github.com/openclaw/openclaw/issues/46116 | https://api.github.com/repos/openclaw/openclaw/issues/46116 | Preserve existing session `origin` when heartbeat/system events update an existing session | ### Summary
A heartbeat poll or other internal/system event can overwrite the persisted `origin` metadata of an existing session entry.
This is especially visible for stable sessions such as `agent:kabot-core:main`: after a heartbeat is delivered into that session, the session still keeps the same `sessionKey`, but it... | closed | completed | true | 3 | [] | [] | 2026-03-14T12:48:44Z | 2026-04-27T04:43:31Z | 2026-04-25T01:09:46Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | kamiwang777 | 266,023,262 | U_kgDOD9sxXg | User | false |
openclaw/openclaw | 4,075,769,672 | I_kwDOQb6kR87y709I | 46,123 | https://github.com/openclaw/openclaw/issues/46123 | https://api.github.com/repos/openclaw/openclaw/issues/46123 | [Bug] chat.send with attachments causes RangeError: Maximum call stack size exceeded | ## Description
`chat.send` via WebSocket fails with `RangeError: Maximum call stack size exceeded` when the `attachments` parameter is present. This affects **all** WebSocket clients, including the built-in Control UI.
## Reproduction Steps
1. Connect to the gateway via WebSocket (any client: Control UI, custom app,... | closed | completed | true | 2 | [] | [] | 2026-03-14T12:55:42Z | 2026-04-27T04:43:28Z | 2026-04-25T01:10:03Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | villenero | 7,782,627 | MDQ6VXNlcjc3ODI2Mjc= | User | false |
openclaw/openclaw | 4,326,441,937 | I_kwDOQb6kR88AAAABAeBD0Q | 71,314 | https://github.com/openclaw/openclaw/issues/71314 | https://api.github.com/repos/openclaw/openclaw/issues/71314 | DeepSeek V4 thinking mode: reasoning_content not echoed back, breaking multi-turn conversations | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
When using DeepSeek V4 with thinking mode (reasoning: true), the first turn works fine, but all subsequent turns fail with HTTP 400. DeepSeek's API requires reasoning_content from previous assistant turns to be passed back in f... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-25T00:46:15Z | 2026-04-27T04:43:27Z | 2026-04-25T01:10:14Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | tryxin | 20,015,204 | MDQ6VXNlcjIwMDE1MjA0 | User | false |
openclaw/openclaw | 4,075,771,892 | I_kwDOQb6kR87y71f0 | 46,125 | https://github.com/openclaw/openclaw/issues/46125 | https://api.github.com/repos/openclaw/openclaw/issues/46125 | [Feature]: Please add streaming output support for WebChat (webchat) | ### Summary
Add streaming output support for WebChat to enable real-time response display.
### Problem to solve
Currently WebChat sends complete responses after generation. Users cannot see AI responses being generated in real-time, unlike Telegram/Discord which support streaming via `streaming: "partial"`.
### Pro... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-03-14T12:57:16Z | 2026-04-27T04:43:27Z | 2026-04-25T01:10:21Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | MrliCST | 190,002,436 | U_kgDOC1M1BA | User | false |
openclaw/openclaw | 4,075,772,936 | I_kwDOQb6kR87y71wI | 46,128 | https://github.com/openclaw/openclaw/issues/46128 | https://api.github.com/repos/openclaw/openclaw/issues/46128 | [Bug]: the dashboard always break, and need to reconnect or refresh many times | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When I ask a question with the dashboard, it begins to deel my request, even 3 seconds after, it shows me a big picture ( I do not know what is that) , and I cannot see the response information. then I need to refresh the session, and login... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-03-14T12:57:54Z | 2026-04-27T04:43:26Z | 2026-04-25T01:10:38Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | RichardQidian | 252,679,334 | U_kgDODw-Upg | User | false |
openclaw/openclaw | 4,075,775,442 | I_kwDOQb6kR87y72XS | 46,129 | https://github.com/openclaw/openclaw/issues/46129 | https://api.github.com/repos/openclaw/openclaw/issues/46129 | [Bug] Cron Tool Not Available - Plugin Not Enabled | ## Bug Description
The cron tool is not available in the current environment. Logs show the tool is not enabled.
## Error Evidence
From `/tmp/openclaw/openclaw-2026-03-14.log`:
```
tools.profile (coding) allowlist contains unknown entries (apply_patch, cron, image). These entries won't match any tool unless the plu... | closed | completed | true | 1 | [] | [] | 2026-03-14T12:59:33Z | 2026-04-27T04:43:25Z | 2026-04-25T01:10:55Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | alpha-evolver | 264,152,151 | U_kgDOD76kVw | User | false |
openclaw/openclaw | 4,175,763,192 | I_kwDOQb6kR8745Rb4 | 58,138 | https://github.com/openclaw/openclaw/issues/58138 | https://api.github.com/repos/openclaw/openclaw/issues/58138 | [Bug] Cron agentTurn jobs timeout even when script executes successfully | ## Bug Description
When using cron jobs with `agentTurn` payload, the job reports "job execution timed out" even though the underlying script executes successfully. The result is never delivered to the configured channel.
## Environment
- OpenClaw version: 2026.3.28
- macOS
- Provider: minimax/MiniMax-M2.7
## Steps... | closed | duplicate | true | 2 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-03-31T04:52:18Z | 2026-04-27T04:43:24Z | 2026-04-25T01:11:05Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | richarduncle | 19,268,769 | MDQ6VXNlcjE5MjY4NzY5 | User | false |
openclaw/openclaw | 4,193,197,070 | I_kwDOQb6kR8757xwO | 59,579 | https://github.com/openclaw/openclaw/issues/59579 | https://api.github.com/repos/openclaw/openclaw/issues/59579 | [Bug]: Scheduled cron runs timeout at ~60s in 2026.4.1 — manual triggers work fine (regression from 2026.3.31) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading from 2026.3.31 to 2026.4.1, all scheduled isolated agentTurn cron jobs timeout at ~60s with profileFailureReason: "timeout" at the Anthropic API call level; manually triggering the same jobs via cron run succ... | closed | duplicate | true | 5 | [
"bug",
"close:duplicate",
"dedupe:child",
"regression"
] | [] | 2026-04-02T09:21:52Z | 2026-04-27T04:43:23Z | 2026-04-25T01:11:09Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | trevenag | 18,563,066 | MDQ6VXNlcjE4NTYzMDY2 | User | false |
openclaw/openclaw | 4,156,998,481 | I_kwDOQb6kR873xsNR | 56,121 | https://github.com/openclaw/openclaw/issues/56121 | https://api.github.com/repos/openclaw/openclaw/issues/56121 | Cron sessions stuck in status=running after completion | ## Bug Description
Cron sessions remain in `status=running` with `endedAt=NONE` in `sessions.json` after the cron run has completed successfully (run log shows `finished/ok`).
## Evidence
- **17 sessions** stuck in `status=running`, oldest 19.9 hours
- The cron run log (`cron/runs/*.jsonl`) correctly records `finish... | closed | duplicate | true | 1 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-03-28T02:22:14Z | 2026-04-27T04:43:23Z | 2026-04-25T01:11:11Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | WaynezProg | 87,684,399 | MDQ6VXNlcjg3Njg0Mzk5 | User | false |
openclaw/openclaw | 4,075,805,257 | I_kwDOQb6kR87y79pJ | 46,145 | https://github.com/openclaw/openclaw/issues/46145 | https://api.github.com/repos/openclaw/openclaw/issues/46145 | [Bug] Control UI image upload not received by Agent sessions | ## Bug Description
When uploading an image in the OpenClaw Control UI (openclaw-control-ui), the image data is not being received by the Agent session. The user sees only the metadata header but no image content.
## Steps to Reproduce
1. Open OpenClaw Control UI at http://127.0.0.1:18789/
2. Try to upload/send an im... | closed | completed | true | 2 | [
"close:duplicate",
"close:already-fixed",
"dedupe:child"
] | [
"vincentkoc"
] | 2026-03-14T13:16:16Z | 2026-04-27T04:31:47Z | 2026-04-25T01:11:12Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | lizhuojunx86 | 45,158,108 | MDQ6VXNlcjQ1MTU4MTA4 | User | false |
openclaw/openclaw | 4,194,976,128 | I_kwDOQb6kR876CkGA | 59,720 | https://github.com/openclaw/openclaw/issues/59720 | https://api.github.com/repos/openclaw/openclaw/issues/59720 | No config path enables exec for agents in isolated sessions — sandbox blocks before exec-approvals | ## Summary
In OpenClaw 2026.3.31 (213a704), **there is no configuration path that enables `exec` for agents running in isolated sessions.** The sandbox policy blocks exec before `exec-approvals.json` is even consulted, and the `elevated` escape hatch doesn't work because isolated sessions report `channel: (unknown)`, ... | closed | completed | true | 3 | [] | [] | 2026-04-02T14:25:54Z | 2026-04-27T04:43:21Z | 2026-04-25T01:11:14Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | optimization-dave | 196,272,915 | U_kgDOC7LjEw | User | false |
openclaw/openclaw | 4,161,362,141 | I_kwDOQb6kR874CVjd | 56,572 | https://github.com/openclaw/openclaw/issues/56572 | https://api.github.com/repos/openclaw/openclaw/issues/56572 | Isolated cron run sessions never transition to 'done' in session store | ## Bug Description
Isolated cron run sessions (`sessionTarget: "isolated"`) are created in the agent session store with `status: "running"` but never transition to `"done"` when the run completes. The cron run log (`cron/runs/<jobId>.jsonl`) correctly records `status: "ok"`, but the corresponding session store entry (... | closed | duplicate | true | 1 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-03-28T19:10:13Z | 2026-04-27T04:43:22Z | 2026-04-25T01:11:14Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | r2dpoogehring | 268,780,991 | U_kgDOEAVFvw | User | false |
openclaw/openclaw | 4,075,810,810 | I_kwDOQb6kR87y7-_6 | 46,147 | https://github.com/openclaw/openclaw/issues/46147 | https://api.github.com/repos/openclaw/openclaw/issues/46147 | openclaw node run hangs silently on macOS after printing PATH, never connects to gateway | ## Bug: node host hangs silently on macOS, never sends WS handshake
### Steps to reproduce
1. Install openclaw via npm on macOS (Node 22.22.1)
2. Set up SSH forward tunnel: `ssh -f -N -L 18789:127.0.0.1:18789 root@<vps>`
3. Verify tunnel works: `curl http://127.0.0.1:18789/health` → `{"ok":true}`
4. Verify WS handsha... | closed | completed | true | 1 | [] | [] | 2026-03-14T13:18:38Z | 2026-04-27T04:43:20Z | 2026-04-25T01:11:30Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Mb29661 | 51,697,539 | MDQ6VXNlcjUxNjk3NTM5 | User | false |
openclaw/openclaw | 4,075,812,620 | I_kwDOQb6kR87y7_cM | 46,148 | https://github.com/openclaw/openclaw/issues/46148 | https://api.github.com/repos/openclaw/openclaw/issues/46148 | Control UI can break when scrolling history during streaming replies | ## Summary
Control UI can break and show a large exclamation-mark screen when the assistant is actively streaming a reply and the user scrolls upward through message history at the same time.
## Environment
- OpenClaw version: 2026.3.13 (observed locally)
- Surface: Control UI / webchat
- OS: Windows
- Model in use: g... | closed | completed | true | 1 | [] | [] | 2026-03-14T13:19:47Z | 2026-04-27T04:43:19Z | 2026-04-25T01:11:47Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | 2116217237 | 230,849,672 | U_kgDODcJ8iA | User | false |
openclaw/openclaw | 4,075,834,858 | I_kwDOQb6kR87y8E3q | 46,169 | https://github.com/openclaw/openclaw/issues/46169 | https://api.github.com/repos/openclaw/openclaw/issues/46169 | [Bug]: 2026.3.13 with docker-setup.sh build error | ### Bug type
Regression (worked before, now fails)
### Summary
I had use 2026.3.11 in docker well, and it's upgrade from 2026.3.8( Yes, I had a successful upgraded).But When I want upgrade it to 2026.3.13, it failed in `docker-setup.sh` progress.
### Steps to reproduce
1. Modify those environment:
HOME_VOLUME_NAM... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-03-14T13:33:29Z | 2026-04-27T04:43:18Z | 2026-04-25T01:12:22Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Chivenh | 28,721,089 | MDQ6VXNlcjI4NzIxMDg5 | User | false |
openclaw/openclaw | 4,326,446,608 | I_kwDOQb6kR88AAAABAeBWEA | 71,316 | https://github.com/openclaw/openclaw/issues/71316 | https://api.github.com/repos/openclaw/openclaw/issues/71316 | [Feature]: adaptive thinking level should map to "dynamic thinking" of gemini thinkingBudget=-1 | ### Summary
Map OpenClaw's 'adaptive' thinking level to the Gemini API's "dynamic thinking" mode by setting `thinkingBudget: -1`.
### Problem to solve
Currently, OpenClaw's `adaptive` thinking level does not leverage the native dynamic resource allocation feature of Gemini models. This means users must manually swi... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-25T00:47:36Z | 2026-04-27T04:43:17Z | 2026-04-25T01:12:27Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | shichuzhu | 6,067,848 | MDQ6VXNlcjYwNjc4NDg= | User | false |
openclaw/openclaw | 4,075,877,421 | I_kwDOQb6kR87y8PQt | 46,201 | https://github.com/openclaw/openclaw/issues/46201 | https://api.github.com/repos/openclaw/openclaw/issues/46201 | [Bug]: agents 页面，有大于一个模型时，更换模型后无法保存 | ### Bug type
Regression (worked before, now fails)
### Summary
agents 页面
### Steps to reproduce
agents 页面 & 大于一个agents
下拉选中其它大模型，
点save
选择还原到更换前的模型
保存无效，实际使用也仍然是进页面时选中的模型
### Expected behavior
期待save后保存选中的模型。
### Actual behavior
实际save后仍然是打开页面时的模型，修改没有生效。
### OpenClaw version
版本 2026.3.13
### Operating syst... | closed | not_planned | true | 3 | [
"bug",
"regression"
] | [] | 2026-03-14T13:59:50Z | 2026-04-27T04:43:17Z | 2026-04-25T01:12:40Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | wattshen | 11,402,816 | MDQ6VXNlcjExNDAyODE2 | User | false |
openclaw/openclaw | 4,075,882,958 | I_kwDOQb6kR87y8QnO | 46,207 | https://github.com/openclaw/openclaw/issues/46207 | https://api.github.com/repos/openclaw/openclaw/issues/46207 | [Bug]: [UI Bug] 聊天框出现无法消除的警示图标，只能通过删除 DOM 元素消除 | ### Bug type
Regression (worked before, now fails)
### Summary
## 问题描述
在 OpenClaw Web UI 聊天框中，有时会出现一个警示图标（感叹号 SVG），刷新页面和重启应用后仍然显示。
## 复现步骤
1. 打开 OpenClaw 控制台
2. 进行某些操作后（暂不明确具体触发条件）
3. 聊天的输入框上方出现警示图标
## 期望行为
图标应该正常消失，或者显示明确的错误信息
## 实际行为
图标一直存在，刷新和重启都无法消除，只能通过浏览器开发者工具删除 DOM 元素才能消除
## 截图/信息
图标 SVG 代码：
<svg class="c... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-03-14T14:03:07Z | 2026-04-27T04:43:16Z | 2026-04-25T01:12:57Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | g932826739 | 267,974,118 | U_kgDOD_j15g | User | false |
openclaw/openclaw | 4,075,897,133 | I_kwDOQb6kR87y8UEt | 46,218 | https://github.com/openclaw/openclaw/issues/46218 | https://api.github.com/repos/openclaw/openclaw/issues/46218 | [Bug] CLI-mode gateway RPC client handshake timeout in v2026.3.13 — connect.challenge never processed | ## Environment
- **OpenClaw:** 2026.3.13 (61d171a)
- **Node:** v22.22.0
- **OS:** Debian 12 (bookworm), x86_64
- **Install:** npm global (`npm i -g openclaw`)
- **Gateway:** systemd user service, loopback bind
## Problem
After upgrading from v2026.3.12 → v2026.3.13, **all CLI commands using the CLI-mode gateway RPC ... | closed | completed | true | 3 | [] | [] | 2026-03-14T14:10:03Z | 2026-04-27T04:43:15Z | 2026-04-25T01:13:14Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | MSCHKY | 177,659,116 | U_kgDOCpbc7A | User | false |
openclaw/openclaw | 4,075,902,322 | I_kwDOQb6kR87y8VVy | 46,219 | https://github.com/openclaw/openclaw/issues/46219 | https://api.github.com/repos/openclaw/openclaw/issues/46219 | 🔥OpenClaw自我进化插件视频教程 | ### Summary
🔥视频教程：https://youtu.be/bhuGrjuCM_g
🔥插件：https://github.com/CortexReach/memory-lancedb-pro
🔥skill：https://github.com/CortexReach/memory-lancedb-pro-skill
### Problem to solve
智能提取记忆
### Proposed solution
智能提取记忆
### Alternatives considered
_No response_
### Impact
智能提取记忆
### Evidence/examples
_... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-14T14:13:01Z | 2026-04-27T04:43:14Z | 2026-04-25T01:13:32Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | stoeng | 129,583,294 | U_kgDOB7lIvg | User | false |
openclaw/openclaw | 4,075,909,010 | I_kwDOQb6kR87y8W-S | 46,223 | https://github.com/openclaw/openclaw/issues/46223 | https://api.github.com/repos/openclaw/openclaw/issues/46223 | [Feature Request] GitHub Webhook Integration for Real-time Agent Task Triggering | ---
title: "[Feature Request] GitHub Webhook Integration for Real-time Agent Task Triggering"
labels: ["enhancement", "discussion"]
assignees: []
---
## 🎯 Problem Statement
Currently, OpenClaw requires polling-based approaches (e.g., `/gh-issues --watch`) to detect new GitHub issues/comments. This introduces:
- **La... | closed | not_planned | true | 1 | [] | [] | 2026-03-14T14:16:42Z | 2026-04-27T04:43:13Z | 2026-04-25T01:13:49Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | tianxiao1430-jpg | 255,869,001 | U_kgDOD0BASQ | User | false |
openclaw/openclaw | 4,075,911,645 | I_kwDOQb6kR87y8Xnd | 46,225 | https://github.com/openclaw/openclaw/issues/46225 | https://api.github.com/repos/openclaw/openclaw/issues/46225 | [Bug]: OpenClaw 子 Agent Workspace 隔离问题报告 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
OpenClaw 子 Agent Workspace 隔离问题报告
### Steps to reproduce
好，修改报告，隐匿个人信息：---
## OpenClaw 子 Agent Workspace 隔离问题报告
### 问题现象
**配置与行为不符：**
在 `openclaw.json` 中配置了子 Agent 的独立 workspace：
```json5
{
agents: {
list: [
{
id: ... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-03-14T14:18:27Z | 2026-04-27T04:43:12Z | 2026-04-25T01:14:06Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | woaw | 32,876,709 | MDQ6VXNlcjMyODc2NzA5 | User | false |
openclaw/openclaw | 4,075,912,954 | I_kwDOQb6kR87y8X76 | 46,226 | https://github.com/openclaw/openclaw/issues/46226 | https://api.github.com/repos/openclaw/openclaw/issues/46226 | [Bug]: Gateway probe shows 3000ms budget but uses 800ms internally — false timeout on healthy local loopback | ## Bug Description
The <code>openclaw gateway probe</code> command displays <code>Probe budget: 3000ms</code> to the user, but internally applies <code>Math.min(800, overallMs)</code> for local loopback connections. This creates a misleading mismatch where:
1. The CLI shows a 3000ms budget
2. The actual connection ti... | closed | completed | true | 1 | [] | [] | 2026-03-14T14:19:19Z | 2026-04-27T04:43:11Z | 2026-04-25T01:14:23Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Mdx2025 | 181,967,107 | U_kgDOCtiZAw | User | false |
openclaw/openclaw | 4,075,937,460 | I_kwDOQb6kR87y8d60 | 46,237 | https://github.com/openclaw/openclaw/issues/46237 | https://api.github.com/repos/openclaw/openclaw/issues/46237 | Telegram: stale old assistant/subagent completion results can resurface later in-thread as if current | ## Summary
In Telegram topic `12088`, old **already-completed assistant/subagent result messages** can resurface later in the same thread as if they were current.
This is **not just duplicate inbound Telegram updates**. The symptom here is that an old internal completion/result message appears again later, out of tim... | closed | completed | true | 1 | [] | [] | 2026-03-14T14:34:06Z | 2026-04-27T04:43:11Z | 2026-04-25T01:14:40Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Jackten | 2,895,479 | MDQ6VXNlcjI4OTU0Nzk= | User | false |
openclaw/openclaw | 4,075,940,487 | I_kwDOQb6kR87y8eqH | 46,240 | https://github.com/openclaw/openclaw/issues/46240 | https://api.github.com/repos/openclaw/openclaw/issues/46240 | Control UI / webchat no longer renders MEDIA replies as attachments; message tool also rejects channel=webchat | ## Summary
In Control UI / webchat, assistant replies containing `MEDIA:<local-path>` no longer render as attachments.
Instead, the UI shows broken text behavior:
- the reply does not produce an attachment bubble
- the `MEDIA:` directive is partially consumed
- in one repro, only `MEDIA` was displayed in the chat tra... | closed | completed | true | 3 | [] | [] | 2026-03-14T14:35:59Z | 2026-04-27T04:43:09Z | 2026-04-25T01:14:58Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Teintes-BI | 73,472,314 | MDQ6VXNlcjczNDcyMzE0 | User | false |
openclaw/openclaw | 4,075,946,926 | I_kwDOQb6kR87y8gOu | 46,241 | https://github.com/openclaw/openclaw/issues/46241 | https://api.github.com/repos/openclaw/openclaw/issues/46241 | [Feature Request] 支持 WebChat 自定义用户头像 | ### Summary
目前 OpenClaw 只能配置助手（Yuni）的头像，但 WebChat 界面右侧「用户头像」无法自定义，只能显示默认灰色 placeholder。
### Problem to solve
Users want to see their own profile photo/avatar instead of the default placeholder on the right side of the chat interface.
### Proposed solution
Allow configuring user avatar in openclaw.json (e.g., unde... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-03-14T14:40:03Z | 2026-04-27T04:43:08Z | 2026-04-25T01:15:15Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | liyinkeji-tech | 262,088,625 | U_kgDOD58nsQ | User | false |
openclaw/openclaw | 4,075,947,429 | I_kwDOQb6kR87y8gWl | 46,242 | https://github.com/openclaw/openclaw/issues/46242 | https://api.github.com/repos/openclaw/openclaw/issues/46242 | [Feature]: Passkey (WebAuthn/FIDO2) support for Control UI authentication | ### Summary
Add Passkey (WebAuthn/FIDO2) as an authentication method for the Control UI, enabling Face ID / Touch ID login on mobile devices.
### Problem to solve
The Control UI currently only supports token-based authentication. On mobile (iOS/Android), users must manually type a long token string every time they ... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-03-14T14:40:23Z | 2026-04-27T04:43:08Z | 2026-04-25T01:15:32Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Anoyou | 13,214,236 | MDQ6VXNlcjEzMjE0MjM2 | User | false |
openclaw/openclaw | 4,075,959,668 | I_kwDOQb6kR87y8jV0 | 46,250 | https://github.com/openclaw/openclaw/issues/46250 | https://api.github.com/repos/openclaw/openclaw/issues/46250 | [Bug][CRITICAL] Session transcript lost after session reset - 6 hours of conversation history missing | ### Bug type
Crash (process/app exits or hangs)
### Summary
Session transcript persistence fails after session reset: 6 hours of conversation history lost despite successful skill file creation, indicating a critical race condition in session management.
On March 13, 2026, a 6-hour conversation session was lost aft... | closed | completed | true | 2 | [
"bug",
"bug:crash"
] | [] | 2026-03-14T14:47:40Z | 2026-04-27T04:43:07Z | 2026-04-25T01:15:50Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | lyemmaly-sudo | 268,187,537 | U_kgDOD_w3kQ | User | false |
openclaw/openclaw | 4,075,969,479 | I_kwDOQb6kR87y8lvH | 46,254 | https://github.com/openclaw/openclaw/issues/46254 | https://api.github.com/repos/openclaw/openclaw/issues/46254 | [Bug] Subagent Docker containers not auto-removed after completion, causing maxConcurrent slot exhaustion | ## Bug Description
Subagent Docker containers are not automatically cleaned up after task completion, causing them to accumulate until maxConcurrent limit is reached and blocking new subagent spawns.
### Observed Behavior
- Subagent containers continue running after the task is complete
- docker ps shows multiple op... | closed | not_planned | true | 3 | [] | [] | 2026-03-14T14:53:32Z | 2026-04-27T04:43:06Z | 2026-04-25T01:16:07Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | 2670044605 | 70,048,042 | MDQ6VXNlcjcwMDQ4MDQy | User | false |
openclaw/openclaw | 4,075,970,587 | I_kwDOQb6kR87y8mAb | 46,256 | https://github.com/openclaw/openclaw/issues/46256 | https://api.github.com/repos/openclaw/openclaw/issues/46256 | [Bug]: openclaw browser start/status fails with handshake timeout on macOS when plugin loading is slow | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Description:
Running openclaw browser start or openclaw browser status consistently fails with:
Error: gateway closed (1000 normal closure): no close reason
Gateway target: ws://127.0.0.1:18789
The gateway itself is healthy (openclaw gate... | closed | completed | true | 6 | [
"bug",
"bug:behavior"
] | [] | 2026-03-14T14:54:07Z | 2026-04-27T04:43:05Z | 2026-04-25T01:16:24Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | lisphi | 1,599,477 | MDQ6VXNlcjE1OTk0Nzc= | User | false |
openclaw/openclaw | 4,075,972,823 | I_kwDOQb6kR87y8mjX | 46,257 | https://github.com/openclaw/openclaw/issues/46257 | https://api.github.com/repos/openclaw/openclaw/issues/46257 | Sandboxed sessions receive inaccessible host skill paths in <available_skills> (read fails with ‘Path escapes sandbox root’) | ## Summary
In sandboxed sessions, OpenClaw exposes skills in `<available_skills>` using host filesystem paths, but the file tools inside the sandbox cannot read those paths.
This creates a mismatch between:
- what the agent is instructed to do (“read the selected `SKILL.md` before replying”), and
- what the sandboxed... | closed | completed | true | 1 | [] | [] | 2026-03-14T14:55:31Z | 2026-04-27T04:43:04Z | 2026-04-25T01:16:41Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | GitHoubi | 99,397,913 | U_kgDOBeyxGQ | User | false |
openclaw/openclaw | 4,075,980,148 | I_kwDOQb6kR87y8oV0 | 46,261 | https://github.com/openclaw/openclaw/issues/46261 | https://api.github.com/repos/openclaw/openclaw/issues/46261 | [Feature]: Integrate pi-multi-pass | ### Summary
There's currently no way to configure multiple OAuth accounts per provider. [pi-multi-pass](https://github.com/hjanuschka/pi-multi-pass) solves this elegantly. Could it be integrated into OpenClaw?
### Problem to solve
When teams need to operate multiple identities under the same OAuth provider (e.g., se... | closed | completed | true | 5 | [
"enhancement"
] | [] | 2026-03-14T14:59:37Z | 2026-04-27T04:43:04Z | 2026-04-25T01:16:59Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | AgatElite | 191,369,180 | U_kgDOC2gP3A | User | false |
openclaw/openclaw | 4,075,992,698 | I_kwDOQb6kR87y8rZ6 | 46,267 | https://github.com/openclaw/openclaw/issues/46267 | https://api.github.com/repos/openclaw/openclaw/issues/46267 | [Bug] Secondary Effect of Account Hijack - Multiple agents respond to single @mention | ## Bug Report - Secondary Issue from Account Hijack
**Submitted by:** OpenClaw Agent (auto-submitted)
**Related Issue:** #45766 (Matrix Message Tool - accountId not working & mention bugs)
---
### Problem Description
This is a **secondary bug** that occurs as a consequence of the original "Account Hijack" bug (#45... | closed | completed | true | 1 | [] | [] | 2026-03-14T15:06:13Z | 2026-04-27T04:43:03Z | 2026-04-25T01:17:16Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | aaavieri | 13,959,552 | MDQ6VXNlcjEzOTU5NTUy | User | false |
openclaw/openclaw | 4,075,995,231 | I_kwDOQb6kR87y8sBf | 46,271 | https://github.com/openclaw/openclaw/issues/46271 | https://api.github.com/repos/openclaw/openclaw/issues/46271 | LiteLLM custom provider times out even though direct /v1/chat/completions succeeds | ## Summary
I’m configuring OpenClaw to use a LiteLLM endpoint as a custom/OpenAI-compatible provider. OpenClaw successfully recognizes the provider and model, but all real agent requests time out with `LLM request timed out`, even though direct HTTPS calls to the same LiteLLM endpoint and model succeed immediately fro... | closed | not_planned | true | 2 | [] | [] | 2026-03-14T15:07:33Z | 2026-04-27T04:43:02Z | 2026-04-25T01:17:33Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | db2guru | 7,860,065 | MDQ6VXNlcjc4NjAwNjU= | User | false |
openclaw/openclaw | 4,076,007,570 | I_kwDOQb6kR87y8vCS | 46,274 | https://github.com/openclaw/openclaw/issues/46274 | https://api.github.com/repos/openclaw/openclaw/issues/46274 | [Feature]: Support WeChat Official Account (微信公众号) Integration | ### Summary
## Feature Description
Add support for WeChat Official Account (微信公众号) integration in OpenClaw
## Use Case
- Publish articles to WeChat Official Account
- Auto-post content to WeChat
- WeChat message handling
## Why Needed
WeChat Official Account is widely used in China for content publishing. Many use... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-14T15:14:52Z | 2026-04-27T04:43:01Z | 2026-04-25T01:17:50Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | 94411449-eng | 266,421,843 | U_kgDOD-FGUw | User | false |
openclaw/openclaw | 4,076,014,940 | I_kwDOQb6kR87y8w1c | 46,277 | https://github.com/openclaw/openclaw/issues/46277 | https://api.github.com/repos/openclaw/openclaw/issues/46277 | [Bug] Browser tool causes gateway to crash in v2026.3.13 | ## Bug Description
In OpenClaw v2026.3.13, any `openclaw browser` command causes the gateway to crash with error: `Error: gateway closed (1000 normal closure): no close reason`
## Steps to Reproduce
1. Run `openclaw gateway start`
2. Start Chrome with remote debugging: `google-chrome --remote-debugging-port=9222`
3.... | closed | completed | true | 1 | [] | [] | 2026-03-14T15:19:18Z | 2026-04-27T04:43:01Z | 2026-04-25T01:18:07Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | KangShuaiFu | 19,404,095 | MDQ6VXNlcjE5NDA0MDk1 | User | false |
openclaw/openclaw | 4,076,024,843 | I_kwDOQb6kR87y8zQL | 46,282 | https://github.com/openclaw/openclaw/issues/46282 | https://api.github.com/repos/openclaw/openclaw/issues/46282 | [Bug]: Telegram media download uses native fetch instead of proxy-aware fetch | ### Bug type
Regression (worked before, now fails)
### Summary
File: auth-profiles-Cm29uFmx.js line ~18114
Impact: sendPhoto and inbound media download fail when proxy is required
### Steps to reproduce
sendPhoto and inbound media download fail when proxy is required
### Expected behavior
sendPhoto and inbound m... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-03-14T15:25:10Z | 2026-04-27T04:43:00Z | 2026-04-25T01:18:24Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | xiaoboan | 80,965,801 | MDQ6VXNlcjgwOTY1ODAx | User | false |
openclaw/openclaw | 4,076,038,572 | I_kwDOQb6kR87y82ms | 46,289 | https://github.com/openclaw/openclaw/issues/46289 | https://api.github.com/repos/openclaw/openclaw/issues/46289 | [Bug]: Duplicate message output in Feishu channel | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Subject: Bug Report: Duplicate message output in Feishu channel
Hi OpenClaw Team,
I'm experiencing a bug where messages are duplicated when sending replies through the Feishu channel.
**Environment:**
- OpenClaw version: 2026.3.8
- OS: W... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-03-14T15:32:03Z | 2026-04-27T04:42:59Z | 2026-04-25T01:18:41Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | fyp1234561 | 266,465,648 | U_kgDOD-HxcA | User | false |
openclaw/openclaw | 4,076,067,897 | I_kwDOQb6kR87y89w5 | 46,297 | https://github.com/openclaw/openclaw/issues/46297 | https://api.github.com/repos/openclaw/openclaw/issues/46297 | Feature: Suppress typing indicator when message matches hold pattern | ## Feature Request
### Summary
When an agent silently holds on a message (returns NO_REPLY), the typing indicator still briefly appears and disappears on the user's side. This leaks the fact that the agent processed the message.
### Use Case
We use a convention where messages ending with `...` (optionally preceded by... | closed | completed | true | 2 | [] | [] | 2026-03-14T15:43:51Z | 2026-04-27T04:42:58Z | 2026-04-25T01:18:58Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | cam-ross | 239,492,553 | U_kgDODkZdyQ | User | false |
openclaw/openclaw | 4,076,072,038 | I_kwDOQb6kR87y8-xm | 46,298 | https://github.com/openclaw/openclaw/issues/46298 | https://api.github.com/repos/openclaw/openclaw/issues/46298 | sessions_yield terminates isolated cron sessions instead of waiting for next turn | ### Bug type
Regression (worked before, now fails)
### Summary
`sessions_yield` (introduced in v2026.3.12 via PR #36537) immediately terminates isolated cron sessions instead of pausing and waiting for a follow-up event (like a sub-agent completion). The description says it should "end the current turn and carry a f... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-03-14T15:45:32Z | 2026-04-27T04:42:58Z | 2026-04-25T01:19:16Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | lnxg33k | 1,170,490 | MDQ6VXNlcjExNzA0OTA= | User | false |
openclaw/openclaw | 4,076,085,238 | I_kwDOQb6kR87y9B_2 | 46,302 | https://github.com/openclaw/openclaw/issues/46302 | https://api.github.com/repos/openclaw/openclaw/issues/46302 | Feature request: expose registerInternalHook / clearBootstrapSnapshot via plugin SDK for bootstrap hot-reload | ## Problem
OpenClaw caches bootstrap files (AGENTS.md, SOUL.md, etc.) in memory at gateway startup. Once cached, edits to these files don't take effect until the user runs `/reset` or `openclaw gateway restart`. This creates friction for active development and configuration iteration.
There is currently no way for a ... | closed | completed | true | 1 | [] | [] | 2026-03-14T15:50:57Z | 2026-04-27T04:42:57Z | 2026-04-25T01:19:33Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | hchen13 | 3,998,350 | MDQ6VXNlcjM5OTgzNTA= | User | false |
openclaw/openclaw | 4,079,093,791 | I_kwDOQb6kR87zIggf | 47,608 | https://github.com/openclaw/openclaw/issues/47608 | https://api.github.com/repos/openclaw/openclaw/issues/47608 | cron: script payload timeoutSeconds not enforced | ## Bug
`CronScriptPayload.timeoutSeconds` is defined in the type but never applied — script jobs always run with the `DEFAULT_JOB_TIMEOUT_MS` (10 min) ceiling regardless of what `timeoutSeconds` is set to.
## Root cause
`resolveCronJobTimeoutMs` in `src/cron/service/timeout-policy.ts` only reads `timeoutSeconds` for... | closed | duplicate | true | 2 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-03-15T20:05:29Z | 2026-04-27T04:42:56Z | 2026-04-25T01:19:46Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Arry8 | 263,019,747 | U_kgDOD61c4w | User | false |
openclaw/openclaw | 4,076,116,708 | I_kwDOQb6kR87y9Jrk | 46,310 | https://github.com/openclaw/openclaw/issues/46310 | https://api.github.com/repos/openclaw/openclaw/issues/46310 | config.patch unconditionally schedules SIGUSR1 restart, ignoring hot-reload capability | ## `config.patch` unconditionally schedules SIGUSR1 restart, ignoring hot-reload capability
**Version:** OpenClaw 2026.3.8 (3caab92)
**Platform:** macOS (LaunchAgent), Node 22
### Summary
The `config.patch` WS API handler unconditionally calls `scheduleGatewaySigusr1Restart()` after writing the config file, even whe... | closed | completed | true | 2 | [] | [] | 2026-03-14T16:04:44Z | 2026-04-27T04:42:55Z | 2026-04-25T01:19:50Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | 0xHubert | 260,107,275 | U_kgDOD4DsCw | User | false |
openclaw/openclaw | 4,076,124,575 | I_kwDOQb6kR87y9Lmf | 46,316 | https://github.com/openclaw/openclaw/issues/46316 | https://api.github.com/repos/openclaw/openclaw/issues/46316 | [Bug]: `devices list` / `nodes status` timeout while `gateway status` shows `RPC probe: ok` (regression in 2026.3.12/2026.3.13) | ### Bug type
Regression (worked before, now fails)
### Summary
On 2026.3.12 and 2026.3.13, `openclaw gateway status` reports `RPC probe: ok` to `ws://127.0.0.1:18789`, but `openclaw devices list` and `openclaw nodes status` intermittently timeout against the same endpoint.
### Steps to reproduce
1. Start gateway n... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-03-14T16:08:07Z | 2026-04-27T04:42:52Z | 2026-04-25T01:20:07Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | dmmmuse | 58,375,502 | MDQ6VXNlcjU4Mzc1NTAy | User | false |
openclaw/openclaw | 4,076,132,313 | I_kwDOQb6kR87y9NfZ | 46,317 | https://github.com/openclaw/openclaw/issues/46317 | https://api.github.com/repos/openclaw/openclaw/issues/46317 | [Feature]: Pre prompt Rag Hook | ### Summary
Feature Request: Pre-prompt RAG hook — auto-inject relevant context before agent response
### Problem to solve
**Problem:** Agents have search tools (memory_search, QMD) but using them is model-discretionary. Sessions frequently skip searching and respond with zero context. The knowledge exists — the... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-03-14T16:10:53Z | 2026-04-27T04:42:49Z | 2026-04-25T01:20:25Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | mel0107 | 7,932,681 | MDQ6VXNlcjc5MzI2ODE= | User | false |
openclaw/openclaw | 4,182,714,071 | I_kwDOQb6kR875TybX | 58,585 | https://github.com/openclaw/openclaw/issues/58585 | https://api.github.com/repos/openclaw/openclaw/issues/58585 | LiveSessionModelSwitchError persists across isolated cron sessions | ## Bug
Isolated cron sessions (`sessionTarget: "isolated"`) should start with a clean model state on every run, but they inherit stale model state from `sessions.json`. This causes `LiveSessionModelSwitchError` when:
1. A cron run falls back to a different model (e.g., Gemini when Anthropic 429s)
2. The fallback mode... | closed | duplicate | true | 2 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-03-31T22:17:42Z | 2026-04-27T04:42:48Z | 2026-04-25T01:20:27Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | biz-biz-Bot | 269,882,212 | U_kgDOEBYTZA | User | false |
openclaw/openclaw | 4,220,589,945 | I_kwDOQb6kR877kRd5 | 62,707 | https://github.com/openclaw/openclaw/issues/62707 | https://api.github.com/repos/openclaw/openclaw/issues/62707 | [Bug]: resolveCronSession leaks cliSessionIds on forceNew — isolated runs carry stale CLI session identity | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
`resolveCronSession` in `src/cron/isolated-agent/session.ts` preserves `cliSessionIds` and `claudeCliSessionId` via `...entry` spread when `isNewSession` is true, meaning isolated cron run session entries conta... | closed | duplicate | true | 1 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-04-07T21:05:14Z | 2026-04-27T04:42:47Z | 2026-04-25T01:20:31Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | alexey-pelykh | 345,102 | MDQ6VXNlcjM0NTEwMg== | User | false |
openclaw/openclaw | 4,175,327,448 | I_kwDOQb6kR8743nDY | 58,083 | https://github.com/openclaw/openclaw/issues/58083 | https://api.github.com/repos/openclaw/openclaw/issues/58083 | [Bug] Cron isolated jobs created from chat sessions may persist current sessionKey and lose true isolation | ## Summary
When creating a cron job from a Discord/chat session with `sessionTarget: "isolated"`, OpenClaw appears to persist the **current chat sessionKey** onto the job even when `sessionKey` was not explicitly provided.
That makes the job run against the chat-bound session instead of the expected isolated `cron:<j... | closed | duplicate | true | 2 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-03-31T02:56:52Z | 2026-04-27T04:42:47Z | 2026-04-25T01:20:35Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | jrsdhr | 4,820,276 | MDQ6VXNlcjQ4MjAyNzY= | User | false |
openclaw/openclaw | 4,093,339,381 | I_kwDOQb6kR87z-2b1 | 49,578 | https://github.com/openclaw/openclaw/issues/49578 | https://api.github.com/repos/openclaw/openclaw/issues/49578 | [Bug]: Cron jobs can contaminate and retain persisted session state | ## Summary
Cron jobs can leave behind unsafe persisted session state in two ways:
1. multiple cron jobs can share one long-lived session and contaminate context
2. disabling/removing cron jobs does not reliably clean up their persisted sessions/transcripts
## Environment
- OpenClaw 2026.3.13
- macOS 26.3.1
- local gat... | closed | duplicate | true | 2 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-03-18T07:38:13Z | 2026-04-27T04:42:46Z | 2026-04-25T01:20:40Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | liaoandi | 31,417,269 | MDQ6VXNlcjMxNDE3MjY5 | User | false |
openclaw/openclaw | 4,076,149,987 | I_kwDOQb6kR87y9Rzj | 46,319 | https://github.com/openclaw/openclaw/issues/46319 | https://api.github.com/repos/openclaw/openclaw/issues/46319 | Context pruning can split tool_use/tool_result pairs, causing API validation errors | ## Bug
Context pruning can split `tool_use`/`tool_result` message pairs, causing Anthropic API validation errors.
## Root Cause Analysis
After tracing the code in `pi-embedded-runner/run/attempt.js` (lines ~431–447), the message pipeline runs in this order:
1. `sanitizeSessionHistory()` — calls `repairToolUseResult... | closed | completed | true | 1 | [] | [] | 2026-03-14T16:16:55Z | 2026-04-27T04:42:45Z | 2026-04-25T01:20:42Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | hazeydata | 250,902,262 | U_kgDODvR29g | User | false |
openclaw/openclaw | 4,121,767,285 | I_kwDOQb6kR871rS11 | 52,975 | https://github.com/openclaw/openclaw/issues/52975 | https://api.github.com/repos/openclaw/openclaw/issues/52975 | [Bug]: The sessionTarget method of using a custom seesinID in a cron scheduled task does not work | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Edit the jobs.json file directly and configure cron jobs as follows.
"sessionTarget:"session:agent:main:direct:main",
"wakeMode": "now",
"payload": {
"kind": "agentTurn",
message: "xxxxxxx"
}
The session:agent:main:direct:main is an existin... | closed | duplicate | true | 4 | [
"bug",
"close:duplicate",
"dedupe:child",
"bug:behavior"
] | [] | 2026-03-23T15:31:38Z | 2026-04-27T04:42:44Z | 2026-04-25T01:20:44Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | senlinzhiyan | 35,173,242 | MDQ6VXNlcjM1MTczMjQy | User | false |
openclaw/openclaw | 4,177,968,173 | I_kwDOQb6kR875Brwt | 58,304 | https://github.com/openclaw/openclaw/issues/58304 | https://api.github.com/repos/openclaw/openclaw/issues/58304 | [Bug]: Cron job with non-isolated sessionTarget causes sessionId/sessionFile mismatch | ### Summary
When a cron job is configured with `sessionTarget: "session:xxx"}` (non-isolated mode, reusing an existing session), it incorrectly updates the sessionId of the linked session entry, causing sessionId and sessionFile to point to different files.
### Steps to reproduce
1. Configure a cron job with `sessio... | closed | duplicate | true | 1 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-03-31T10:27:32Z | 2026-04-27T04:42:43Z | 2026-04-25T01:20:54Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Kerberos255 | 97,336,831 | U_kgDOBc09_w | User | false |
openclaw/openclaw | 4,111,951,073 | I_kwDOQb6kR871F2Th | 51,530 | https://github.com/openclaw/openclaw/issues/51530 | https://api.github.com/repos/openclaw/openclaw/issues/51530 | [Bug]: Cron job agentTurn model override to Google/Gemini failing, defaulting to Anthropic/Claude | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Hi team,
I'm experiencing an issue with cron jobs running on OpenClaw v2026.3.13.
Problem:
When setting payload.kind: "agentTurn" jobs with model: "google/gemini-2.5-flash" in the payload, the jobs consistently execute using anthropic/cla... | closed | duplicate | true | 3 | [
"bug",
"close:duplicate",
"dedupe:child",
"bug:behavior"
] | [] | 2026-03-21T09:42:04Z | 2026-04-27T04:42:43Z | 2026-04-25T01:20:58Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | airellang | 269,267,157 | U_kgDOEAyw1Q | User | false |
openclaw/openclaw | 4,076,169,056 | I_kwDOQb6kR87y9Wdg | 46,326 | https://github.com/openclaw/openclaw/issues/46326 | https://api.github.com/repos/openclaw/openclaw/issues/46326 | Audio transcription stopped working after update to 2026.3.12/3.13 | ## Description
Audio transcription for inbound voice messages (Signal) stopped working after updating from a pre-2026.3.12 version to 2026.3.12, and persists on 2026.3.13.
## Environment
- **OS:** macOS (Apple Silicon, Mac mini M4)
- **Version:** 2026.3.13 (also broken on 2026.3.12)
- **Previous version:** Was workin... | closed | completed | true | 1 | [] | [] | 2026-03-14T16:23:23Z | 2026-04-27T04:42:42Z | 2026-04-25T01:20:59Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | daliboy | 81,185,334 | MDQ6VXNlcjgxMTg1MzM0 | User | false |
openclaw/openclaw | 4,207,987,834 | I_kwDOQb6kR8760Mx6 | 61,414 | https://github.com/openclaw/openclaw/issues/61414 | https://api.github.com/repos/openclaw/openclaw/issues/61414 | [Bug]: Cron delivery ignores explicit delivery.channel when sessionTarget=isolated | # Bug Report for OpenClaw
**Bug type:** Behavior bug (incorrect output/state without crash)
**Beta release blocker:** No
---
## Summary
Cron jobs with explicit `delivery.channel` setting (e.g., `"webchat"`) deliver to wrong channel when `sessionTarget="isolated"`. The configured channel is ignored and delivery fal... | closed | duplicate | true | 1 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-04-05T16:08:29Z | 2026-04-27T04:42:41Z | 2026-04-25T01:21:03Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | big-bear-dweller | 271,713,466 | U_kgDOEDIEug | User | false |
openclaw/openclaw | 4,166,418,028 | I_kwDOQb6kR874Vn5s | 57,392 | https://github.com/openclaw/openclaw/issues/57392 | https://api.github.com/repos/openclaw/openclaw/issues/57392 | Bug: cron delivery with `channel: "last"` silently fails in isolated sessions | ## Summary
Cron jobs configured with `delivery.channel: "last"` silently fail to deliver results when running in isolated sessions. No error is logged, no fallback is attempted — the cron job completes successfully but the output is never delivered to the user.
## Environment
- OpenClaw: 2026.3.24 / 2026.3.28
- OS: ... | closed | duplicate | true | 1 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-03-30T01:39:53Z | 2026-04-27T04:42:40Z | 2026-04-25T01:21:07Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | zeeyuu222 | 267,768,244 | U_kgDOD_XRtA | User | false |
openclaw/openclaw | 4,221,577,114 | I_kwDOQb6kR877oCea | 62,830 | https://github.com/openclaw/openclaw/issues/62830 | https://api.github.com/repos/openclaw/openclaw/issues/62830 | Cron isolated session exec tool result not delivered to parent model after upgrade to 2026.4.5 | ## Bug Description
After upgrading from 2026.3.28 to 2026.4.5, cron jobs using isolated session + agentTurn payload fail consistently with timeout errors, even though the underlying script executes successfully.
## Steps to Reproduce
1. Set up a cron job with sessionTarget: \"isolated\" and agentTurn payload that ru... | closed | duplicate | true | 2 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-04-08T01:53:59Z | 2026-04-27T04:42:40Z | 2026-04-25T01:21:12Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | JudgeAllen | 4,287,323 | MDQ6VXNlcjQyODczMjM= | User | false |
openclaw/openclaw | 4,076,169,097 | I_kwDOQb6kR87y9WeJ | 46,327 | https://github.com/openclaw/openclaw/issues/46327 | https://api.github.com/repos/openclaw/openclaw/issues/46327 | [Bug]: Gateway event loop freezes on 2026.3.13 — Slack socket mode events not dispatched (Linux/amd64, 2GB VPS) | ### Bug type
Regression (worked before, now fails)
### Summary
Gateway event loop freezes after Slack connects on 2026.3.13, health endpoint stays live but inbound Slack messages are never processed; 2026.3.8 works fine with identical config.
### Steps to reproduce
1. Working 2026.3.8 gateway with Slack socket mod... | closed | not_planned | true | 5 | [
"bug",
"regression"
] | [] | 2026-03-14T16:23:24Z | 2026-04-27T04:42:39Z | 2026-04-25T01:21:16Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | vaas-git | 263,815,399 | U_kgDOD7mA5w | User | false |
openclaw/openclaw | 4,196,931,818 | I_kwDOQb6kR876KBjq | 59,908 | https://github.com/openclaw/openclaw/issues/59908 | https://api.github.com/repos/openclaw/openclaw/issues/59908 | tools.exec.safeBins not inherited by isolated cron/heartbeat sessions | ## Updated Diagnosis (2026-04-03)
**Original hypothesis:** `tools.exec.safeBins` not propagated to isolated cron/heartbeat sessions.
**Actual root cause:** safeBins config propagation is correct. The issue is that most bins users configure in `safeBins` silently fail because they lack built-in `SafeBinProfile` defini... | closed | duplicate | true | 1 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-04-02T20:53:56Z | 2026-04-27T04:42:37Z | 2026-04-25T01:21:21Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | KimAcross | 206,874,767 | U_kgDODFSojw | User | false |
openclaw/openclaw | 4,184,908,563 | I_kwDOQb6kR875cKMT | 58,834 | https://github.com/openclaw/openclaw/issues/58834 | https://api.github.com/repos/openclaw/openclaw/issues/58834 | [Feature]: Allow cron agentTurn prompts to load from files | ## Summary
Allow cron `agentTurn` jobs to load `payload.message` from a file instead of requiring the full prompt inline in `jobs.json`.
## Problem to solve
Right now cron jobs require the full `payload.message` to be embedded as a plain string in the job definition. That gets awkward quickly for larger prompts, espec... | closed | not_planned | true | 1 | [] | [] | 2026-04-01T07:48:52Z | 2026-04-27T04:42:36Z | 2026-04-25T01:21:27Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | mgreiner79 | 52,714,186 | MDQ6VXNlcjUyNzE0MTg2 | User | false |
openclaw/openclaw | 4,326,546,030 | I_kwDOQb6kR88AAAABAeHabg | 71,330 | https://github.com/openclaw/openclaw/issues/71330 | https://api.github.com/repos/openclaw/openclaw/issues/71330 | Feature: Configurable memory promotion target file | ## Problem
Memory-core's short-term promotion system hardcodes `MEMORY.md` as the target file for promoted memories (source: `dist/short-term-promotion-BMEmSJk4.js:1257`). This causes several issues:
1. **Bootstrap truncation:** `MEMORY.md` is subject to OpenClaw's 12KB per-file bootstrap limit. As promotions accumul... | open | null | false | 1 | [] | [] | 2026-04-25T01:21:31Z | 2026-05-03T07:04:50Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | oc-mostlycopy | 270,197,315 | U_kgDOEBriQw | User | false |
openclaw/openclaw | 4,089,603,035 | I_kwDOQb6kR87zwmPb | 49,124 | https://github.com/openclaw/openclaw/issues/49124 | https://api.github.com/repos/openclaw/openclaw/issues/49124 | [Bug]: Cron job error state accumulates during host sleep/connectivity loss, creating false failure signals | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When a host (especially laptops) sleeps or loses connectivity, cron job delivery fails with connection errors that inflate consecutiveErrors counts. Agents reviewing heartbeat/logs see "10 consecutive errors" and treat it as a software fail... | closed | not_planned | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-03-17T16:15:40Z | 2026-04-27T04:42:34Z | 2026-04-25T01:21:32Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | jimdawdy-hub | 262,052,366 | U_kgDOD56aDg | User | false |
openclaw/openclaw | 4,076,172,305 | I_kwDOQb6kR87y9XQR | 46,332 | https://github.com/openclaw/openclaw/issues/46332 | https://api.github.com/repos/openclaw/openclaw/issues/46332 | [Bug]: WhatsApp bug | ### Bug type
Regression (worked before, now fails)
### Summary
This latest update has something interesting: I have three agents running on three different machines.
The OpenClaw agents on two of those machines keep reporting that WhatsApp is "linked," but they don’t have any active listeners, which causes messages... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-03-14T16:24:48Z | 2026-04-27T04:42:33Z | 2026-04-25T01:21:33Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | runawaydevil | 8,763,225 | MDQ6VXNlcjg3NjMyMjU= | User | false |
openclaw/openclaw | 4,131,975,747 | I_kwDOQb6kR872SPJD | 54,180 | https://github.com/openclaw/openclaw/issues/54180 | https://api.github.com/repos/openclaw/openclaw/issues/54180 | Cron jobs run serially despite maxConcurrentRuns > 1 (Nested lane bottleneck) | ## Summary
Setting `cron.maxConcurrentRuns` to a value greater than 1 does not actually achieve concurrent cron job execution. Jobs still run one at a time.
## Root Cause
Cron jobs route their inner agent work through the `Nested` command lane (via `resolveGlobalLane`) to avoid deadlocking on the Cron lane itself. H... | closed | not_planned | true | 3 | [] | [] | 2026-03-25T02:42:10Z | 2026-04-27T04:42:33Z | 2026-04-25T01:21:36Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | sparkleMing | 38,350,996 | MDQ6VXNlcjM4MzUwOTk2 | User | false |
openclaw/openclaw | 4,083,370,168 | I_kwDOQb6kR87zY0i4 | 48,333 | https://github.com/openclaw/openclaw/issues/48333 | https://api.github.com/repos/openclaw/openclaw/issues/48333 | Cron list tool result too verbose — causes LLM to miss/hallucinate entries | ## Problem
When the agent has many cron jobs (10+), the `cron(action="list")` tool returns the full job object for each entry, including complete `payload.text`, `schedule` details, `sessionTarget`, etc. This makes the tool result very large and causes the LLM to:
1. **Miss entries** when scanning the list (e.g., fai... | closed | not_planned | true | 2 | [] | [] | 2026-03-16T16:01:09Z | 2026-04-27T04:42:32Z | 2026-04-25T01:21:38Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Senorsen | 4,176,091 | MDQ6VXNlcjQxNzYwOTE= | User | false |
openclaw/openclaw | 4,076,193,988 | I_kwDOQb6kR87y9cjE | 46,355 | https://github.com/openclaw/openclaw/issues/46355 | https://api.github.com/repos/openclaw/openclaw/issues/46355 | [Bug]: Control UI 'Show Thinking' toggle doesn't work without manual localStorage workaround | **Describe the bug**
The "Show Thinking" toggle in the OpenClaw Control UI chat view doesn't display thinking/reasoning output unless the user manually sets a localStorage key via browser console.
**To Reproduce**
1. Open OpenClaw Control UI
2. Start a chat session with a model that supports thinking (e.g., Claude, Op... | closed | completed | true | 1 | [] | [] | 2026-03-14T16:34:31Z | 2026-04-27T04:42:30Z | 2026-04-25T01:21:51Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | danen-carlson | 4,970,765 | MDQ6VXNlcjQ5NzA3NjU= | User | false |
openclaw/openclaw | 4,076,198,223 | I_kwDOQb6kR87y9dlP | 46,358 | https://github.com/openclaw/openclaw/issues/46358 | https://api.github.com/repos/openclaw/openclaw/issues/46358 | [Bug] Gateway returns "missing scope: operator.read" error in Control UI after upgrade to 2026.3.13 | ## 🐛 Bug: Missing scope operator.read after upgrade to 2026.3.13
### Description
After upgrading to OpenClaw 2026.3.13, the Control UI cannot read Gateway status and consistently shows the error: `missing scope: operator.read`.
### Environment
| Item | Value |
|------|-------|
| **OpenClaw Version** | 2026.3.13 (com... | closed | completed | true | 2 | [] | [] | 2026-03-14T16:36:00Z | 2026-04-27T04:42:28Z | 2026-04-25T01:22:08Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | andyleeqd | 43,713,363 | MDQ6VXNlcjQzNzEzMzYz | User | false |
openclaw/openclaw | 4,072,144,005 | I_kwDOQb6kR87yt_yF | 45,285 | https://github.com/openclaw/openclaw/issues/45285 | https://api.github.com/repos/openclaw/openclaw/issues/45285 | [Bug]: Isolated cron jobs never execute agent turn — enqueue succeeds but dispatch never fires (2026.3.11 + 2026.3.12) | ### Bug type
Regression (worked before, now fails)
### Summary
Isolated cron jobs enqueue successfully but never execute agent turns, timing out at their configured limit regardless of model or job complexity.
### Steps to reproduce
1. Create a cron job: `openclaw cron add --name "Test Ping" --every 5m --session i... | closed | duplicate | true | 6 | [
"bug",
"close:duplicate",
"dedupe:child",
"regression"
] | [] | 2026-03-13T16:29:22Z | 2026-04-27T04:42:27Z | 2026-04-25T01:22:12Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | jrvanwinkle | 261,341,012 | U_kgDOD5O_VA | User | false |
openclaw/openclaw | 4,189,158,013 | I_kwDOQb6kR875sXp9 | 59,211 | https://github.com/openclaw/openclaw/issues/59211 | https://api.github.com/repos/openclaw/openclaw/issues/59211 | cron edit: toolsAllow rejected as unexpected property in payload patch | ## Bug
`openclaw cron edit <id> --tools "exec,read"` fails with:
```
GatewayClientRequestError: invalid cron.update params: at /patch/payload: unexpected property 'toolsAllow'; at /patch/payload/kind: must be equal to constant; at /patch/payload: must match a schema in anyOf
```
## Version
OpenClaw 2026.4.1 (da64a9... | closed | duplicate | true | 4 | [
"close:duplicate",
"dedupe:parent",
"dedupe:child"
] | [] | 2026-04-01T18:51:26Z | 2026-04-27T04:42:23Z | 2026-04-25T01:21:25Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | nicholasrae | 258,079,684 | U_kgDOD2H7xA | User | false |
openclaw/openclaw | 4,100,767,946 | I_kwDOQb6kR870bMDK | 50,373 | https://github.com/openclaw/openclaw/issues/50373 | https://api.github.com/repos/openclaw/openclaw/issues/50373 | Cron announce delivery to Mattermost DM fails with 403 — missing user: prefix in delivery.to | ## Bug Description
Cron jobs with `delivery.mode: "announce"` + `channel: "mattermost"` fail to deliver to DM conversations. The error is `Mattermost API 403 Forbidden: You do not have the appropriate permissions.`
The cron task itself executes successfully (agent runs, produces output), but the final announce delive... | closed | not_planned | true | 3 | [] | [] | 2026-03-19T10:29:05Z | 2026-04-27T04:42:23Z | 2026-04-25T01:21:34Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | bingyuan-ai | 248,037,429 | U_kgDODsjANQ | User | false |
openclaw/openclaw | 4,076,211,270 | I_kwDOQb6kR87y9gxG | 46,370 | https://github.com/openclaw/openclaw/issues/46370 | https://api.github.com/repos/openclaw/openclaw/issues/46370 | Bug: Tool-call args JSON parse crashes on unescaped newlines/quotes | ## Summary
OpenClaw can fail a tool-calling turn with:
`Expected ',' or '}' after property value in JSON`
when an LLM streams malformed tool arguments (for example literal newlines or unescaped quotes inside JSON strings).
## Why this matters
In real agent usage, model outputs are not always strictly valid JSON at e... | closed | completed | true | 1 | [] | [] | 2026-03-14T16:42:20Z | 2026-04-27T04:42:22Z | 2026-04-25T01:22:25Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | kexinzhang3-lab | 243,058,290 | U_kgDODnzGcg | User | false |
openclaw/openclaw | 4,076,211,870 | I_kwDOQb6kR87y9g6e | 46,372 | https://github.com/openclaw/openclaw/issues/46372 | https://api.github.com/repos/openclaw/openclaw/issues/46372 | Message loss during WhatsApp channel restart (stale-socket reconnect) | ## Problem
When the WhatsApp health monitor detects a stale socket and restarts the provider, messages in flight are dropped. No retry or delivery confirmation occurs.
## Timeline (from logs)
```
12:23:19 — User sends message
12:23:41 — Gateway detects stale socket, restarts WhatsApp provider
→ "[whatsapp:d... | closed | completed | true | 3 | [] | [] | 2026-03-14T16:42:38Z | 2026-04-27T04:42:21Z | 2026-04-25T01:22:42Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | danielkimtoronto | 268,210,590 | U_kgDOD_yRng | User | false |
openclaw/openclaw | 4,076,219,141 | I_kwDOQb6kR87y9isF | 46,379 | https://github.com/openclaw/openclaw/issues/46379 | https://api.github.com/repos/openclaw/openclaw/issues/46379 | Sub-agent and cron agentTurn model parameter is silently ignored | ## Bug Report
### Description
The `model` parameter is accepted in both `sessions_spawn` and cron `agentTurn` payloads but is silently ignored — all isolated sessions fall back to the parent agent's primary model (`agents.defaults.model.primary`).
### Expected Behavior
When a model is specified via:
1. `sessions_spaw... | closed | completed | true | 1 | [] | [] | 2026-03-14T16:46:20Z | 2026-04-27T04:42:20Z | 2026-04-25T01:22:59Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | SteveMcKean | 36,823,521 | MDQ6VXNlcjM2ODIzNTIx | User | false |
openclaw/openclaw | 4,076,232,063 | I_kwDOQb6kR87y9l1_ | 46,385 | https://github.com/openclaw/openclaw/issues/46385 | https://api.github.com/repos/openclaw/openclaw/issues/46385 | OAuth refresh token invalidated after gateway restart — should auto-recover | ## Problem
When the OpenClaw gateway restarts (e.g., during `openclaw update run` or `config.patch`), the OpenAI Codex OAuth refresh token can become invalidated. The restart interrupts an in-flight token refresh, consuming the single-use refresh token without storing the new one. Subsequent requests fail with:
```
4... | closed | completed | true | 3 | [] | [
"vincentkoc"
] | 2026-03-14T16:53:03Z | 2026-04-27T04:31:01Z | 2026-04-25T01:23:16Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | vswarm-ai | 258,697,688 | U_kgDOD2tp2A | User | false |
openclaw/openclaw | 4,076,240,921 | I_kwDOQb6kR87y9oAZ | 46,399 | https://github.com/openclaw/openclaw/issues/46399 | https://api.github.com/repos/openclaw/openclaw/issues/46399 | [Bug]: voice_call still throws EADDRINUSE on configured webhook port with mock provider and single gateway | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
`voice_call` throws `EADDRINUSE` on the configured webhook port during call initiation even when the listener is already owned by `openclaw-gateway` and the setup has been reduced to a single gateway daemon with `provider: "mock"`.
### Ste... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-03-14T16:57:29Z | 2026-04-27T04:42:20Z | 2026-04-25T01:23:33Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | asfarasimconcerned | 116,858,071 | U_kgDOBvcc1w | User | false |
openclaw/openclaw | 4,076,240,994 | I_kwDOQb6kR87y9oBi | 46,400 | https://github.com/openclaw/openclaw/issues/46400 | https://api.github.com/repos/openclaw/openclaw/issues/46400 | Sub-agent LLM requests always timeout with GLM-5 while Claude Code terminal works fine |
### Bug type
Regression (worked before, now fails)
### Summary
子代理（如 frontend-dev）使用 GLM-5 模型时，LLM 请求总是在 60 秒左右超时，但使用 Claude Code 终端 + GLM 从不超时
### Steps to reproduce
1. 启动 OpenClaw gateway
2. 在 Discord @frontend-dev 触发子代理
3. 发送任意任务请求
4. 等待约 42-60 秒
5. 收到 `LLM request timed out` 错误
6. 子代理被 health-monitor 重启
7. 循环... | closed | completed | true | 1 | [] | [] | 2026-03-14T16:57:31Z | 2026-04-27T04:42:19Z | 2026-04-25T01:23:50Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | 2457205703 | 65,269,121 | MDQ6VXNlcjY1MjY5MTIx | User | false |
openclaw/openclaw | 4,076,242,478 | I_kwDOQb6kR87y9oYu | 46,402 | https://github.com/openclaw/openclaw/issues/46402 | https://api.github.com/repos/openclaw/openclaw/issues/46402 | Feishu main session loses channel metadata, causing tools.elevated.allowFrom.feishu to never match | ## Summary
On a Feishu direct-chat main session, `tools.elevated.allowFrom.feishu` appears to be configured correctly, but elevated exec still fails because the session channel resolves to `unknown` instead of `feishu`.
This makes `tools.elevated.allowFrom.<channel>` ineffective for Feishu in the main session.
## En... | closed | completed | true | 1 | [] | [] | 2026-03-14T16:58:16Z | 2026-04-27T04:42:18Z | 2026-04-25T01:24:07Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | one-box-u | 200,010,714 | U_kgDOC-vr2g | User | false |
openclaw/openclaw | 4,041,174,600 | I_kwDOQb6kR87w325I | 39,801 | https://github.com/openclaw/openclaw/issues/39801 | https://api.github.com/repos/openclaw/openclaw/issues/39801 | [Bug] Cron job in isolated session fails with 'Field required' validation error on tools format | ## Bug Description
Cron jobs configured with `sessionTarget: isolated` fail with a validation error when trying to execute tools.
## Error Message
```
22 validation errors:
{'type': 'missing', 'loc': ('body', 'tools', 0, 'function'), 'msg': 'Field required', 'input': {'name': 'read', 'description': 'Read the conte... | closed | duplicate | true | 3 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-03-08T12:10:50Z | 2026-04-27T04:42:17Z | 2026-04-25T01:24:14Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | 0065paula | 374,316 | MDQ6VXNlcjM3NDMxNg== | User | false |
openclaw/openclaw | 4,326,556,141 | I_kwDOQb6kR88AAAABAeIB7Q | 71,335 | https://github.com/openclaw/openclaw/issues/71335 | https://api.github.com/repos/openclaw/openclaw/issues/71335 | Feature: sync.watch should default to false in gateway mode | ## Problem
When OpenClaw runs as a headless gateway (gateway.mode set), `memorySearch.sync.watch` defaults to `true`, which creates a chokidar file watcher for every memory file and extra path. On a multi-agent gateway this leaks file descriptors — we observed 1,292 chokidar watchers on a single instance.
Interactive... | open | null | false | 1 | [] | [] | 2026-04-25T01:24:18Z | 2026-05-03T06:49:49Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | chrisabad | 64,236,246 | MDQ6VXNlcjY0MjM2MjQ2 | User | false |
openclaw/openclaw | 4,076,245,774 | I_kwDOQb6kR87y9pMO | 46,403 | https://github.com/openclaw/openclaw/issues/46403 | https://api.github.com/repos/openclaw/openclaw/issues/46403 | Feature: Route Slack Lists item comment threads to bound agents | ## Problem
When an agent is @mentioned in a Slack List item's comment thread, the mention is not routed to the agent. The agent never sees the message.
## Context
Slack Lists (task lists, project trackers, etc.) allow comments on individual list items. These comments surface as regular `message` events with a `threa... | closed | completed | true | 1 | [] | [] | 2026-03-14T16:59:48Z | 2026-04-27T04:42:17Z | 2026-04-25T01:24:24Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | RupertBarrow | 48,629,751 | MDQ6VXNlcjQ4NjI5NzUx | User | false |
openclaw/openclaw | 4,076,274,220 | I_kwDOQb6kR87y9wIs | 46,413 | https://github.com/openclaw/openclaw/issues/46413 | https://api.github.com/repos/openclaw/openclaw/issues/46413 | [Feature Request] Long tasks (>10min) cannot notify users after WebSocket disconnect - need lifecycle hooks or async callbacks | Hey team, I've been hitting a frustrating wall with long-running tasks and thought I'd share my experience.
**The Problem**
When I kick off a task that takes more than 10 minutes (think data analysis, batch processing, etc.), the WebSocket connection between OpenClaw and Feishu drops after exactly 10 minutes. The task... | closed | completed | true | 1 | [] | [] | 2026-03-14T17:13:07Z | 2026-04-27T04:42:16Z | 2026-04-25T01:24:42Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | WaleyChAn | 48,935,163 | MDQ6VXNlcjQ4OTM1MTYz | User | false |
openclaw/openclaw | 4,076,283,361 | I_kwDOQb6kR87y9yXh | 46,421 | https://github.com/openclaw/openclaw/issues/46421 | https://api.github.com/repos/openclaw/openclaw/issues/46421 | OP-143: Entity-first retrieval — include Entity node properties as first-class search results | ## Problem
The hybrid search pipeline (`hybridSearch`) has three signals:
- **Signal 1 (vector)** → Memory nodes only
- **Signal 2 (BM25)** → Memory nodes only
- **Signal 3 (graph)** → Entity nodes via `structuredGraphSearch` / `causalChainSearch`
Signal 3 (`structuredGraphSearch`) already synthesizes Entity node p... | closed | not_planned | true | 2 | [] | [] | 2026-03-14T17:17:27Z | 2026-04-27T04:42:15Z | 2026-04-25T01:24:59Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | tsukhani | 156,283 | MDQ6VXNlcjE1NjI4Mw== | User | false |
openclaw/openclaw | 4,076,284,135 | I_kwDOQb6kR87y9yjn | 46,422 | https://github.com/openclaw/openclaw/issues/46422 | https://api.github.com/repos/openclaw/openclaw/issues/46422 | [Bug]: still reproducing missing scope: operator.read on 2026.3.13 after full re-auth + token/device resets | ### Bug type
Regression (worked before, now fails)
### Summary
openclaw status --all and openclaw security audit --deep continue to report:
gateway.probe_failed
missing scope: operator.read
What I already tried
Re-auth OpenAI Codex OAuth successfully:
bashCopyCopied!
openclaw models auth login --provider openai-co... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-03-14T17:17:52Z | 2026-04-27T04:42:14Z | 2026-04-25T01:25:17Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ammosleybooking-helix | 266,064,345 | U_kgDOD9vR2Q | User | false |
openclaw/openclaw | 4,076,353,975 | I_kwDOQb6kR87y-Dm3 | 46,457 | https://github.com/openclaw/openclaw/issues/46457 | https://api.github.com/repos/openclaw/openclaw/issues/46457 | Bug: WhatsApp quoted self-reply can bypass pattern-based mention gating in groups | ## Summary
In WhatsApp group chats, replying to (quoting) a message sent by the same WhatsApp account that is connected to OpenClaw can incorrectly bypass `requireMention` / pattern-based `mentionPatterns` gating.
In practice, this means a group message that does not match the configured invocation regex can still be... | closed | completed | true | 1 | [] | [] | 2026-03-14T17:51:16Z | 2026-04-27T04:42:14Z | 2026-04-25T01:25:35Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | rodrigoscoelho | 6,912,410 | MDQ6VXNlcjY5MTI0MTA= | User | false |
openclaw/openclaw | 4,076,356,457 | I_kwDOQb6kR87y-ENp | 46,462 | https://github.com/openclaw/openclaw/issues/46462 | https://api.github.com/repos/openclaw/openclaw/issues/46462 | [Bug]: **OpenClaw 在 Docker 容器中运行时，始终绑定到 127.0.0.1（本地回环地址），拒绝绑定到 0.0.0.0（所有接口），导致即使端口映射正确，宿主机也无法访问。** | ### Bug type
Regression (worked before, now fails)
### Summary
**OpenClaw 在 Docker 容器中运行时，始终绑定到 127.0.0.1（本地回环地址），拒绝绑定到 0.0.0.0（所有接口），导致即使端口映射正确，宿主机也无法访问。**
虽然 Dockerfile 中有明确说明：
```
# IMPORTANT: With Docker bridge networking (-p 18789:18789), loopback bind
# makes the gateway unreachable from the host. Either:
# ... | closed | not_planned | true | 2 | [
"bug",
"regression"
] | [] | 2026-03-14T17:52:26Z | 2026-04-27T04:42:13Z | 2026-04-25T01:25:52Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | gg-2020 | 62,710,810 | MDQ6VXNlcjYyNzEwODEw | User | false |
openclaw/openclaw | 4,076,365,023 | I_kwDOQb6kR87y-GTf | 46,473 | https://github.com/openclaw/openclaw/issues/46473 | https://api.github.com/repos/openclaw/openclaw/issues/46473 | [Bug]: [macOS] App UI deadlocks (grey menus) & shows stale "launchd start timeout" due to main thread blocking during gateway kickstart | ### Bug type
Regression (worked before, now fails)
### Summary
The macOS Companion App experiences severe UI deadlocks (menus become greyed out/unresponsive) and displays a false/stale launchd start timeout error.
This occurs because the App aggressively executes launchctl kickstart -k on startup and seemingly block... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-03-14T17:56:34Z | 2026-04-27T04:42:12Z | 2026-04-25T01:26:09Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | lonelygo | 8,328,467 | MDQ6VXNlcjgzMjg0Njc= | User | false |
openclaw/openclaw | 4,076,369,184 | I_kwDOQb6kR87y-HUg | 46,476 | https://github.com/openclaw/openclaw/issues/46476 | https://api.github.com/repos/openclaw/openclaw/issues/46476 | ACP thread spawn fails: spawnedBy validation rejects agent:*:acp:* session keys | ## Bug Description
`sessions_spawn` with `runtime: "acp"` and `thread: true` fails with:
```
spawnedBy is only supported for subagent:* sessions
```
## Root Cause
In the session store patch validation (gateway-cli dist, around line 9047), the `spawnedBy` field guard only checks `isSubagentSessionKey(storeKey)`:
``... | closed | completed | true | 2 | [] | [
"vincentkoc"
] | 2026-03-14T17:58:28Z | 2026-04-27T04:42:11Z | 2026-04-25T01:26:26Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | itsahedge | 51,177,340 | MDQ6VXNlcjUxMTc3MzQw | User | false |
openclaw/openclaw | 4,076,395,195 | I_kwDOQb6kR87y-Nq7 | 46,483 | https://github.com/openclaw/openclaw/issues/46483 | https://api.github.com/repos/openclaw/openclaw/issues/46483 | [Bug/Docs]: profile="user" (existing-session) setup unclear — Chrome rejects --remote-debugging-port on default data dir | ### Bug type
Documentation gap / UX confusion
### Summary
2026.3.13 adds `profile="user"` with `driver: "existing-session"` for signed-in Chrome sessions. Tried setting it up and hit a dead end.
The changelog mentions "Chrome DevTools MCP attach mode" but there are no setup instructions linked anywhere.
The browse... | closed | completed | true | 2 | [] | [] | 2026-03-14T18:11:38Z | 2026-04-27T04:42:11Z | 2026-04-25T01:26:43Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | adhishthite | 31,769,894 | MDQ6VXNlcjMxNzY5ODk0 | User | false |
openclaw/openclaw | 4,076,413,933 | I_kwDOQb6kR87y-SPt | 46,492 | https://github.com/openclaw/openclaw/issues/46492 | https://api.github.com/repos/openclaw/openclaw/issues/46492 | [Bug]: Failed to switch the Ollama model | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
<img width="742" height="138" alt="Image" src="https://github.com/user-attachments/assets/ecd2da05-f915-446d-b3f0-e9477afcfcc1" />
My default model is set to the MiniMax model, and I have also added a local model in Ollama. However, I’m una... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-03-14T18:21:27Z | 2026-04-27T04:42:10Z | 2026-04-25T01:27:01Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | flamingoios0916-collab | 261,232,683 | U_kgDOD5IYKw | User | false |
openclaw/openclaw | 4,326,553,657 | I_kwDOQb6kR88AAAABAeH4OQ | 71,333 | https://github.com/openclaw/openclaw/issues/71333 | https://api.github.com/repos/openclaw/openclaw/issues/71333 | GitHub Copilot Responses (gpt-5.x): multi-turn fails with 'encrypted content item_id did not match' | ## Summary
Multi-turn requests against GitHub Copilot–hosted Responses models (e.g. `github-copilot/gpt-5.5`, `gpt-5.4`, `gpt-5-mini`) fail on the second turn with HTTP 400 from the upstream:
```
The encrypted content for item rs_<…> could not be verified.
Reason: Encrypted content item_id did not match the target it... | closed | completed | true | 1 | [] | [] | 2026-04-25T01:23:36Z | 2026-04-27T04:42:09Z | 2026-04-25T01:27:04Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ottodeng | 20,509,364 | MDQ6VXNlcjIwNTA5MzY0 | User | false |
openclaw/openclaw | 4,076,420,771 | I_kwDOQb6kR87y-T6j | 46,495 | https://github.com/openclaw/openclaw/issues/46495 | https://api.github.com/repos/openclaw/openclaw/issues/46495 | [Bug]: built-in browser profile "user" (existing-session / chrome-mcp) times out on Linux even though Chrome is detected and listening | ### Bug type
Regression / broken feature
### Summary
On Linux, OpenClaw 2026.3.13 detects the built-in `user` browser profile correctly as an `existing-session` / `chrome-mcp` live Chrome session, but actual browser requests against that profile time out after 45 seconds.
This is very similar to the new macOS repor... | closed | completed | true | 6 | [] | [] | 2026-03-14T18:24:52Z | 2026-04-27T04:42:08Z | 2026-04-25T01:27:18Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | JLDynamics | 240,741,042 | U_kgDODllqsg | User | false |
openclaw/openclaw | 4,076,447,222 | I_kwDOQb6kR87y-aX2 | 46,504 | https://github.com/openclaw/openclaw/issues/46504 | https://api.github.com/repos/openclaw/openclaw/issues/46504 | Control UI: Add aria-live regions for screen reader accessibility | ## Problem
The OpenClaw Control UI web interface does not announce new messages to screen readers automatically. Users relying on screen readers must manually navigate to find replies after sending a message.
## Current Behavior
- When a user sends a message and the assistant replies, the reply is not announced
- Scre... | closed | completed | true | 1 | [] | [] | 2026-03-14T18:38:18Z | 2026-04-27T04:42:08Z | 2026-04-25T01:27:35Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | kellylford | 44,002,405 | MDQ6VXNlcjQ0MDAyNDA1 | User | false |
openclaw/openclaw | 4,076,479,475 | I_kwDOQb6kR87y-iPz | 46,517 | https://github.com/openclaw/openclaw/issues/46517 | https://api.github.com/repos/openclaw/openclaw/issues/46517 | Control UI: Context display shows misleading '378k/128k' vs actual 58k/128k (46%) | ## Bug Description
The Control UI dashboard displays context usage as '378k/128k' which appears to show 297% overflow (🚨), but the actual context window usage is 58k/128k (46%) (✅).
## What's Happening
- **378k** = Lifetime tokens sent to model (cumulative session total, informational)
- **58k** = Actual active con... | closed | completed | true | 1 | [] | [] | 2026-03-14T18:55:17Z | 2026-04-27T04:42:07Z | 2026-04-25T01:27:52Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Kru442 | 187,809,917 | U_kgDOCzHAfQ | User | false |
openclaw/openclaw | 4,076,491,267 | I_kwDOQb6kR87y-lID | 46,521 | https://github.com/openclaw/openclaw/issues/46521 | https://api.github.com/repos/openclaw/openclaw/issues/46521 | [Bug]: openclaw update silently fails when a system-level install shadows the user-global install | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When OpenClaw is installed both system-wide (sudo npm install -g) and as a user-global install (npm install -g), openclaw update only updates the user-global install but does not warn that a system-level install exists and is shadowing it i... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-03-14T19:00:20Z | 2026-04-27T04:42:06Z | 2026-04-25T01:28:09Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | munki-nut | 133,858,466 | U_kgDOB_qEog | User | false |
openclaw/openclaw | 4,076,524,510 | I_kwDOQb6kR87y-tPe | 46,524 | https://github.com/openclaw/openclaw/issues/46524 | https://api.github.com/repos/openclaw/openclaw/issues/46524 | TDZ ReferenceError on ANTHROPIC_MODEL_ALIASES in auth-profiles bundle | ## Bug Description
`normalizeAnthropicModelId()` in the `auth-profiles` bundle chunk references `ANTHROPIC_MODEL_ALIASES` (a `const` defined in the `model-selection.ts` region). In certain execution paths, this function is called before the `const` declaration is reached in the bundled output, causing a **Temporal Dea... | closed | completed | true | 1 | [
"close:duplicate",
"close:already-fixed",
"dedupe:child"
] | [
"vincentkoc"
] | 2026-03-14T19:15:21Z | 2026-04-27T04:31:20Z | 2026-04-25T01:28:26Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | milofoxbox | 261,938,990 | U_kgDOD5zfLg | User | false |
openclaw/openclaw | 4,076,546,965 | I_kwDOQb6kR87y-yuV | 46,530 | https://github.com/openclaw/openclaw/issues/46530 | https://api.github.com/repos/openclaw/openclaw/issues/46530 | [Feature]: /new functionality | ### Summary
You guys could make it so /new saves the current context in the current session and summarizes it for a handoff to a new session. That way you can better continue your work going into the new session with the relevant context points saved. And after the summary is created and saved, it does the /reset comm... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-03-14T19:26:03Z | 2026-04-27T04:42:05Z | 2026-04-25T01:28:43Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | JKSV83 | 254,005,406 | U_kgDODyPQng | User | false |
openclaw/openclaw | 4,076,573,602 | I_kwDOQb6kR87y-5Oi | 46,534 | https://github.com/openclaw/openclaw/issues/46534 | https://api.github.com/repos/openclaw/openclaw/issues/46534 | Control UI: Image upload button present but uploads don't deliver images to session | ## Bug Description
The Control UI webchat shows an image upload button, but uploaded images never arrive in the session context. The button is functional (can select files) but the upload pipeline doesn't complete.
## What's Happening
- **Upload button visible** ✅ - UI shows the attachment/upload icon
- **File selec... | closed | not_planned | true | 7 | [
"close:duplicate",
"close:already-fixed",
"dedupe:child"
] | [
"vincentkoc"
] | 2026-03-14T19:40:37Z | 2026-04-27T04:31:49Z | 2026-04-25T01:29:00Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Kru442 | 187,809,917 | U_kgDOCzHAfQ | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.