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 | 4,295,275,587 | I_kwDOQb6kR88AAAABAAS0Qw | 69,309 | https://github.com/openclaw/openclaw/issues/69309 | https://api.github.com/repos/openclaw/openclaw/issues/69309 | fix: MEDIA directive silently drops media without user notification | ## Problem
When an agent replies with `MEDIA:/path/to/file.png` and the path fails validation (not in `localRoots` allowlist, file not found, not a regular file, etc.), OpenClaw Core silently drops the media item. Neither the user nor the agent receives any feedback that the attachment was not sent.
### 3 silent-swal... | closed | completed | true | 1 | [] | [] | 2026-04-20T11:25:48Z | 2026-04-28T04:49:25Z | 2026-04-25T19:07:14Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Jerry-Xin | 3,401,616 | MDQ6VXNlcjM0MDE2MTY= | User | false |
openclaw/openclaw | 4,295,283,446 | I_kwDOQb6kR88AAAABAATS9g | 69,311 | https://github.com/openclaw/openclaw/issues/69311 | https://api.github.com/repos/openclaw/openclaw/issues/69311 | fix: MEDIA: directive false-positive extraction from indented code blocks | ## Problem
When an agent's reply contains `MEDIA:` inside a **4-space / tab indented code block** (CommonMark indented code block), `splitMediaFromOutput()` incorrectly extracts it as a real media directive. This causes:
1. The code line containing `MEDIA:` is silently removed from the output text
2. A spurious media... | closed | not_planned | true | 1 | [] | [] | 2026-04-20T11:27:24Z | 2026-04-29T04:35:50Z | 2026-04-26T14:55:56Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Jerry-Xin | 3,401,616 | MDQ6VXNlcjM0MDE2MTY= | User | false |
openclaw/openclaw | 4,295,296,606 | I_kwDOQb6kR88AAAABAAUGXg | 69,314 | https://github.com/openclaw/openclaw/issues/69314 | https://api.github.com/repos/openclaw/openclaw/issues/69314 | [Feature]: Add optional OPENCLAW_DOCKER_PLATFORM support to ./scripts/docker/setup.sh for cross-platform builds | ### Summary
### Title
Add optional `OPENCLAW_DOCKER_PLATFORM` support to `setup.sh` for cross-platform builds
---
### Description
Currently, `scripts/docker/setup.sh` does not support passing a `--platform` argument to `docker build`.
This makes it difficult to build images for a different architecture than the ho... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-20T11:30:03Z | 2026-04-28T04:49:25Z | 2026-04-25T19:07:21Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | safrano9999 | 240,768,512 | U_kgDODlnWAA | User | false |
openclaw/openclaw | 4,295,363,854 | I_kwDOQb6kR88AAAABAAYNDg | 69,315 | https://github.com/openclaw/openclaw/issues/69315 | https://api.github.com/repos/openclaw/openclaw/issues/69315 | Feature request: non-blocking / async TTS delivery | ## Summary
When TTS is enabled, OpenClaw generates audio for each reply chunk **sequentially and synchronously** before processing the next inbound message. This creates a noticeable delay — the user sees all text replies arrive quickly, then voice notes attach one by one, and any new message they send is blocked unti... | open | null | false | 1 | [] | [] | 2026-04-20T11:42:46Z | 2026-05-02T13:08:03Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | IGI30 | 10,922,046 | MDQ6VXNlcjEwOTIyMDQ2 | User | false |
openclaw/openclaw | 4,295,604,735 | I_kwDOQb6kR88AAAABAAm5_w | 69,326 | https://github.com/openclaw/openclaw/issues/69326 | https://api.github.com/repos/openclaw/openclaw/issues/69326 | [Bug] Replies not routed back to WeChat when message originates from webchat UI | ## Bug Description
When a message is sent from the **browser webchat UI** to the WeChat-connected session, the assistant's reply appears in the browser but is **NOT delivered back to the WeChat mobile app**.
However, when a message is sent directly from the **WeChat mobile app**, replies are correctly delivered back t... | closed | not_planned | true | 1 | [] | [] | 2026-04-20T12:22:01Z | 2026-05-02T04:28:56Z | 2026-04-29T08:22:17Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | HOHOgithub | 73,624,796 | MDQ6VXNlcjczNjI0Nzk2 | User | false |
openclaw/openclaw | 4,295,685,637 | I_kwDOQb6kR88AAAABAAr2BQ | 69,329 | https://github.com/openclaw/openclaw/issues/69329 | https://api.github.com/repos/openclaw/openclaw/issues/69329 | [Bug]: exec runtime surfaces completed from lifecycle alone with no seam for artifact-gated closure | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The exec-driven runtime can surface completed from process lifecycle alone, with no post-run artifact verification, wrong-tree rejection, or closure-state distinction for artifact-dependent workflow consumers.
... | closed | not_planned | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-20T12:35:10Z | 2026-04-28T04:49:24Z | 2026-04-25T19:07:34Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | JB357345688 | 205,907,616 | U_kgDODEXmoA | User | false |
openclaw/openclaw | 4,295,724,783 | I_kwDOQb6kR88AAAABAAuO7w | 69,330 | https://github.com/openclaw/openclaw/issues/69330 | https://api.github.com/repos/openclaw/openclaw/issues/69330 | MCP server leak: duplicate processes accumulate on gateway reconnect (WhatsApp 499) | ## Bug: MCP servers leak on gateway reconnect
### Summary
Every time the WhatsApp gateway disconnects and reconnects (status 499 → 200), OpenClaw spawns a new set of MCP server processes without terminating the previous ones. Over time, this causes hundreds of duplicate MCP processes to accumulate, exhausting system ... | closed | completed | true | 1 | [] | [] | 2026-04-20T12:41:25Z | 2026-04-28T04:49:23Z | 2026-04-25T19:07:41Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | davidmedel | 52,670,482 | MDQ6VXNlcjUyNjcwNDgy | User | false |
openclaw/openclaw | 4,295,727,147 | I_kwDOQb6kR88AAAABAAuYKw | 69,331 | https://github.com/openclaw/openclaw/issues/69331 | https://api.github.com/repos/openclaw/openclaw/issues/69331 | [Feature]: Add Venice.ai as an image_generate tool provider | ### Summary
Add Venice.ai as a provider for the `image_generate` tool, alongside OpenAI, Google, fal, MiniMax, ComfyUI, and Vydra.
### Problem to solve
Venice is already a first-class **chat** provider in OpenClaw (`venice/kimi-k2-5`, `venice/claude-opus-4-6`, etc.), but Venice's image generation API is not wired in... | closed | not_planned | true | 1 | [] | [] | 2026-04-20T12:41:45Z | 2026-04-30T04:38:32Z | 2026-04-26T05:44:07Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | marclawclaw | 267,796,554 | U_kgDOD_ZASg | User | false |
openclaw/openclaw | 4,295,792,354 | I_kwDOQb6kR88AAAABAAyW4g | 69,333 | https://github.com/openclaw/openclaw/issues/69333 | https://api.github.com/repos/openclaw/openclaw/issues/69333 | Todo: improve Hubitat installed-app config discovery and handle browser-unavailable fallback | ## Summary
While investigating Hubitat installed app configuration, two concrete limitations were hit that should be tracked as follow-up work.
## Problem 1: Direct HTTP/curl fetch of Hubitat Installed Apps UI is insufficient for config discovery
**What happened**
Fetching pages like `/installedapp/list` with curl onl... | closed | not_planned | true | 2 | [] | [] | 2026-04-20T12:53:00Z | 2026-04-29T04:36:06Z | 2026-04-26T12:20:35Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | mason-loki | 272,005,004 | U_kgDOEDZ3jA | User | false |
openclaw/openclaw | 4,295,731,964 | I_kwDOQb6kR88AAAABAAuq_A | 69,332 | https://github.com/openclaw/openclaw/issues/69332 | https://api.github.com/repos/openclaw/openclaw/issues/69332 | [Bug]: AWS EC2 Linux instance fails to install on one line install. | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
I am running at Linux machine on AWS EC2 and the details are below re: architecture:
Linux.ec2.internal 6.1.166-197.305.amzn2023.x86_64 #1 SMP PREEMPT_DYNAMIC Mon Mar 23 09:53:26 UTC 2026 x86_64 x86_64 x86_64 GNU/Linux
I trie... | closed | completed | true | 2 | [
"bug",
"bug:crash"
] | [] | 2026-04-20T12:42:31Z | 2026-04-28T04:49:22Z | 2026-04-25T19:07:47Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kingkpink | 223,677,524 | U_kgDODVUMVA | User | false |
openclaw/openclaw | 4,296,048,271 | I_kwDOQb6kR88AAAABABB-jw | 69,341 | https://github.com/openclaw/openclaw/issues/69341 | https://api.github.com/repos/openclaw/openclaw/issues/69341 | channels status docs mention --deep, but the flag does not exist | On OpenClaw 2026.4.15 (041266a), `openclaw channels status` and the parent `openclaw channels` help text both mention `status --deep`, but the flag is not actually accepted by the CLI.
Repro:
```bash
openclaw channels status --help
openclaw channels status
openclaw channels status --deep
```
Observed:
- `openclaw cha... | closed | not_planned | true | 1 | [] | [] | 2026-04-20T13:29:54Z | 2026-04-29T04:37:07Z | 2026-04-26T10:50:15Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | fabianfabian | 343,640 | MDQ6VXNlcjM0MzY0MA== | User | false |
openclaw/openclaw | 4,082,292,085 | I_kwDOQb6kR87zUtV1 | 48,205 | https://github.com/openclaw/openclaw/issues/48205 | https://api.github.com/repos/openclaw/openclaw/issues/48205 | [Bug]: # Gateway Restarts Every Avg.~50 Minutes with "reason=none" | ### Bug type
Crash (process/app exits or hangs)
### Summary
1. Run OpenClaw gateway
2. Wait - restarts happen automatically without user intervention
## Frequency (March 16, 2026)
| Time (UTC) | Time (CET) | Interval |
|------------|------------|----------|
| 04:40:15 | 05:40 | ~1h 20m |
| 06:03:54 | 07:03 | ~1h 23... | closed | not_planned | true | 21 | [
"bug",
"bug:crash"
] | [] | 2026-03-16T12:49:46Z | 2026-04-23T04:22:36Z | 2026-04-20T13:36:05Z | NONE | null | 20260423T060020Z | 2026-04-23T06:00:20Z | WaMaSeDu | 265,785,875 | U_kgDOD9eSEw | User | false |
openclaw/openclaw | 3,895,287,204 | I_kwDOQb6kR87oLV2k | 8,673 | https://github.com/openclaw/openclaw/issues/8673 | https://api.github.com/repos/openclaw/openclaw/issues/8673 | Add retry logic to OAuth token refresh | ## Problem
When the OAuth token refresh API call fails transiently (network blip, API timeout), the gateway immediately throws an error without retrying. This can cause agents to fail even when the underlying issue is temporary.
### Observed behavior
```
2026-02-04T08:14:05.225Z [agents/auth-profiles] wrote refreshe... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-04T08:23:10Z | 2026-05-04T23:10:24Z | null | NONE | null | 20260505T000034Z | 2026-05-05T00:00:34Z | skyblue-will | 118,227,962 | U_kgDOBwwD-g | User | false |
openclaw/openclaw | 4,296,127,921 | I_kwDOQb6kR88AAAABABG1sQ | 69,347 | https://github.com/openclaw/openclaw/issues/69347 | https://api.github.com/repos/openclaw/openclaw/issues/69347 | [Bug]: sessions.json.estimatedCostUsd inflated 1x-72x by three persist sites that accumulate cost while snapshotting tokens | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
`sessions.json.estimatedCostUsd` is inflated 1×–72× over the real OpenRouter cost (the `usage.cost.total` field inside the JSONL transcript) because three persist sites snapshot tokens but accumulate cost.
###... | closed | completed | true | 0 | [] | [] | 2026-04-20T13:41:15Z | 2026-04-23T04:22:05Z | 2026-04-21T02:44:27Z | NONE | null | 20260423T060020Z | 2026-04-23T06:00:20Z | antongoryachev | 237,813,605 | U_kgDODiy_ZQ | User | false |
openclaw/openclaw | 4,064,291,925 | I_kwDOQb6kR87yQCxV | 44,028 | https://github.com/openclaw/openclaw/issues/44028 | https://api.github.com/repos/openclaw/openclaw/issues/44028 | shouldSyncSessions skips session data during full reindex triggered by session-start | # Bug Report: `shouldSyncSessions` skips session data during full reindex triggered by `session-start`
## Summary
When a full reindex is triggered during a `session-start` sync (e.g., due to model mismatch after gateway restart), session transcript data is silently dropped from the index. The metadata still reports `... | closed | completed | true | 8 | [] | [] | 2026-03-12T12:23:18Z | 2026-04-27T04:49:43Z | 2026-04-24T22:24:13Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | mrdeeds89 | 13,260,378 | MDQ6VXNlcjEzMjYwMzc4 | User | false |
openclaw/openclaw | 4,296,182,044 | I_kwDOQb6kR88AAAABABKJHA | 69,351 | https://github.com/openclaw/openclaw/issues/69351 | https://api.github.com/repos/openclaw/openclaw/issues/69351 | Telegram native approval handler: self-signed certificate errors | ## Problem
Telegram plugin logs repeated errors every ~1 second:
```
[telegram] connect error: self-signed certificate; if the root CA is installed locally, try running Node.js with --use-system-ca
[telegram] failed to start native approval handler: Error: self-signed certificate
```
## Root Cause
The native approv... | closed | not_planned | true | 1 | [] | [] | 2026-04-20T13:49:50Z | 2026-04-30T04:38:31Z | 2026-04-26T05:44:16Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | 73ttv627re77 | 112,801,096 | U_kgDOBrk1SA | User | false |
openclaw/openclaw | 4,166,041,041 | I_kwDOQb6kR874UL3R | 57,334 | https://github.com/openclaw/openclaw/issues/57334 | https://api.github.com/repos/openclaw/openclaw/issues/57334 | [Bug]: Memory-core session indexer skips .jsonl.reset and .jsonl.deleted files — most past sessions invisible to search | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The `listSessionFilesForAgent` function in `query-expansion-CeNhqo71.js` filters session files with `.filter((name) => name.endsWith(".jsonl"))`, which excludes `.jsonl.reset.*` and `.jsonl.deleted.*` files. After the daily ... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-03-29T23:15:21Z | 2026-04-28T05:18:45Z | 2026-04-25T12:06:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | christian-lallo | 147,324,086 | U_kgDOCMf8tg | User | false |
openclaw/openclaw | 4,296,217,641 | I_kwDOQb6kR88AAAABABMUKQ | 69,353 | https://github.com/openclaw/openclaw/issues/69353 | https://api.github.com/repos/openclaw/openclaw/issues/69353 | [Bug]: `context1m` config flag ineffective for non-`anthropic` providers | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The 1M context window path in `resolveContextTokensForModel` requires both `context1m === true` in config AND `isAnthropic1MModel()` returning true. But `isAnthropic1MModel` hardcodes `if (provider !== "anthrop... | closed | not_planned | true | 2 | [
"bug"
] | [] | 2026-04-20T13:55:28Z | 2026-04-30T04:38:30Z | 2026-04-26T05:44:21Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | laserlemon | 34,264 | MDQ6VXNlcjM0MjY0 | User | true |
openclaw/openclaw | 4,296,259,881 | I_kwDOQb6kR88AAAABABO5KQ | 69,354 | https://github.com/openclaw/openclaw/issues/69354 | https://api.github.com/repos/openclaw/openclaw/issues/69354 | Dreaming diary language unstable: sometimes English, sometimes Chinese | ## Problem
The Dreaming module generates diary entries with unstable language. Sometimes the output is in English, sometimes in Chinese, with no consistent language configuration.
## Expected Behavior
Users should be able to configure a language preference (e.g., `language: "zh"` in the dreaming config), so all diar... | closed | not_planned | true | 2 | [] | [] | 2026-04-20T14:02:32Z | 2026-04-29T04:41:25Z | 2026-04-26T08:21:56Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | zhong18804784882 | 268,618,080 | U_kgDOEALJYA | User | false |
openclaw/openclaw | 4,116,842,485 | I_kwDOQb6kR871Ygf1 | 52,441 | https://github.com/openclaw/openclaw/issues/52441 | https://api.github.com/repos/openclaw/openclaw/issues/52441 | 5分鐘防蠢決策SOP，可能對OpenClaw使用者有幫助 | ### Summary
一套極簡決策檢查表，可在重要任務前快速避坑。
### Problem to solve
我最近做了一個「5分鐘防蠢決策SOP」，專門幫開發者、團隊在重要決策前避坑。可能對 OpenClaw 的使用者有幫助，歡迎參考：
https://github.com/dabaozheng/anti-stupid-sop
### Proposed solution
提供這份 SOP 給大家參考，希望能幫助 OpenClaw 使用者在重要決策前快速避坑。
https://github.com/dabaozheng/anti-stupid-sop
### Alternatives considered
_No... | closed | completed | true | 0 | [
"enhancement"
] | [] | 2026-03-22T19:06:22Z | 2026-04-23T04:22:35Z | 2026-04-20T14:07:24Z | NONE | null | 20260423T060020Z | 2026-04-23T06:00:20Z | dabaozheng | 269,720,734 | U_kgDOEBOcng | User | false |
openclaw/openclaw | 4,296,292,643 | I_kwDOQb6kR88AAAABABQ5Iw | 69,358 | https://github.com/openclaw/openclaw/issues/69358 | https://api.github.com/repos/openclaw/openclaw/issues/69358 | [Bug] Control-UI channel: pure image messages fail with "I didn't receive any text" | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
When sending a pure image (without text caption) to OpenClaw via control-ui/webchat channel, the agent always responds with "I didn't receive any text in your message. Please resend or add a caption." The image is never proc... | closed | completed | true | 0 | [
"bug"
] | [] | 2026-04-20T14:07:54Z | 2026-04-24T04:26:34Z | 2026-04-21T05:04:35Z | NONE | null | 20260424T060021Z | 2026-04-24T06:00:21Z | batuhanforras903-stack | 276,621,764 | U_kgDOEHzpxA | User | false |
openclaw/openclaw | 4,296,380,134 | I_kwDOQb6kR88AAAABABWO5g | 69,359 | https://github.com/openclaw/openclaw/issues/69359 | https://api.github.com/repos/openclaw/openclaw/issues/69359 | Webchat image upload: sends CDN URL instead of file, image inaccessible to model | ## Bug type
Behavior bug (incorrect output/state without crash)
## Beta release blocker
Yes (images cannot be analyzed)
## Summary
In beta.2, when a user sends an image in webchat (via Ctrl+V paste or drag-and-drop), the webchat uploads it to MiniMax CDN and sends the CDN URL to the model instead of the actual ima... | closed | completed | true | 1 | [] | [] | 2026-04-20T14:21:26Z | 2026-04-28T04:49:20Z | 2026-04-25T19:08:14Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zhong18804784882 | 268,618,080 | U_kgDOEALJYA | User | false |
openclaw/openclaw | 4,296,383,136 | I_kwDOQb6kR88AAAABABWaoA | 69,360 | https://github.com/openclaw/openclaw/issues/69360 | https://api.github.com/repos/openclaw/openclaw/issues/69360 | [Bug]: isolated cron agentTurn returns "Agent couldn't generate a response" despite successful execution on 2026.4.15 (Google Gemini) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On OpenClaw 2026.4.15, isolated cron jobs with payload.kind: "agentTurn" intermittently fail with Agent couldn't generate a response even when the underlying script executes successfully and the agent produces ... | closed | duplicate | true | 3 | [
"bug",
"close:duplicate",
"dedupe:child",
"bug:behavior"
] | [] | 2026-04-20T14:21:51Z | 2026-04-27T04:42:53Z | 2026-04-25T01:20:04Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | fenglanhua | 257,674,093 | U_kgDOD1vLbQ | User | false |
openclaw/openclaw | 4,227,538,167 | I_kwDOQb6kR877-xz3 | 63,357 | https://github.com/openclaw/openclaw/issues/63357 | https://api.github.com/repos/openclaw/openclaw/issues/63357 | Agent runtime per-turn startup overhead is ~17s on CPU-only systems, dominating end-to-end latency for fast cloud models | ## Bug: Agent runtime per-turn startup overhead is ~17s on CPU-only systems, dominating end-to-end latency for fast cloud models
### Summary
On a CPU-only host, every agent turn pays a ~17 second per-turn initialization cost between the channel adapter receiving the inbound message and the user message being written ... | closed | completed | true | 3 | [] | [] | 2026-04-08T20:44:43Z | 2026-04-29T04:33:41Z | 2026-04-26T04:39:27Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | tmote | 21,976,980 | MDQ6VXNlcjIxOTc2OTgw | User | false |
openclaw/openclaw | 4,046,706,754 | I_kwDOQb6kR87xM9hC | 41,278 | https://github.com/openclaw/openclaw/issues/41278 | https://api.github.com/repos/openclaw/openclaw/issues/41278 | [Bug]: update to 3.8 got error | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
npm error code 128
npm error An unknown git error occurred
npm error command git --no-replace-objects ls-remote ssh://git@github.com/whiskeysockets/libsignal-node.git
npm error git@github.com: Permission denied (publickey).
... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-03-09T16:49:30Z | 2026-04-23T04:22:31Z | 2026-04-20T14:37:45Z | NONE | null | 20260423T060020Z | 2026-04-23T06:00:20Z | lukehzg | 11,779,804 | MDQ6VXNlcjExNzc5ODA0 | User | false |
openclaw/openclaw | 4,222,348,115 | I_kwDOQb6kR877q-tT | 62,952 | https://github.com/openclaw/openclaw/issues/62952 | https://api.github.com/repos/openclaw/openclaw/issues/62952 | postinstall-bundled-plugins fails silently without SSH client, leaving runtime deps missing | ## Description
`postinstall-bundled-plugins.mjs` fails silently when the container image lacks an SSH client (`openssh-client`). Some transitive dependencies use `git+ssh://` URLs (e.g., `whiskeysockets/libsignal-node`). When SSH is unavailable, the postinstall catches the error and continues, leaving critical runtime... | closed | completed | true | 1 | [] | [] | 2026-04-08T05:32:17Z | 2026-04-23T04:22:30Z | 2026-04-20T14:37:47Z | NONE | null | 20260423T060020Z | 2026-04-23T06:00:20Z | achandmsft | 14,188,700 | MDQ6VXNlcjE0MTg4NzAw | User | false |
openclaw/openclaw | 4,296,518,225 | I_kwDOQb6kR88AAAABABeqUQ | 69,364 | https://github.com/openclaw/openclaw/issues/69364 | https://api.github.com/repos/openclaw/openclaw/issues/69364 | [Feature]: Control UI: no cross-agent view when configuring or debugging multi-agent teams | ### Summary
Control UI presents agents as a flat list. For teams coordinating via `sessions_spawn` + `shared-workspace` symlinks + `TASK.json`, reasoning about any single cross-agent relationship currently requires opening **5–10 files across 5 directories**, because the UI surfaces each file individually but never th... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-20T14:42:48Z | 2026-05-02T16:01:49Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | ccc330 | 56,498,783 | MDQ6VXNlcjU2NDk4Nzgz | User | false |
openclaw/openclaw | 4,036,574,625 | I_kwDOQb6kR87wmT2h | 38,387 | https://github.com/openclaw/openclaw/issues/38387 | https://api.github.com/repos/openclaw/openclaw/issues/38387 | Deprecated transitive dependencies and outdated npm in Docker build | During Docker build (Step 6/14), `npm i -g openclaw` produces multiple deprecation warnings for transitive dependencies and an npm self-update notice.
**Deprecated dependencies:**
- `inflight@1.0.6` — memory leak, replace with `lru-cache`
- `are-we-there-yet@2.0.0` — no longer supported
- `gauge@3.0.2` — no longer sup... | closed | completed | true | 2 | [] | [] | 2026-03-06T23:10:18Z | 2026-04-23T04:22:28Z | 2026-04-20T14:45:00Z | NONE | null | 20260423T060020Z | 2026-04-23T06:00:20Z | serdarsemen | 3,353,767 | MDQ6VXNlcjMzNTM3Njc= | User | false |
openclaw/openclaw | 4,295,520,545 | I_kwDOQb6kR88AAAABAAhxIQ | 69,320 | https://github.com/openclaw/openclaw/issues/69320 | https://api.github.com/repos/openclaw/openclaw/issues/69320 | [Bug] Dreaming cron hits hard-coded 10-minute timeout; no config override for systemEvent jobs | ## Summary
The Dreaming cron (`Memory Dreaming Promotion`, managed by `memory-core.short-term-promotion`) hits the hard-coded 600s (10-minute) cron runner timeout and gets marked `error` even though the promotion work itself completes successfully. There's no way to raise this timeout for `systemEvent` payloads.
## E... | closed | completed | true | 7 | [] | [] | 2026-04-20T12:08:27Z | 2026-04-28T04:49:21Z | 2026-04-25T19:07:28Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | bradfreels | 179,389,909 | U_kgDOCrFF1Q | User | false |
openclaw/openclaw | 4,162,136,643 | I_kwDOQb6kR874FSpD | 56,692 | https://github.com/openclaw/openclaw/issues/56692 | https://api.github.com/repos/openclaw/openclaw/issues/56692 | Group chat context handling can blur which messages are addressed to the current agent | ## Summary
In Telegram group chats with multiple agents participating, the current agent can sometimes mis-handle context and respond to messages that were actually directed at a different agent. This creates confusion and causes the assistant to pick up the wrong sub-thread.
## Problem
When several assistants are p... | open | null | false | 5 | [] | [] | 2026-03-29T01:18:47Z | 2026-05-03T07:44:33Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Svtter | 5,856,931 | MDQ6VXNlcjU4NTY5MzE= | User | false |
openclaw/openclaw | 4,296,473,106 | I_kwDOQb6kR88AAAABABb6Eg | 69,361 | https://github.com/openclaw/openclaw/issues/69361 | https://api.github.com/repos/openclaw/openclaw/issues/69361 | channels: remove core duplicate channel schemas in zod-schema.providers-core.ts (post-#69193) | ## Context
PR #69193 added `channels.bluebubbles.sendTimeoutMs` to the **plugin-owned** schema at `extensions/bluebubbles/src/config-schema.ts`. End-to-end testing on a live install (Lobster) revealed the field was rejected at runtime by `openclaw config set` with:
```
Error: Config validation failed: channels.bluebu... | open | null | false | 0 | [
"maintainer"
] | [
"omarshahine"
] | 2026-04-20T14:35:43Z | 2026-04-20T14:49:21Z | null | CONTRIBUTOR | null | 20260420T180040Z | 2026-04-20T18:00:40Z | omarshahine | 10,343,873 | MDQ6VXNlcjEwMzQzODcz | User | false |
openclaw/openclaw | 4,296,087,847 | I_kwDOQb6kR88AAAABABEZJw | 69,343 | https://github.com/openclaw/openclaw/issues/69343 | https://api.github.com/repos/openclaw/openclaw/issues/69343 | [Bug]: openclaw message thread create --channel telegram fails with Unsupported Telegram action: thread-create | ## Summary
Telegram forum topic creation is supported by the Telegram action runtime (`topic-create` exists), but the generic CLI command:
- `openclaw message thread create --channel telegram ...`
still dispatches `thread-create` instead of translating it into Telegram's native `topic-create` action.
Result:
- Tele... | closed | completed | true | 2 | [] | [] | 2026-04-20T13:35:33Z | 2026-04-28T04:49:22Z | 2026-04-25T19:07:54Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jarvisz8 | 260,156,352 | U_kgDOD4GrwA | User | false |
openclaw/openclaw | 4,296,712,096 | I_kwDOQb6kR88AAAABABqfoA | 69,368 | https://github.com/openclaw/openclaw/issues/69368 | https://api.github.com/repos/openclaw/openclaw/issues/69368 | Model fallback chain not attempted on HTTP 502 / "terminated" / infrastructure errors from primary provider | ## Summary
When a primary provider fails with a raw HTTP infrastructure error — 502 Bad Gateway, connection `terminated`, etc. — the configured fallback chain is **never attempted**. The error propagates up as a bare `Request failed` and, on cron lanes, surfaces to the channel as a failure notification. This is exactl... | closed | completed | true | 0 | [] | [] | 2026-04-20T15:14:06Z | 2026-04-26T04:30:06Z | 2026-04-23T16:58:21Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | chrispatil | 116,297,895 | U_kgDOBu6Qpw | User | false |
openclaw/openclaw | 4,296,728,131 | I_kwDOQb6kR88AAAABABreQw | 69,369 | https://github.com/openclaw/openclaw/issues/69369 | https://api.github.com/repos/openclaw/openclaw/issues/69369 | [Bug]: Regression of #9348: file tools still reject custom sandbox bind mounts even though exec and Docker can access them | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
This looks like a regression, or at least an incomplete recurrence, of [#9348](https://github.com/
openclaw/openclaw/issues/9348).
In my current setup, a custom sandbox Docker bind mount is definitely present in the liv... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-20T15:16:34Z | 2026-04-28T04:49:19Z | 2026-04-25T19:08:28Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | skylarbpayne | 4,830,598 | MDQ6VXNlcjQ4MzA1OTg= | User | false |
openclaw/openclaw | 4,296,789,181 | I_kwDOQb6kR88AAAABABvMvQ | 69,371 | https://github.com/openclaw/openclaw/issues/69371 | https://api.github.com/repos/openclaw/openclaw/issues/69371 | [Enhancement] Improve stability on ARM64 edge devices (Raspberry Pi 5 / Rock5) | ## Problem Summary
Running OpenClaw on ARM64 single-board computers (e.g., Raspberry Pi 5, Rock5) reveals several stability and performance issues that significantly degrade the experience on resource-constrained devices.
---
## Observed Issues
### 1. CLI commands timeout on ARM64 (critical)
**Impact:** `openclaw s... | open | null | false | 1 | [] | [] | 2026-04-20T15:25:44Z | 2026-05-02T16:41:05Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | shinfon | 92,320,216 | U_kgDOBYCx2A | User | false |
openclaw/openclaw | 4,295,642,896 | I_kwDOQb6kR88AAAABAApPEA | 69,327 | https://github.com/openclaw/openclaw/issues/69327 | https://api.github.com/repos/openclaw/openclaw/issues/69327 | [Bug]: Subagent sandbox does not propagate sandbox.docker.env and may reuse stale workspace/bootstrap state after recreate | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Per-agent sandbox configuration appears correct in ~/.openclaw/openclaw.json, but fresh sandboxed subagent runs do not reliably receive:
- sandbox.docker.env variables
- current workspace file state
- expected ... | open | null | false | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-20T12:28:21Z | 2026-05-02T16:55:37Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | mozdyn | 58,273,958 | MDQ6VXNlcjU4MjczOTU4 | User | false |
openclaw/openclaw | 4,068,301,419 | I_kwDOQb6kR87yfVpr | 44,533 | https://github.com/openclaw/openclaw/issues/44533 | https://api.github.com/repos/openclaw/openclaw/issues/44533 | [BUG] delivery.mode = "none" incorrectly reports "not-delivered" instead of "not-requested" | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When `delivery.mode = "none"` is set in a cron job config, the system incorrectly reports `lastDeliveryStatus: "not-delivered"` instead of `"not-requested"`.
### Steps to reproduce
1. Create a cron job with `delivery: { mode: "none" }`
2.... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-03-13T01:59:42Z | 2026-04-27T04:48:08Z | 2026-04-24T23:29:04Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | WaMaSeDu | 265,785,875 | U_kgDOD9eSEw | User | false |
openclaw/openclaw | 4,209,795,037 | I_kwDOQb6kR8767F_d | 61,701 | https://github.com/openclaw/openclaw/issues/61701 | https://api.github.com/repos/openclaw/openclaw/issues/61701 | v2026.4.5: gateway process 100% CPU after upgrade from v2026.4.2 | ## Description
After upgrading from v2026.4.2 to v2026.4.5, the gateway process immediately consumes 100%+ CPU and never settles. Rolling back to v2026.4.2 resolves the issue — CPU drops to ~1% after startup.
## Environment
- Host: Ubuntu 24.04, x86_64, HP Pavilion laptop (NAS)
- Docker: containerized deployment (se... | closed | completed | true | 8 | [] | [] | 2026-04-06T05:36:45Z | 2026-04-30T04:30:17Z | 2026-04-28T02:25:33Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | ENAwareness | 166,632,095 | U_kgDOCe6anw | User | false |
openclaw/openclaw | 4,296,861,345 | I_kwDOQb6kR88AAAABABzmoQ | 69,374 | https://github.com/openclaw/openclaw/issues/69374 | https://api.github.com/repos/openclaw/openclaw/issues/69374 | Expose WhatsApp forwardingScore / isForwarded to plugin inbound_claim hook | ### Summary
The WhatsApp adapter already extracts `contextInfo` from inbound Baileys envelopes, but the forwarding fields (`forwardingScore`, `isForwarded`) are dropped during normalization and never reach plugins. This makes it impossible for a plugin to distinguish a user's own words from a message they forwarded fr... | open | null | false | 1 | [] | [] | 2026-04-20T15:36:05Z | 2026-05-02T13:08:04Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | omerekin | 36,278,311 | MDQ6VXNlcjM2Mjc4MzEx | User | false |
openclaw/openclaw | 4,035,904,109 | I_kwDOQb6kR87wjwJt | 38,300 | https://github.com/openclaw/openclaw/issues/38300 | https://api.github.com/repos/openclaw/openclaw/issues/38300 | Subagent completion announce silently downgrades to channel-only, parent agent never continues | ## Summary
When a subagent spawned via `sessions_spawn` (mode="run") completes, the announce mechanism is supposed to inject the completion back into the parent agent session, triggering a continuation turn. In practice, the direct-announce path frequently fails under load, and the fallback silently posts the result t... | closed | completed | true | 4 | [] | [
"vincentkoc"
] | 2026-03-06T19:54:23Z | 2026-04-27T05:01:50Z | 2026-04-24T19:32:24Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | henryclaw007-lgtm | 260,290,683 | U_kgDOD4O4ew | User | false |
openclaw/openclaw | 4,166,355,571 | I_kwDOQb6kR874VYpz | 57,385 | https://github.com/openclaw/openclaw/issues/57385 | https://api.github.com/repos/openclaw/openclaw/issues/57385 | Sub-agent completion announces don't route to originating Telegram session | ## Bug
When a sub-agent is spawned from a main session that's actively receiving messages via Telegram, the completion announce routes to the webchat/desktop surface instead of the Telegram channel where the conversation is happening.
### Expected behavior
Sub-agent completion events should announce back to the same ... | closed | completed | true | 3 | [] | [] | 2026-03-30T01:20:51Z | 2026-04-28T05:18:38Z | 2026-04-25T12:07:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Adam-Researchh | 261,827,313 | U_kgDOD5sq8Q | User | false |
openclaw/openclaw | 4,225,275,042 | I_kwDOQb6kR8772JSi | 63,188 | https://github.com/openclaw/openclaw/issues/63188 | https://api.github.com/repos/openclaw/openclaw/issues/63188 | [Feature]: Agent Runtime: Configurable stopReason="length" catch-and-continue for mode:"run" sessions | ### Summary
When a subagent running in `mode: "run"` hits the provider's maximum output token limit mid-generation, OpenClaw receives `stopReason: "length"` from the API but the agent session is immediately terminated. The agent never sees the error, cannot recover, and leaves an incomplete output on disk with no sign... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-08T14:25:31Z | 2026-05-02T12:18:14Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | AndrejsLiepinieks | 37,855,598 | MDQ6VXNlcjM3ODU1NTk4 | User | false |
openclaw/openclaw | 4,225,579,654 | I_kwDOQb6kR8773TqG | 63,210 | https://github.com/openclaw/openclaw/issues/63210 | https://api.github.com/repos/openclaw/openclaw/issues/63210 | [Feature]: Detect and recover from output truncation (stopReason:"length") in main agent sessions | ### Summary
When the model hits its `max_output_tokens` limit mid-generation (especially during tool_use with large parameters), the agent runtime silently proceeds with truncated output. There is no detection, no recovery, and no signal to the LLM that its output was incomplete.
This is a complementary request to #6... | open | null | false | 2 | [] | [] | 2026-04-08T15:13:35Z | 2026-05-03T03:22:34Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | yilong016 | 120,642,887 | U_kgDOBzDdRw | User | false |
openclaw/openclaw | 4,231,477,146 | I_kwDOQb6kR878Nzea | 63,689 | https://github.com/openclaw/openclaw/issues/63689 | https://api.github.com/repos/openclaw/openclaw/issues/63689 | assistant returns content=[] with usage=0 causing silent no-reply to user | ## Problem Description
When the model returns a valid stop stopReason but produces an empty assistant message (content: [] or content: "" with usage.totalTokens === 0), OpenClaw currently forwards nothing to the channel. The user experiences a complete silence with no indication anything happened.
## Environment
- O... | closed | completed | true | 3 | [] | [] | 2026-04-09T10:36:01Z | 2026-04-28T05:02:59Z | 2026-04-25T15:45:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | xion516 | 187,609,864 | U_kgDOCy6zCA | User | false |
openclaw/openclaw | 4,296,924,494 | I_kwDOQb6kR88AAAABAB3dTg | 69,380 | https://github.com/openclaw/openclaw/issues/69380 | https://api.github.com/repos/openclaw/openclaw/issues/69380 | Image tool fails with reasoning-capable vision models that return empty content | ## Summary
The OpenClaw image tool fails with `"Image model returned no text"` when using any OpenAI-compatible reasoning-capable vision model that returns its response in `reasoning_content` instead of `content`.
## Environment
- OpenClaw version: latest (container)
- Provider: Custom OpenAI-compatible provider
- M... | closed | completed | true | 0 | [] | [] | 2026-04-20T15:46:15Z | 2026-04-23T04:21:57Z | 2026-04-21T04:20:24Z | NONE | null | 20260423T060020Z | 2026-04-23T06:00:20Z | haseo-ai | 265,708,334 | U_kgDOD9ZjLg | User | false |
openclaw/openclaw | 4,297,005,360 | I_kwDOQb6kR88AAAABAB8ZMA | 69,384 | https://github.com/openclaw/openclaw/issues/69384 | https://api.github.com/repos/openclaw/openclaw/issues/69384 | [Bug]: Assistant can answer a previous user question after a long multi-tool turn | ## Summary
In a live chat, after a long multi-tool turn, the assistant can resume the previously active reasoning branch and answer an older user question instead of the newest incoming one.
This seems especially likely when:
- the assistant is doing long `exec` / `process poll` work,
- new user messages arrive while ... | closed | not_planned | true | 1 | [] | [] | 2026-04-20T15:59:35Z | 2026-04-28T04:49:19Z | 2026-04-25T19:08:35Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | gchahrouri | 37,080,676 | MDQ6VXNlcjM3MDgwNjc2 | User | false |
openclaw/openclaw | 4,297,018,430 | I_kwDOQb6kR88AAAABAB9MPg | 69,386 | https://github.com/openclaw/openclaw/issues/69386 | https://api.github.com/repos/openclaw/openclaw/issues/69386 | CRITICAL: Agent ignores stop commands during exec denial - infinite retry loop, delayed command flood, and queue overflow | ## Bug Report: Agent Ignores Stop Commands During Exec Approval Loop
### Severity
🔴 Critical
### Environment
- **OpenClaw Version**: 2026.4.15
- **Platform**: Windows 11 (NT 10.0.26200, x64)
- **Node**: v25.9.0
- **Model**: alibailian/qwen3.6-plus
- **Channel**: webchat (Control UI)
- **Agent**: trade
---
### Repr... | open | null | false | 1 | [] | [] | 2026-04-20T16:01:46Z | 2026-05-02T16:41:06Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | xiaobu1112 | 267,266,960 | U_kgDOD-4rkA | User | false |
openclaw/openclaw | 4,297,116,357 | I_kwDOQb6kR88AAAABACDKxQ | 69,390 | https://github.com/openclaw/openclaw/issues/69390 | https://api.github.com/repos/openclaw/openclaw/issues/69390 | [Bug]: Ollama local model still hits ~5 minute "Headers Timeout Error" on 2026.4.15 despite increased OpenClaw timeouts | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
When running OpenClaw against a slow local Ollama model, the request still appears to die at about 5 minutes with:
`fetch failed | Headers Timeout Error`
This happens even after increasing the OpenClaw timeout, and on a versi... | closed | completed | true | 2 | [
"bug",
"close:duplicate",
"dedupe:child"
] | [
"vincentkoc"
] | 2026-04-20T16:17:24Z | 2026-04-30T04:44:17Z | 2026-04-25T07:52:22Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | vnitu02 | 30,450,325 | MDQ6VXNlcjMwNDUwMzI1 | User | false |
openclaw/openclaw | 4,041,725,873 | I_kwDOQb6kR87w59ex | 40,089 | https://github.com/openclaw/openclaw/issues/40089 | https://api.github.com/repos/openclaw/openclaw/issues/40089 | Gateway restart (update.run) can leave service dead: bootout without bootstrap | ## Summary
After `update.run` triggers a gateway restart, the service can end up in a broken state where `launchctl bootout` runs successfully but `launchctl bootstrap` never executes — leaving the gateway permanently dead until manual intervention.
## Environment
- macOS 15 (arm64)
- LaunchAgent with `KeepAlive: tr... | closed | not_planned | true | 9 | [] | [] | 2026-03-08T17:07:06Z | 2026-04-28T04:42:35Z | 2026-04-26T02:42:45Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | gandalf-at-lerian | 264,490,208 | U_kgDOD8PM4A | User | false |
openclaw/openclaw | 4,217,620,704 | I_kwDOQb6kR877Y8jg | 62,478 | https://github.com/openclaw/openclaw/issues/62478 | https://api.github.com/repos/openclaw/openclaw/issues/62478 | [Feature]: Background sub-agent spawn mode — parent session completes immediately while children run async | ## Summary
Add a sub-agent spawn mode where the **parent's run completes immediately** after dispatching children, leaving the parent session free to accept new user messages while sub-agents work in the background. Results are delivered asynchronously via injected messages, notifications, or kanban updates.
## Probl... | closed | completed | true | 2 | [] | [] | 2026-04-07T12:24:35Z | 2026-04-28T05:06:07Z | 2026-04-25T14:37:13Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | bizmindx | 28,684,645 | MDQ6VXNlcjI4Njg0NjQ1 | User | false |
openclaw/openclaw | 4,297,198,910 | I_kwDOQb6kR88AAAABACINPg | 69,396 | https://github.com/openclaw/openclaw/issues/69396 | https://api.github.com/repos/openclaw/openclaw/issues/69396 | config.write clobbers symlinked openclaw.json and serializes resolved SecretRef plaintext to disk | ## Summary
`config.write` in the gateway has two related bugs that together cause plaintext secrets to leak into `~/.openclaw/openclaw.json` on every config touch:
1. **Atomic rename replaces symlinked config with a regular file.** If `~/.openclaw/openclaw.json` is a symlink (e.g. to a git-tracked workspace file), ev... | closed | completed | true | 2 | [] | [] | 2026-04-20T16:31:02Z | 2026-04-25T04:16:18Z | 2026-04-22T22:45:12Z | NONE | null | 20260425T060021Z | 2026-04-25T06:00:21Z | quintonm | 991,413 | MDQ6VXNlcjk5MTQxMw== | User | false |
openclaw/openclaw | 4,297,163,806 | I_kwDOQb6kR88AAAABACGEHg | 69,392 | https://github.com/openclaw/openclaw/issues/69392 | https://api.github.com/repos/openclaw/openclaw/issues/69392 | [bug] acpx runtime: persistent ACP client not returned to pool after runTurn, causing per-send cold-start and making acp.runtime.ttlMinutes a dead config | # [bug] acpx runtime: persistent ACP client not returned to pool after `runTurn`, causing per-send cold-start and making `acp.runtime.ttlMinutes` a dead config
## Version
- openclaw `2026.4.15` (npm global install)
- Observed with `claude-agent-acp` and `gemini --acp` adapters; reproducible across all ACP agents
## ... | closed | not_planned | true | 1 | [] | [] | 2026-04-20T16:25:14Z | 2026-04-24T04:26:14Z | 2026-04-21T08:57:21Z | NONE | null | 20260424T060021Z | 2026-04-24T06:00:21Z | Sway-Chan | 15,033,876 | MDQ6VXNlcjE1MDMzODc2 | User | false |
openclaw/openclaw | 4,297,269,969 | I_kwDOQb6kR88AAAABACMi0Q | 69,398 | https://github.com/openclaw/openclaw/issues/69398 | https://api.github.com/repos/openclaw/openclaw/issues/69398 | [Bug]: image tool fails in 2026.4.15 — broken ajv import path in @mariozechner/pi-ai | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
image tool fails on all models due to broken ajv import path in @mariozechner/pi-ai
### Steps to reproduce
1. Install OpenClaw 2026.4.15
2. Use the image tool (send any image to the assistant)
3. Observe the error in too... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-20T16:44:25Z | 2026-04-23T04:22:17Z | 2026-04-20T22:23:36Z | NONE | null | 20260423T060020Z | 2026-04-23T06:00:20Z | jtyun | 16,197,533 | MDQ6VXNlcjE2MTk3NTMz | User | false |
openclaw/openclaw | 4,296,169,211 | I_kwDOQb6kR88AAAABABJW-w | 69,350 | https://github.com/openclaw/openclaw/issues/69350 | https://api.github.com/repos/openclaw/openclaw/issues/69350 | [Bug]: OpenRouter API key is configured and validated (works with curl), but OpenClaw returns empty responses when using openrouter/google/gemma-4-26b-a4b-it | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Description:
OpenRouter API key is configured and validated (works with curl), but OpenClaw returns empty responses when using openrouter/google/gemma-4-26b-a4b-it.
Setup:
OpenRouter API key: ✅ valid (tested with curl)
openc... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-20T13:47:33Z | 2026-04-30T04:38:32Z | 2026-04-26T05:44:11Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | Autom8ted54 | 268,345,669 | U_kgDOD_6hRQ | User | false |
openclaw/openclaw | 4,297,397,474 | I_kwDOQb6kR88AAAABACUU4g | 69,400 | https://github.com/openclaw/openclaw/issues/69400 | https://api.github.com/repos/openclaw/openclaw/issues/69400 | Feature: honor OPENCLAW_WRAPPER (or --wrapper) when generating LaunchAgent ProgramArguments | ## Problem
When users wrap the openclaw gateway in a shell script (commonly to inject environment from a secrets manager like Doppler/SOPS/1Password), they have to manually re-point `~/Library/LaunchAgents/ai.openclaw.gateway.plist` at the wrapper after every install. `openclaw service install` (and the equivalent pat... | closed | completed | true | 2 | [] | [] | 2026-04-20T17:08:28Z | 2026-04-29T04:32:18Z | 2026-04-27T02:40:34Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | willtmc | 6,836,068 | MDQ6VXNlcjY4MzYwNjg= | User | false |
openclaw/openclaw | 4,297,422,089 | I_kwDOQb6kR88AAAABACV1CQ | 69,401 | https://github.com/openclaw/openclaw/issues/69401 | https://api.github.com/repos/openclaw/openclaw/issues/69401 | Feature Request: Untrusted content wrapping for inbound channel messages | ## Summary
Inbound messages from channels (Slack, Telegram, etc.) arrive in the agent context without structural separation between message content and system instructions. While sender metadata is correctly marked as "untrusted metadata," the message body itself is presented as raw text that the model processes along... | open | null | false | 2 | [] | [] | 2026-04-20T17:13:26Z | 2026-05-02T16:41:08Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | msanguily | 246,621,548 | U_kgDODrMlbA | User | false |
openclaw/openclaw | 4,297,488,770 | I_kwDOQb6kR88AAAABACZ5gg | 69,406 | https://github.com/openclaw/openclaw/issues/69406 | https://api.github.com/repos/openclaw/openclaw/issues/69406 | [Bug]: Chat history disappears from control UI on re-render — only most recent message reappears -- still having this issue 4.24.26 | ## Summary
Chat history periodically vanishes from the webchat UI. After the history clears, only the most recent message reappears. All prior messages are gone until page reload. Backend session is unaffected — messages are received and processed correctly.
## Steps to Reproduce
1. Open chat at `127.0.0.1:18789`
2.... | closed | completed | true | 5 | [] | [] | 2026-04-20T17:26:19Z | 2026-04-28T04:49:17Z | 2026-04-25T19:08:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | KMMossy | 270,393,824 | U_kgDOEB3h4A | User | false |
openclaw/openclaw | 4,297,490,901 | I_kwDOQb6kR88AAAABACaB1Q | 69,407 | https://github.com/openclaw/openclaw/issues/69407 | https://api.github.com/repos/openclaw/openclaw/issues/69407 | [tools] browser: add subagent-side timeout so lane is not held when browser hangs | **Version:** v2026.4.19-beta.1
**Symptom**
When `[tools] browser` hangs (network, page crash, Playwright internal stall), the subagent that called it keeps the execution lane claimed indefinitely. On a single-lane agent workspace, this blocks all subsequent turns for the duration of the hang. Observed lane-wait spike... | closed | completed | true | 2 | [] | [] | 2026-04-20T17:26:42Z | 2026-04-28T04:49:16Z | 2026-04-25T19:08:54Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | cjplanted | 176,929,257 | U_kgDOCou56Q | User | false |
openclaw/openclaw | 4,297,491,111 | I_kwDOQb6kR88AAAABACaCpw | 69,408 | https://github.com/openclaw/openclaw/issues/69408 | https://api.github.com/repos/openclaw/openclaw/issues/69408 | [agents/loop-detection]: downgrade Loop warning log level to info for isolated sessions with lifetime <= cron interval | **Version:** v2026.4.19-beta.1
**Symptom**
`[agents/loop-detection] Loop warning: exec called Ntimes` fills `gateway.err.log` with ~50 warnings per 48h, roughly 4 per 15min cron tick. The warnings come from isolated-session cron jobs (each run is a fresh ephemeral session), not from stale or polluted interactive chan... | open | null | false | 1 | [] | [] | 2026-04-20T17:26:43Z | 2026-05-02T11:36:26Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | cjplanted | 176,929,257 | U_kgDOCou56Q | User | false |
openclaw/openclaw | 4,297,498,414 | I_kwDOQb6kR88AAAABACafLg | 69,409 | https://github.com/openclaw/openclaw/issues/69409 | https://api.github.com/repos/openclaw/openclaw/issues/69409 | Feature request: Approval gate for restart-required config changes | ## Problem
When a restart-required config field (e.g. `plugins.allow`) is modified via direct file write (not the `config.patch` API), the gateway auto-detects the change and executes a restart without any human approval gate.
This is a problem in environments where:
- The config file is synced via Dropbox/iCloud/etc... | closed | completed | true | 2 | [] | [] | 2026-04-20T17:27:59Z | 2026-04-29T04:33:26Z | 2026-04-27T00:01:57Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | BennyB90 | 198,494,167 | U_kgDOC9TH1w | User | false |
openclaw/openclaw | 4,297,500,877 | I_kwDOQb6kR88AAAABACaozQ | 69,410 | https://github.com/openclaw/openclaw/issues/69410 | https://api.github.com/repos/openclaw/openclaw/issues/69410 | [Bug]: Custom OpenAI-compatible provider returns valid assistant content, but embedded agent ends with `payloads=0` / "Agent couldn't generate a response" | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
With a custom OpenAI-compatible provider, OpenClaw receives a valid assistant response from the backend, but the embedded agent still ends the turn with payloads=0 and shows “Agent couldn't generate a response.”
### Steps t... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-20T17:28:25Z | 2026-04-28T04:49:16Z | 2026-04-25T19:09:01Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Greensports4810 | 104,763,699 | U_kgDOBj6RMw | User | false |
openclaw/openclaw | 4,297,203,984 | I_kwDOQb6kR88AAAABACIhEA | 69,397 | https://github.com/openclaw/openclaw/issues/69397 | https://api.github.com/repos/openclaw/openclaw/issues/69397 | Internal tool (cron) misidentified as remote client, causing unresolvable 'pairing required' on scope-upgrade | ## Summary
Internal tools (e.g. `cron`) connecting from localhost with valid token auth are classified as `remote` clients, triggering a scope-upgrade pairing flow that cannot be resolved through the Web UI since the Control UI has no pairing approval interface for scope-upgrade requests.
## Environment
- **OpenClaw... | closed | completed | true | 2 | [] | [] | 2026-04-20T16:31:58Z | 2026-04-23T04:22:10Z | 2026-04-21T02:10:36Z | NONE | null | 20260423T060020Z | 2026-04-23T06:00:20Z | bailang-ycl | 188,980,250 | U_kgDOC0OcGg | User | false |
openclaw/openclaw | 4,297,570,723 | I_kwDOQb6kR88AAAABACe5ow | 69,412 | https://github.com/openclaw/openclaw/issues/69412 | https://api.github.com/repos/openclaw/openclaw/issues/69412 | [Bug]: `openclaw update` runs full npm install even when already on latest version | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
`openclaw update` does not check whether the installed version matches the registry version before invoking the package manager. This causes unnecessary work, wastes time, and can be dangerous on resource-const... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-20T17:41:30Z | 2026-04-25T04:16:38Z | 2026-04-22T21:03:31Z | NONE | null | 20260425T060021Z | 2026-04-25T06:00:21Z | Inhum | 75,725,438 | MDQ6VXNlcjc1NzI1NDM4 | User | false |
openclaw/openclaw | 4,297,474,897 | I_kwDOQb6kR88AAAABACZDUQ | 69,405 | https://github.com/openclaw/openclaw/issues/69405 | https://api.github.com/repos/openclaw/openclaw/issues/69405 | [Bug]: Image attachment send silently blocked after first successful image message in control UI | ## Summary
After sending one image attachment successfully in the webchat, all subsequent image sends are silently dropped — no console error, no network request, nothing reaches the gateway. Text-only messages continue to work fine.
## Steps to Reproduce
1. Open chat at `127.0.0.1:18789`
2. Attach an image and send... | closed | completed | true | 2 | [] | [] | 2026-04-20T17:23:46Z | 2026-04-28T04:49:18Z | 2026-04-25T19:08:41Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | KMMossy | 270,393,824 | U_kgDOEB3h4A | User | false |
openclaw/openclaw | 4,297,616,350 | I_kwDOQb6kR88AAAABAChr3g | 69,414 | https://github.com/openclaw/openclaw/issues/69414 | https://api.github.com/repos/openclaw/openclaw/issues/69414 | [Bug]: models auth login --provider openai-codex writes to :email profile instead of updating :default, and clobbers agents.defaults.models block (v2026.4.15) | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
## Bug type
Behavior bug (incorrect output/state without crash) + Data loss
## Version / Environment
- OpenClaw: 2026.4.15 (container: openclaw-custom:latest, base ghcr.io/openclaw/openclaw:2026.4.15@sha256:9d5f1df...)
- OS: U... | closed | completed | true | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-20T17:49:37Z | 2026-04-26T04:30:27Z | 2026-04-23T06:16:25Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | nik-greek | 82,338,475 | MDQ6VXNlcjgyMzM4NDc1 | User | false |
openclaw/openclaw | 4,297,634,490 | I_kwDOQb6kR88AAAABACiyug | 69,415 | https://github.com/openclaw/openclaw/issues/69415 | https://api.github.com/repos/openclaw/openclaw/issues/69415 | [Bug]: Web UI 粘贴图片结果为空消息 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
---
## 📝 Issue 摘要
**标题**：Web UI 粘贴图片结果为空消息
**核心问题**：
- 粘贴图片后预览正常
- 点击发送后消息为空
- 图片附件没有正确传输
**根因分析**：
代码中使用了 **非标准的图片格式**（Anthropic Claude 格式），但后端可能期望 **OpenAI 标准格式**。
| 当前代码（有问题） | 应该是（标准） |
|------------... | closed | completed | true | 0 | [
"bug",
"bug:behavior"
] | [] | 2026-04-20T17:52:43Z | 2026-04-24T04:26:33Z | 2026-04-21T05:04:36Z | NONE | null | 20260424T060021Z | 2026-04-24T06:00:21Z | majinggui | 187,953,390 | U_kgDOCzPw7g | User | false |
openclaw/openclaw | 4,297,657,360 | I_kwDOQb6kR88AAAABACkMEA | 69,418 | https://github.com/openclaw/openclaw/issues/69418 | https://api.github.com/repos/openclaw/openclaw/issues/69418 | [Bug]:Screenshots arrive as file paths — vision content not passed to model | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Screenshots arrive as file paths — vision content not passed to model despite imageModel config
### Steps to reproduce
Screenshots arrive as file paths — vision content not passed to model despite imageModel config
Steps ... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-04-20T17:56:54Z | 2026-04-28T04:49:14Z | 2026-04-25T19:09:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jwurban | 266,721,274 | U_kgDOD-XX-g | User | false |
openclaw/openclaw | 4,216,016,012 | I_kwDOQb6kR877S0yM | 62,345 | https://github.com/openclaw/openclaw/issues/62345 | https://api.github.com/repos/openclaw/openclaw/issues/62345 | [Bug] Discord REST final-reply and slash-command failures after undici v7 → v8 bump (2026.4.5) | ### Environment
- **OpenClaw version**: 2026.4.5 (broken) / 2026.4.2 (working)
- **Node.js**: v25.9.0
- **OS**: macOS (Apple Silicon)
- **Proxy**: HTTP proxy configured via `channels.discord.proxy` + `proxy-preload.cjs` (`--require` injected in LaunchAgent plist)
### Description
After upgrading from 2026.4.2 to 2026... | closed | completed | true | 4 | [] | [] | 2026-04-07T06:57:57Z | 2026-05-03T04:36:05Z | 2026-04-30T16:16:46Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | SirisZheng | 70,536,096 | MDQ6VXNlcjcwNTM2MDk2 | User | false |
openclaw/openclaw | 4,297,771,760 | I_kwDOQb6kR88AAAABACrK8A | 69,423 | https://github.com/openclaw/openclaw/issues/69423 | https://api.github.com/repos/openclaw/openclaw/issues/69423 | [Bug]: A single broken plugin's malformed tool schema crashes every Anthropic agent run | [Bug]: A single broken plugin's malformed tool schema crashes every agent run (no per-tool isolation in Anthropic adapter)
## Summary
When any loaded plugin registers a tool whose schema is stored under the wrong
key (e.g., `inputSchema` instead of `parameters`), the Anthropic tool-conversion
adapter throws `TypeErro... | closed | completed | true | 1 | [] | [] | 2026-04-20T18:18:09Z | 2026-05-01T04:40:33Z | 2026-04-29T03:21:46Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | jmnickels | 3,698,642 | MDQ6VXNlcjM2OTg2NDI= | User | false |
openclaw/openclaw | 4,297,794,057 | I_kwDOQb6kR88AAAABACsiCQ | 69,424 | https://github.com/openclaw/openclaw/issues/69424 | https://api.github.com/repos/openclaw/openclaw/issues/69424 | docs-search skill broken after 2026.4.15: local docs bundle reduced from 375 files to 13 templates | ## Summary
Starting with `2026.4.15`, the local docs bundle at `/opt/homebrew/lib/node_modules/openclaw/docs/` was restructured from a full documentation set (~375 markdown files, ~3425 chunks) down to **13 template files** under `docs/reference/templates/`.
This breaks the `openclaw-docs` CLI and any skill/workflow ... | closed | completed | true | 1 | [] | [] | 2026-04-20T18:22:10Z | 2026-04-30T04:38:29Z | 2026-04-26T05:44:25Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | spark-of-imagination[bot] | 271,815,129 | BOT_kgDOEDOR2Q | Bot | false |
openclaw/openclaw | 4,297,757,360 | I_kwDOQb6kR88AAAABACqSsA | 69,422 | https://github.com/openclaw/openclaw/issues/69422 | https://api.github.com/repos/openclaw/openclaw/issues/69422 | Google Chat outbound: thread/reply routing metadata leaks across streamed block replies | ## Summary
In the `googlechat` outbound channel, chunks of a single reply can land outside the thread of the inbound message (or split between thread and top-level of the space). The pipeline preserves the chat destination but not the thread target across chunks.
Affects: `v2026.4.12` (likely earlier). Reproducible i... | open | null | false | 4 | [] | [] | 2026-04-20T18:15:41Z | 2026-05-02T15:23:14Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | rubensandrade-sketch | 264,141,151 | U_kgDOD755Xw | User | false |
openclaw/openclaw | 4,297,853,195 | I_kwDOQb6kR88AAAABACwJCw | 69,427 | https://github.com/openclaw/openclaw/issues/69427 | https://api.github.com/repos/openclaw/openclaw/issues/69427 | Regression: Screenshot/image upload broken in webchat on 2026.4.15 | ## Bug Report
**Version:** 2026.4.15 (041266a)
**Previous working version:** 2026.4.5
**Platform:** macOS 26.2 (arm64), Node 22.22.1
**Channel:** webchat (Control UI)
## Description
Image/screenshot uploads via the webchat Control UI stopped working after upgrading from 2026.4.5 to 2026.4.15. The attachment button i... | closed | completed | true | 0 | [] | [] | 2026-04-20T18:33:15Z | 2026-04-24T04:26:33Z | 2026-04-21T05:04:36Z | NONE | null | 20260424T060021Z | 2026-04-24T06:00:21Z | drkyleahuggins | 262,856,883 | U_kgDOD6rgsw | User | false |
openclaw/openclaw | 4,080,640,622 | I_kwDOQb6kR87zOaJu | 47,975 | https://github.com/openclaw/openclaw/issues/47975 | https://api.github.com/repos/openclaw/openclaw/issues/47975 | Subagent sessions persist after completion, main session becomes unresponsive | ## Bug Report: Subagent sessions persist after completion, main session becomes unresponsive
**Version:** OpenClaw 2026.3.13 (stable channel)
**Environment:** WSL2, Ubuntu 24.04.4, Node v24.14.0
**Channel:** webchat (openclaw-control-ui dashboard)
### Summary
After spawning multiple subagent sessions (particularly w... | open | null | false | 7 | [] | [] | 2026-03-16T07:03:19Z | 2026-05-02T01:29:09Z | null | NONE | null | 20260502T030029Z | 2026-05-02T03:00:29Z | magnusbonnevier | 1,112,476 | MDQ6VXNlcjExMTI0NzY= | User | false |
openclaw/openclaw | 4,043,331,229 | I_kwDOQb6kR87xAFad | 40,604 | https://github.com/openclaw/openclaw/issues/40604 | https://api.github.com/repos/openclaw/openclaw/issues/40604 | Feature: sessions_kill — force-terminate sub-agent sessions | ## Problem
When a parent agent spawns a sub-agent via `sessions_spawn`, there is **no way to force-terminate it**. The only option is `sessions_send` with a "STOP" message, which:
1. Is just a chat message — the agent processes it only after its current tool call completes
2. The agent **decides** whether to obey — i... | closed | completed | true | 2 | [] | [] | 2026-03-09T04:50:22Z | 2026-04-27T04:57:30Z | 2026-04-24T20:51:20Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | antojk | 1,898,247 | MDQ6VXNlcjE4OTgyNDc= | User | false |
openclaw/openclaw | 4,114,520,411 | I_kwDOQb6kR871Pplb | 52,037 | https://github.com/openclaw/openclaw/issues/52037 | https://api.github.com/repos/openclaw/openclaw/issues/52037 | OpenAI Codex OAuth token refresh not persisting to disk | # OpenAI Codex OAuth Token Persistence Bug
**Reported:** 2026-03-22 03:14 UTC
**OpenClaw Version:** 2026.3.13 (61d171a)
**Environment:** Ubuntu VPS, systemd gateway service
---
## Summary
OpenAI Codex OAuth tokens are **not persisting to disk after refresh**. This causes:
1. Refresh tokens to be reused repeated... | closed | completed | true | 14 | [] | [] | 2026-03-22T03:15:14Z | 2026-04-30T04:49:53Z | 2026-04-25T05:47:53Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | rory78669-bot | 260,316,639 | U_kgDOD4Qd3w | User | false |
openclaw/openclaw | 4,184,224,598 | I_kwDOQb6kR875ZjNW | 58,758 | https://github.com/openclaw/openclaw/issues/58758 | https://api.github.com/repos/openclaw/openclaw/issues/58758 | Feature: compactionNotices config option for compaction-only notifications | ## Problem
When auto-compaction triggers during a conversation, the agent goes silent on messaging surfaces (Discord, Telegram, etc.) with no indication of what's happening. The user sends a message, context limit is hit, compaction runs, and there's just... silence. No feedback that anything is happening.
## Current... | closed | completed | true | 0 | [] | [] | 2026-04-01T05:27:28Z | 2026-04-23T04:22:25Z | 2026-04-20T18:55:19Z | NONE | null | 20260423T060020Z | 2026-04-23T06:00:20Z | brendongl | 31,434,186 | MDQ6VXNlcjMxNDM0MTg2 | User | false |
openclaw/openclaw | 4,297,992,355 | I_kwDOQb6kR88AAAABAC4oow | 69,434 | https://github.com/openclaw/openclaw/issues/69434 | https://api.github.com/repos/openclaw/openclaw/issues/69434 | Heartbeat: support model fallbacks (single point of failure when provider quota is exhausted) | ## Problem
The current schema for `agents.*.heartbeat` accepts a single `model` string (e.g. `"oc/kimi-k2.5"`) and no `fallbacks` array. Unlike `agents.defaults.model`, which supports `{ primary, fallbacks }`, the heartbeat model is a **single point of failure**.
When the chosen heartbeat provider returns `429` (quot... | closed | not_planned | true | 1 | [] | [] | 2026-04-20T18:58:44Z | 2026-04-29T04:36:27Z | 2026-04-26T12:05:48Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | zote | 35,069 | MDQ6VXNlcjM1MDY5 | User | false |
openclaw/openclaw | 4,297,993,058 | I_kwDOQb6kR88AAAABAC4rYg | 69,435 | https://github.com/openclaw/openclaw/issues/69435 | https://api.github.com/repos/openclaw/openclaw/issues/69435 | TTS auto-reply generates MP3 only — WhatsApp cannot play as voice note (needs OGG/Opus) | ## Summary
When TTS is enabled with `messages.tts.auto: "always"`, the gateway generates speech audio in **MP3 format** regardless of the target channel. This works fine on Telegram, but **WhatsApp** cannot play MP3 files as voice notes — it requires **OGG/Opus** format. As a result, WhatsApp users see "this audio is ... | closed | completed | true | 2 | [] | [] | 2026-04-20T18:58:51Z | 2026-04-28T04:56:12Z | 2026-04-25T17:35:31Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Fr1day-git | 257,967,343 | U_kgDOD2BE7w | User | false |
openclaw/openclaw | 4,298,016,994 | I_kwDOQb6kR88AAAABAC6I4g | 69,439 | https://github.com/openclaw/openclaw/issues/69439 | https://api.github.com/repos/openclaw/openclaw/issues/69439 | Feature request: Per-agent tool restrictions and multi-tenant isolation | ## Caso de uso
Sistema multi-agente para familia con 13 agentes configurados:
- Peter (admin): acceso total
- Lola, Carlos, Jorge (familia): uso general sin privilegios
- Javier: necesita instalar paquetes de programación
## Necesidades específicas
### 1. Restricciones por agente en tools.exec
Configuración deseada... | closed | completed | true | 1 | [] | [] | 2026-04-20T19:03:35Z | 2026-04-29T04:34:38Z | 2026-04-26T18:57:15Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | urrielu | 213,940,341 | U_kgDODMB4dQ | User | false |
openclaw/openclaw | 4,297,996,141 | I_kwDOQb6kR88AAAABAC43bQ | 69,436 | https://github.com/openclaw/openclaw/issues/69436 | https://api.github.com/repos/openclaw/openclaw/issues/69436 | Feature: enforce response invariants against same-turn tool evidence (block unsupported blocker/safety claims) | ## Summary
OpenClaw currently has strong prompt-time guidance, but no enforceable runtime seam for agent claims like:
- "sandbox blocked this"
- "permissions prevented this"
- ritual file-safety disclaimers such as "not malware / not code"
In live use, the agent can emit these statements with zero supporting tool evid... | closed | completed | true | 2 | [] | [] | 2026-04-20T18:59:31Z | 2026-04-28T04:49:13Z | 2026-04-25T19:09:26Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kenners22 | 267,410,298 | U_kgDOD_Bbeg | User | false |
openclaw/openclaw | 4,298,018,833 | I_kwDOQb6kR88AAAABAC6QEQ | 69,440 | https://github.com/openclaw/openclaw/issues/69440 | https://api.github.com/repos/openclaw/openclaw/issues/69440 | [Bug]: startup race can downgrade agent thinkingDefault=xhigh to high and persist the downgrade into the session | ### What happened?
An agent configured with `thinkingDefault: "xhigh"` can start a new session at `high` without any user override.
In my case this happened on an `invoice` agent using `openai-codex/gpt-5.4`. The downgrade happened before the first user message, and the downgraded level was then persisted into the se... | closed | completed | true | 1 | [] | [] | 2026-04-20T19:03:58Z | 2026-04-29T04:42:44Z | 2026-04-26T06:58:36Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | sunnydongbo | 249,433,142 | U_kgDODt4MNg | User | false |
openclaw/openclaw | 4,297,996,388 | I_kwDOQb6kR88AAAABAC44ZA | 69,437 | https://github.com/openclaw/openclaw/issues/69437 | https://api.github.com/repos/openclaw/openclaw/issues/69437 | Feature: built-in stale watched-file / live-logging guard tied to active session recency | ## Summary
OpenClaw can detect stale live logging externally, but there is no built-in runtime enforcement to keep an agent's required working-context / daily-note updates in sync with active work.
In practice this means the system can know an agent is actively working while the vault/log files are stale by 20m, 90m, ... | closed | not_planned | true | 2 | [] | [] | 2026-04-20T18:59:34Z | 2026-04-28T04:49:13Z | 2026-04-25T19:09:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kenners22 | 267,410,298 | U_kgDOD_Bbeg | User | false |
openclaw/openclaw | 4,298,063,491 | I_kwDOQb6kR88AAAABAC8-gw | 69,441 | https://github.com/openclaw/openclaw/issues/69441 | https://api.github.com/repos/openclaw/openclaw/issues/69441 | Discord guild channel ignores `requireMention: false` and still only replies on @mention | **Version**
OpenClaw `2026.4.15`
**Summary**
In a Discord guild channel, OpenClaw only replies when the bot is explicitly mentioned, even though the configured channel has `requireMention: false`.
**Environment**
- OpenClaw version: `2026.4.15`
- Channel: Discord
- Bot: `@Ziggy`
- Channel name: `Gryph Labs / #gen... | closed | completed | true | 0 | [] | [] | 2026-04-20T19:12:16Z | 2026-04-25T04:17:05Z | 2026-04-22T19:29:53Z | NONE | null | 20260425T060021Z | 2026-04-25T06:00:21Z | WMNau | 25,082,027 | MDQ6VXNlcjI1MDgyMDI3 | User | false |
openclaw/openclaw | 4,298,081,014 | I_kwDOQb6kR88AAAABAC-C9g | 69,442 | https://github.com/openclaw/openclaw/issues/69442 | https://api.github.com/repos/openclaw/openclaw/issues/69442 | [Feature]: Documentation map file for LLM navigation | ### Summary
Why: OpenClaw has deep docs (https://docs.openclaw.ai/) across automation/, channels/, cli/, concepts/, gateway/, install/, nodes/, plugins/, providers/, tools/, web/, and more. An LLM currently has to discover this tree through trial and error. A single map fixes that. Example: [code.claude.com/docs/en/cl... | open | reopened | false | 1 | [
"enhancement"
] | [] | 2026-04-20T19:15:12Z | 2026-05-02T13:08:05Z | null | CONTRIBUTOR | null | 20260502T150030Z | 2026-05-02T15:00:30Z | GHesericsu | 60,202,455 | MDQ6VXNlcjYwMjAyNDU1 | User | false |
openclaw/openclaw | 4,298,203,392 | I_kwDOQb6kR88AAAABADFhAA | 69,443 | https://github.com/openclaw/openclaw/issues/69443 | https://api.github.com/repos/openclaw/openclaw/issues/69443 | [Bug] Subagent RPC callback to WeChat session key routed to main session instead |
## Bug Description
When a subagent (熊猫/Panda) uses `gateway call agent` RPC with `to` set to a WeChat session key (e.g., `agent:main:openclaw-weixin:direct:o9cq802hhmfc_1uryu_d4mwmrkdu@im.wechat`), the message is incorrectly routed to the main session (`agent:main:main`) instead of the specified WeChat session.
## S... | open | null | false | 1 | [] | [] | 2026-04-20T19:37:27Z | 2026-05-02T16:55:38Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | chen11221 | 74,538,310 | MDQ6VXNlcjc0NTM4MzEw | User | false |
openclaw/openclaw | 4,182,895,007 | I_kwDOQb6kR875Uemf | 58,602 | https://github.com/openclaw/openclaw/issues/58602 | https://api.github.com/repos/openclaw/openclaw/issues/58602 | Discord /new slash command times out - needs deferReply() | ## Problem
The `/new` slash command in Discord consistently fails with "application did not respond in time." This is the Discord 3-second interaction timeout -- the gateway takes longer than 3 seconds to process the session reset and Discord gives up before getting a response.
## Expected Behavior
`/new` should res... | closed | completed | true | 8 | [] | [] | 2026-03-31T23:06:39Z | 2026-04-28T05:15:44Z | 2026-04-25T12:42:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Skippy-the-Magnificent-one | 269,547,516 | U_kgDOEBD3_A | User | false |
openclaw/openclaw | 4,298,268,740 | I_kwDOQb6kR88AAAABADJgRA | 69,445 | https://github.com/openclaw/openclaw/issues/69445 | https://api.github.com/repos/openclaw/openclaw/issues/69445 | Feature: per-agent status card field suppression (hide usage/billing from specific agents) | ## Problem
The `/status` command (session_status tool) exposes provider usage percentages (e.g. "Premium 7% left") to all agents equally. In multi-agent setups where different agents serve different users, this creates an information leak — a secondary agent's user can see the account owner's subscription usage, which... | closed | not_planned | true | 1 | [] | [] | 2026-04-20T19:49:15Z | 2026-04-29T04:35:24Z | 2026-04-26T16:03:09Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | OctavianTocan | 55,554,752 | MDQ6VXNlcjU1NTU0NzUy | User | false |
openclaw/openclaw | 4,298,303,355 | I_kwDOQb6kR88AAAABADLnew | 69,447 | https://github.com/openclaw/openclaw/issues/69447 | https://api.github.com/repos/openclaw/openclaw/issues/69447 | Web UI attachments are image-only end-to-end (UI picker + Gateway) | ## Summary
The control-ui chat input only accepts image files, and the Gateway-side attachment normalizer silently drops anything non-image. Users cannot share PDFs, text files, or any non-image document through chat — even though the underlying Anthropic API supports document blocks for PDFs, and OpenClaw already imp... | closed | not_planned | true | 1 | [] | [] | 2026-04-20T19:55:54Z | 2026-04-29T04:37:05Z | 2026-04-26T10:50:20Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jeremyschissler-crypto | 277,866,602 | U_kgDOEI_oag | User | false |
openclaw/openclaw | 4,298,341,104 | I_kwDOQb6kR88AAAABADN68A | 69,448 | https://github.com/openclaw/openclaw/issues/69448 | https://api.github.com/repos/openclaw/openclaw/issues/69448 | Message duplication: agent runner retries full turns on tool errors, resending text to user | ## Bug: Duplicate messages sent to user on tool errors
### Description
When a tool call fails with an error (e.g., "guildId required" on Discord `message` send, or "node not connected"), the agent runner retries the entire turn. Since the model generates both text output AND a tool call in the same turn, the text gets... | closed | completed | true | 1 | [] | [] | 2026-04-20T20:02:39Z | 2026-04-28T04:49:12Z | 2026-04-25T19:09:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | frank-bot07 | 259,001,718 | U_kgDOD3ANdg | User | false |
openclaw/openclaw | 4,298,419,457 | I_kwDOQb6kR88AAAABADStAQ | 69,450 | https://github.com/openclaw/openclaw/issues/69450 | https://api.github.com/repos/openclaw/openclaw/issues/69450 | [PERF] Gateway scans ALL agent sessions for EVERY message — defeats agent isolation, prevents scaling | ### Summary
When a user messages a specific agent (e.g., main), the Gateway currently scans session files across ALL agents, not just the target agent. This creates a critical scalability bottleneck.
### Problem to solve
User messages (main agent)
↓
Gateway scans:
✅ /agents/main/sessions/*.jsonl (needed)
❌ /ag... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-20T20:16:33Z | 2026-04-29T04:35:49Z | 2026-04-26T14:56:01Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | benjaminlau | 1,013,826 | MDQ6VXNlcjEwMTM4MjY= | User | false |
openclaw/openclaw | 4,298,446,459 | I_kwDOQb6kR88AAAABADUWew | 69,451 | https://github.com/openclaw/openclaw/issues/69451 | https://api.github.com/repos/openclaw/openclaw/issues/69451 | Memory: Session files loaded entirely into memory via readFileSync | ## Problem
OpenClaw gateway loads ALL session files into memory at startup:
1. `sessions.json` (42MB in our case) is loaded via `readFileSync` + `JSON.parse`
2. All JSONL session files (831 files, 389MB) are loaded via `readFileSync` per session access
3. No lazy loading or streaming mechanism exists
This causes the... | open | null | false | 1 | [] | [] | 2026-04-20T20:21:00Z | 2026-05-02T16:55:39Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | ksxh0524 | 40,232,908 | MDQ6VXNlcjQwMjMyOTA4 | User | false |
openclaw/openclaw | 4,298,505,219 | I_kwDOQb6kR88AAAABADX8Aw | 69,453 | https://github.com/openclaw/openclaw/issues/69453 | https://api.github.com/repos/openclaw/openclaw/issues/69453 | GPT-5.4 context window resolves inconsistently after reboot/fresh session (200k / 272k / 1050k disagreement) | ## Summary
On macOS with OpenClaw `2026.4.15`, `openai-codex/gpt-5.4` can resolve to different context limits depending on which layer you look at, especially after reboot or fresh session creation.
Observed split:
- fresh Telegram DM session seeded at `200000`
- Mac app model catalog showed `272000` for `openai-cod... | closed | completed | true | 3 | [] | [] | 2026-04-20T20:31:12Z | 2026-04-29T04:34:37Z | 2026-04-26T18:57:19Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | marklights54-byte | 264,699,524 | U_kgDOD8b-hA | User | false |
openclaw/openclaw | 4,222,914,272 | I_kwDOQb6kR877tI7g | 63,008 | https://github.com/openclaw/openclaw/issues/63008 | https://api.github.com/repos/openclaw/openclaw/issues/63008 | [Bug]: video_generate: Google Veo 3.1 returns 404 'No such model' due to wrong API endpoint | ## Problem
The `video_generate` tool fails when using `google/veo-3.1-fast-generate-preview` model with error:
```
404 NOT_FOUND
No such model \"veo-3.1-fast-generate-preview\" at this location
```
## Environment
- OpenClaw version: 2026.4.8 (also affected 2026.4.5 when feature was introduced)
- Google plugin with ... | closed | completed | true | 3 | [] | [] | 2026-04-08T07:42:53Z | 2026-04-28T05:25:43Z | 2026-04-25T10:11:58Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kamusis | 1,849,506 | MDQ6VXNlcjE4NDk1MDY= | User | false |
openclaw/openclaw | 4,298,637,607 | I_kwDOQb6kR88AAAABADgBJw | 69,456 | https://github.com/openclaw/openclaw/issues/69456 | https://api.github.com/repos/openclaw/openclaw/issues/69456 | [Bug]: Windows install regressions slipped past current CI coverage | ## Summary
On Windows 11, we reproduced two install/build regressions locally that required fixes outside current `main`, even though the repo already has a Windows CI lane.
The two observed regressions were:
1. `scripts/pnpm-runner.mjs` needed a Windows-native direct-exec path when `npm_execpath` pointed to `pnpm.e... | closed | completed | true | 1 | [] | [] | 2026-04-20T20:53:04Z | 2026-04-29T04:42:04Z | 2026-04-26T07:15:24Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | igormf | 8,070,635 | MDQ6VXNlcjgwNzA2MzU= | User | false |
openclaw/openclaw | 4,298,671,485 | I_kwDOQb6kR88AAAABADiFfQ | 69,460 | https://github.com/openclaw/openclaw/issues/69460 | https://api.github.com/repos/openclaw/openclaw/issues/69460 | Auto-compaction fails on Claude Opus 4.7 — thinking.type.enabled unsupported (use adaptive + output_config.effort) | **Version:** OpenClaw 2026.4.15 (041266a)
**Model:** `anthropic/claude-opus-4-7`
**OS:** Linux 6.6.87.2-microsoft-standard-WSL2 (WSL2, x64)
## What happened
When a session exceeded the 200K context window, auto-compaction triggered and failed with:
```
Compaction failed: Summarization failed: 400 {"type":"error","er... | closed | completed | true | 2 | [] | [] | 2026-04-20T20:59:27Z | 2026-04-28T04:49:11Z | 2026-04-25T19:09:46Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | mckoss | 82,750 | MDQ6VXNlcjgyNzUw | User | false |
openclaw/openclaw | 4,298,697,705 | I_kwDOQb6kR88AAAABADjr6Q | 69,462 | https://github.com/openclaw/openclaw/issues/69462 | https://api.github.com/repos/openclaw/openclaw/issues/69462 | [Bug]: `openclaw status` / `openclaw health` take 40s / 17s — worker thread pegged in `jiti`'s `normalizeAliases` | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Summary
On a fresh, correctly-installed `openclaw@2026.4.15` (commit `041266a`), one-shot informational commands take tens of seconds:
| command | real | user | sys |
|---|---|---|---|
| `openclaw --version` | 0.18s | 0... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-20T21:04:14Z | 2026-04-24T04:25:19Z | 2026-04-22T02:37:25Z | NONE | null | 20260424T060021Z | 2026-04-24T06:00:21Z | chsbusch-dot | 265,684,486 | U_kgDOD9YGBg | User | false |
openclaw/openclaw | 4,298,722,957 | I_kwDOQb6kR88AAAABADlOjQ | 69,465 | https://github.com/openclaw/openclaw/issues/69465 | https://api.github.com/repos/openclaw/openclaw/issues/69465 | Bug: MCP stdio server processes orphaned on session end — 2.4GB RAM leak from context7-mcp | ## Summary
MCP stdio servers using `npx` are spawned per-session but never cleaned up when sessions end. Over time, orphaned processes accumulate and consume significant RAM.
### Observed Behavior
On a 30GB RAM server, **41 orphaned context7-mcp processes** were found consuming ~2.4GB of RAM total (~60-100MB each). ... | closed | completed | true | 4 | [] | [] | 2026-04-20T21:08:40Z | 2026-04-25T04:16:40Z | 2026-04-22T21:00:16Z | NONE | null | 20260425T060021Z | 2026-04-25T06:00:21Z | Mertolgahan | 36,914,136 | MDQ6VXNlcjM2OTE0MTM2 | User | false |
openclaw/openclaw | 4,298,727,525 | I_kwDOQb6kR88AAAABADlgZQ | 69,466 | https://github.com/openclaw/openclaw/issues/69466 | https://api.github.com/repos/openclaw/openclaw/issues/69466 | Bug: `openclaw security audit` hangs indefinitely (no output) | ## Description
The `openclaw security audit` command hangs indefinitely and never returns output. Tested with multiple flags:
- `openclaw security audit` — hangs
- `openclaw security audit --deep` — hangs
- `openclaw security audit --json` — hangs
All variations were killed by timeout (15-20s). No output is produc... | closed | completed | true | 1 | [] | [] | 2026-04-20T21:09:28Z | 2026-04-29T04:36:26Z | 2026-04-26T12:05:52Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | luisgennari | 5,799,798 | MDQ6VXNlcjU3OTk3OTg= | User | false |
openclaw/openclaw | 4,298,755,533 | I_kwDOQb6kR88AAAABADnNzQ | 69,467 | https://github.com/openclaw/openclaw/issues/69467 | https://api.github.com/repos/openclaw/openclaw/issues/69467 | [Bug]: one-shot cron deleteAfterRun can delete and rebind a live non-default DM session when sessionTarget resolves to current/session:* | ## Anonymization map
- `AGENT_A`: affected non-default agent
- `USER_A`: the human chatting with `AGENT_A`
- `LIVE_DM_KEY`: `agent:<agent-a>:signal:default:direct:<redacted-recipient>`
- `SESSION_A`: populated live DM session before reminder execution
- `SESSION_B`: fresh session created after the first delete
- `SESS... | open | null | false | 1 | [] | [] | 2026-04-20T21:15:41Z | 2026-05-02T17:31:35Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | bricolage | 2,019 | MDQ6VXNlcjIwMTk= | User | false |
openclaw/openclaw | 4,298,769,703 | I_kwDOQb6kR88AAAABADoFJw | 69,469 | https://github.com/openclaw/openclaw/issues/69469 | https://api.github.com/repos/openclaw/openclaw/issues/69469 | Hot reload of agents.defaults.model.primary overrides channels.modelByChannel for existing sessions | ## Summary
When `agents.defaults.model.primary` is changed in `openclaw.json`, the gateway hot-reload path applies the new default directly to already-running sessions **without rechecking `channels.modelByChannel`**. This means a per-channel model override (e.g. pinning a Telegram DM to Opus while the global default ... | closed | not_planned | true | 1 | [] | [] | 2026-04-20T21:18:32Z | 2026-04-28T04:49:10Z | 2026-04-25T19:09:52Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | koskol | 36,731,165 | MDQ6VXNlcjM2NzMxMTY1 | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.