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 | 4,200,061,352 | I_kwDOQb6kR876V9mo | 60,244 | https://github.com/openclaw/openclaw/issues/60244 | https://api.github.com/repos/openclaw/openclaw/issues/60244 | QMD memory backend dirty detection not working - always shows 'Dirty: no' | ## Problem
When using QMD as memory backend, modifying memory files (MEMORY.md, memory/*.md) does not trigger dirty detection. The `openclaw memory status` command always shows `Dirty: no`, even after file modifications.
## Expected Behavior
- After modifying memory files, `Dirty` should show `yes` (or similar indic... | closed | completed | true | 1 | [] | [] | 2026-04-03T11:10:34Z | 2026-04-30T04:33:59Z | 2026-04-27T13:20:37Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | xinzf | 7,508,359 | MDQ6VXNlcjc1MDgzNTk= | User | false |
openclaw/openclaw | 4,336,077,870 | I_kwDOQb6kR88AAAABAnNMLg | 72,824 | https://github.com/openclaw/openclaw/issues/72824 | https://api.github.com/repos/openclaw/openclaw/issues/72824 | github-copilot: Claude Opus 1m variants treated as 128K context (and xhigh blocked for 4.7-1m-internal) | # Issue: GitHub Copilot Claude Opus "1m" variants get 128K context window instead of 1M
## Summary
OpenClaw's GitHub Copilot extension treats Copilot's long-context Claude Opus variants (`claude-opus-4.6-1m`, `claude-opus-4.7-1m-internal`) as if they had a 128K context window, even though Copilot's `/models` endpoint... | open | null | false | 3 | [] | [] | 2026-04-27T13:24:37Z | 2026-05-03T12:31:44Z | null | CONTRIBUTOR | null | 20260503T150031Z | 2026-05-03T15:00:31Z | jakepresent | 31,285,242 | MDQ6VXNlcjMxMjg1MjQy | User | false |
openclaw/openclaw | 4,336,067,757 | I_kwDOQb6kR88AAAABAnMkrQ | 72,823 | https://github.com/openclaw/openclaw/issues/72823 | https://api.github.com/repos/openclaw/openclaw/issues/72823 | sessions_spawn tool not available - multi-agent dispatch blocked | ## Issue
sessions_spawn is documented in SOUL.md but the tool is not actually available in the agent tool list. This blocks native multi-agent dispatch.
## Environment
- OpenClaw: 2026.4.22+
- Channel: Feishu
- Model: minimax/MiniMax-M2.7
## Expected Behavior
sessions_spawn should be available as a native tool for di... | closed | completed | true | 1 | [] | [] | 2026-04-27T13:23:04Z | 2026-04-30T04:33:57Z | 2026-04-27T13:31:32Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | hahahahahahhhah | 127,914,068 | U_kgDOB5_QVA | User | false |
openclaw/openclaw | 4,336,120,147 | I_kwDOQb6kR88AAAABAnPxUw | 72,828 | https://github.com/openclaw/openclaw/issues/72828 | https://api.github.com/repos/openclaw/openclaw/issues/72828 | Feishu WebSocket reconnect exhausted causes permanent connection freeze | ## Bug Report
**Version:** OpenClaw 2026.4.24 (cbcfdf6)
**OS:** macOS Darwin 25.3.0 (arm64)
## Symptom
Feishu WebSocket disconnects every 30-40 minutes and never auto-recovers. Only a Gateway restart restores the connection.
## Root Cause (traced to source)
1. Feishu server sends `reconnectCount=1` in the pong conf... | closed | not_planned | true | 1 | [] | [] | 2026-04-27T13:31:43Z | 2026-05-03T04:36:46Z | 2026-04-30T14:53:43Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | babelee0805 | 262,993,834 | U_kgDOD6z3qg | User | false |
openclaw/openclaw | 4,336,135,891 | I_kwDOQb6kR88AAAABAnQu0w | 72,830 | https://github.com/openclaw/openclaw/issues/72830 | https://api.github.com/repos/openclaw/openclaw/issues/72830 | TTS mode:final causes duplicate text message delivery on chat channels | ### Description
When `messages.tts.mode` is set to `"final"`, every assistant reply is delivered **twice** as two separate messages on chat channels (confirmed on webchat and Feishu). The first message contains text only, and the second (arriving 6-20 seconds later) contains the same text plus the TTS audio attachment... | closed | completed | true | 2 | [] | [] | 2026-04-27T13:34:22Z | 2026-05-01T04:44:39Z | 2026-04-28T07:35:51Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | lhtpluto | 132,045,861 | U_kgDOB97cJQ | User | false |
openclaw/openclaw | 4,336,181,339 | I_kwDOQb6kR88AAAABAnTgWw | 72,833 | https://github.com/openclaw/openclaw/issues/72833 | https://api.github.com/repos/openclaw/openclaw/issues/72833 | feat: [Image Gen] Support full FAL.ai model catalog — expose all 1000+ models as selectable providers | ## Problem
The current `fal` image generation provider in OpenClaw only exposes 2 hardcoded models (`flux/dev` and `flux/dev/image-to-image`). FAL.ai has 1,094+ models including:
- Text-to-image: FLUX.1 [schnell/dev/pro], Stable Diffusion, Imagen
- Image-to-video: Kling v3, Seedance 2, Veo 3, Wan, Sora
- Image-to-ima... | closed | not_planned | true | 1 | [] | [] | 2026-04-27T13:41:26Z | 2026-05-05T04:26:52Z | 2026-05-02T15:49:25Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | AS76 | 22,101,267 | MDQ6VXNlcjIyMTAxMjY3 | User | false |
openclaw/openclaw | 4,336,219,364 | I_kwDOQb6kR88AAAABAnV05A | 72,835 | https://github.com/openclaw/openclaw/issues/72835 | https://api.github.com/repos/openclaw/openclaw/issues/72835 | [Bug]: PR #71694 runtime cron service fallback ineffective on Linux v2026.4.25 — `gatewayContext.getCron()` returns null | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
PR #71694 ("fix(memory-core): add runtime cron service fallback for dreaming reconciliation", merged 2026-04-25) intends to recover the dreaming cron reconciliation when the startup capture is null by falling b... | closed | completed | true | 1 | [] | [] | 2026-04-27T13:47:03Z | 2026-04-30T04:33:33Z | 2026-04-27T19:26:38Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | RayWoo | 2,201,861 | MDQ6VXNlcjIyMDE4NjE= | User | false |
openclaw/openclaw | 4,336,117,666 | I_kwDOQb6kR88AAAABAnPnog | 72,827 | https://github.com/openclaw/openclaw/issues/72827 | https://api.github.com/repos/openclaw/openclaw/issues/72827 | [Bug]: subagents.allowAgents not enforced on sessions_spawn — agents can spawn any agentId regardless of allowlist | ## Summary
`subagents.allowAgents` in `openclaw.json` is not validated by the gateway on `sessions_spawn` calls. An agent can spawn any agentId — including ones not in its allowlist — and the gateway tracks the spawn without rejecting it.
## Reproduction
Config:
```json
// task-manager agent entry in openclaw.json
"... | closed | completed | true | 1 | [] | [] | 2026-04-27T13:31:17Z | 2026-04-30T04:33:53Z | 2026-04-27T13:53:31Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | oiGaDio | 265,029,916 | U_kgDOD8wJHA | User | false |
openclaw/openclaw | 4,336,109,226 | I_kwDOQb6kR88AAAABAnPGqg | 72,826 | https://github.com/openclaw/openclaw/issues/72826 | https://api.github.com/repos/openclaw/openclaw/issues/72826 | Gateway sessionStore maintenance synchronously blocks event loop for 30-60s, causes GC starvation and OOM | ## Symptoms
OpenClaw gateway exhibits monotonic heap growth and crashes with V8 fatal error `Reached heap limit Allocation failed - JavaScript heap out of memory`, on the order of every ~90 minutes on a single-host install with ~50 cron jobs running.
- Heap RSS grows linearly at ~11 MB/min until OOM
- V8 stacks at th... | closed | completed | true | 1 | [] | [] | 2026-04-27T13:29:53Z | 2026-04-30T04:33:53Z | 2026-04-27T13:54:32Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | jared-rebel | 147,876,109 | U_kgDOCNBpDQ | User | false |
openclaw/openclaw | 4,336,219,856 | I_kwDOQb6kR88AAAABAnV20A | 72,836 | https://github.com/openclaw/openclaw/issues/72836 | https://api.github.com/repos/openclaw/openclaw/issues/72836 | Bug: v2026.4.25 still triggers WhatsApp runtime deps install via persisted auth state, causing libsignal-node SSH failure on startup | ## Summary
After upgrading to OpenClaw `2026.4.25`, Gateway startup still triggers WhatsApp runtime dependency installation even though WhatsApp is not explicitly enabled in config.
This leads to npm trying to resolve Baileys -> `libsignal-node` from GitHub over SSH, which fails on hosts without a GitHub SSH key:
- ... | closed | completed | true | 2 | [] | [] | 2026-04-27T13:47:07Z | 2026-04-30T04:33:30Z | 2026-04-27T13:55:11Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | xiao398008 | 211,473,177 | U_kgDODJrTGQ | User | false |
openclaw/openclaw | 4,336,307,968 | I_kwDOQb6kR88AAAABAnbPAA | 72,841 | https://github.com/openclaw/openclaw/issues/72841 | https://api.github.com/repos/openclaw/openclaw/issues/72841 | [Bug]: memory-core: dreaming cron never registers — startup timing + reconciliation skip bug | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
memory-core dreaming cron job fails to register at gateway startup and never retries, causing the Memory Dreaming Promotion task to never execute.
### Steps to reproduce
1. Ensure memory-core plugin is enabled and dreaming... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-27T14:00:35Z | 2026-05-01T04:43:44Z | 2026-04-28T11:15:24Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | RayySummers | 106,456,682 | U_kgDOBlhmag | User | false |
openclaw/openclaw | 4,336,273,976 | I_kwDOQb6kR88AAAABAnZKOA | 72,839 | https://github.com/openclaw/openclaw/issues/72839 | https://api.github.com/repos/openclaw/openclaw/issues/72839 | memory-lancedb: OpenAI SDK v6.34+ defaults to encoding_format=base64, breaking local embedding providers (LM Studio, Ollama) | ## Problem
OpenAI SDK v6.34.0 changed the default `encoding_format` for embeddings from `float` to `base64` (see [openai-node PR](https://github.com/openclaw/openai-node/pull/XXX)). This breaks local embedding providers like LM Studio and Ollama that may not handle `base64` encoding correctly.
### Observed behavior
... | closed | completed | true | 1 | [] | [] | 2026-04-27T13:55:20Z | 2026-04-30T04:33:51Z | 2026-04-27T14:01:14Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | le910n7 | 278,591,303 | U_kgDOEJr3Rw | User | false |
openclaw/openclaw | 4,336,389,749 | I_kwDOQb6kR88AAAABAngOdQ | 72,845 | https://github.com/openclaw/openclaw/issues/72845 | https://api.github.com/repos/openclaw/openclaw/issues/72845 | [Bug]: axios，这个插件怎么又改回来了，哪位大哥在这瞎搞啊 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
<img width="964" height="393" alt="Image" src="https://github.com/user-attachments/assets/a989e8e2-fbf3-4815-9d7c-be74fdd42394" />
### Steps to reproduce
<img width="964" height="393" alt="Image" src="https://github.com/us... | closed | completed | true | 0 | [
"bug",
"regression"
] | [] | 2026-04-27T14:13:33Z | 2026-04-30T04:33:48Z | 2026-04-27T14:13:53Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | zyy88-code | 52,097,530 | MDQ6VXNlcjUyMDk3NTMw | User | false |
openclaw/openclaw | 4,336,421,772 | I_kwDOQb6kR88AAAABAniLjA | 72,849 | https://github.com/openclaw/openclaw/issues/72849 | https://api.github.com/repos/openclaw/openclaw/issues/72849 | Cron jobs without `id` field cause runtime state to collapse into first job (find-collision) | # Cron jobs without `id` field cause runtime state to collapse into first job (find-collision)
## Summary
When `~/.openclaw/cron/jobs.json` contains jobs missing the `id` field, the scheduler's `Array#find()` matches the first entry every time, causing every job's runtime state (errors, lastRunAtMs, nextRunAtMs, cons... | open | null | false | 2 | [] | [] | 2026-04-27T14:18:28Z | 2026-05-03T20:25:08Z | null | NONE | null | 20260503T210034Z | 2026-05-03T21:00:34Z | the-Nile-jones | 191,544,035 | U_kgDOC2q64w | User | false |
openclaw/openclaw | 4,336,414,201 | I_kwDOQb6kR88AAAABAnht-Q | 72,848 | https://github.com/openclaw/openclaw/issues/72848 | https://api.github.com/repos/openclaw/openclaw/issues/72848 | [Bug]: Openclaw no longer works after "Openclaw Update" from 4.23 to either 4.24 or 4.25. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
On 2012 Intel based Mac running Sequoia and have been unable to update since 4.23. Tried 4.24 and gateway timeouts; reverted to 4.23; tried 4.25 today and same result; reverted to 4.23 again.
### Steps to reproduce
1. On... | closed | completed | true | 8 | [
"bug",
"regression"
] | [] | 2026-04-27T14:17:17Z | 2026-04-30T04:33:23Z | 2026-04-27T20:06:25Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | tess020126-cmyk | 259,169,068 | U_kgDOD3KbLA | User | false |
openclaw/openclaw | 4,336,309,267 | I_kwDOQb6kR88AAAABAnbUEw | 72,842 | https://github.com/openclaw/openclaw/issues/72842 | https://api.github.com/repos/openclaw/openclaw/issues/72842 | [Bug]: Why have the last two upgrades both reported the error: Completion cache update failed: Error: spawnSync /usr/bin/node ETIMEDOUT? | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Updating OpenClaw...
│
◇ ✓ Updating via package manager (23.21s)
│
◇ ✓ Running doctor checks (73.38s)
Update Result: OK
Root: /usr/lib/node_modules/openclaw
Before: 2026.4.25
After: 2026.4.25
Total time: 97.01s
U... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-27T14:00:46Z | 2026-04-30T04:33:19Z | 2026-04-27T20:19:20Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | mapleshadow | 525,478 | MDQ6VXNlcjUyNTQ3OA== | User | false |
openclaw/openclaw | 4,336,516,073 | I_kwDOQb6kR88AAAABAnn76Q | 72,851 | https://github.com/openclaw/openclaw/issues/72851 | https://api.github.com/repos/openclaw/openclaw/issues/72851 | [Bug]: [Bug]: Ollama provider hangs or fails on infer model run --local across 2026.4.20 / 2026.4.24 / 2026.4.25 with three distinct failure modes | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The Ollama provider does not deliver responses to the CLI on `openclaw infer model run --local` and `openclaw agent`; reproduced across versions 2026.4.20, 2026.4.24, and 2026.4.25 against the same healthy loca... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-27T14:32:26Z | 2026-04-30T04:32:59Z | 2026-04-27T21:39:13Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | TotalRes2020 | 66,965,950 | MDQ6VXNlcjY2OTY1OTUw | User | false |
openclaw/openclaw | 4,111,642,035 | I_kwDOQb6kR871Eq2z | 51,469 | https://github.com/openclaw/openclaw/issues/51469 | https://api.github.com/repos/openclaw/openclaw/issues/51469 | CLI gateway handshake timeout too short for cold-start module compilation | ## Summary
CLI commands that connect to the gateway (e.g. `openclaw cron list`) fail with `gateway closed (1000 normal closure): no close reason` on systems where Node.js ESM module compilation takes longer than the gateway's 3-second WebSocket handshake timeout.
## Environment
- OpenClaw: 2026.3.13 (61d171a)
- Node... | closed | completed | true | 6 | [] | [] | 2026-03-21T06:43:02Z | 2026-05-01T04:42:01Z | 2026-04-28T19:58:31Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | GothicFox | 8,213,373 | MDQ6VXNlcjgyMTMzNzM= | User | false |
openclaw/openclaw | 4,336,220,062 | I_kwDOQb6kR88AAAABAnV3ng | 72,837 | https://github.com/openclaw/openclaw/issues/72837 | https://api.github.com/repos/openclaw/openclaw/issues/72837 | [Bug]: Dreaming narrative generation times out for all 3 phases on v2026.4.25 warm gateway — #69487 fix appears regressed/incomplete | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On v2026.4.25 stable, `memory-core` narrative generation reliably times out for all three dreaming phases (light, rem, deep) regardless of gateway warm-up state. This contradicts the resolution of #69487 ("Firs... | closed | completed | true | 1 | [] | [] | 2026-04-27T13:47:09Z | 2026-04-30T04:33:17Z | 2026-04-27T20:19:40Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | RayWoo | 2,201,861 | MDQ6VXNlcjIyMDE4NjE= | User | false |
openclaw/openclaw | 4,336,413,821 | I_kwDOQb6kR88AAAABAnhsfQ | 72,847 | https://github.com/openclaw/openclaw/issues/72847 | https://api.github.com/repos/openclaw/openclaw/issues/72847 | [Bug] edit tool returns success but file is not actually modified - causes infinite loop | ## Bug Description
The \edit\ tool reports success when it should, but the file content is not actually modified. This causes the agent to enter an infinite loop: it keeps calling \edit\ with the same \oldText\ (which matches an outdated footer), gets a success response, but the file remains unchanged. On the next rea... | closed | completed | true | 2 | [] | [] | 2026-04-27T14:17:13Z | 2026-05-03T04:36:46Z | 2026-04-30T14:53:50Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | qq550842026-sudo | 269,676,242 | U_kgDOEBLu0g | User | false |
openclaw/openclaw | 4,336,561,619 | I_kwDOQb6kR88AAAABAnqt0w | 72,856 | https://github.com/openclaw/openclaw/issues/72856 | https://api.github.com/repos/openclaw/openclaw/issues/72856 | Cron isolated sessions have no access to main session history for non-default agents | ## Problem
When using `--session isolated` for cron jobs on non-default agents, the isolated session starts fresh with no access to the main conversation history. This makes it impossible for the cron job to summarize or reference past conversations.
## Use Case
I want to create a daily memory cron job that runs at ... | closed | completed | true | 2 | [] | [] | 2026-04-27T14:38:31Z | 2026-05-03T04:36:45Z | 2026-04-30T14:53:55Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | oottlee | 270,504,191 | U_kgDOEB-Q_w | User | false |
openclaw/openclaw | 4,336,590,722 | I_kwDOQb6kR88AAAABAnsfgg | 72,858 | https://github.com/openclaw/openclaw/issues/72858 | https://api.github.com/repos/openclaw/openclaw/issues/72858 | Bug: WebChat/Control UI exec remains denied with allowlist miss even after elevated and provider tool policy are configured | ## Summary
In WebChat / Control UI, `exec` remains unavailable with:
```text
exec denied: allowlist miss
```
even after configuring both:
1. elevated sender allow rules for the WebChat sender identity
2. provider-level tool policy allowing `exec` and `process` for `webchat`
This suggests WebChat / Control UI exec ... | open | null | false | 2 | [] | [
"BunsDev"
] | 2026-04-27T14:42:13Z | 2026-05-03T01:38:06Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | xiao398008 | 211,473,177 | U_kgDODJrTGQ | User | false |
openclaw/openclaw | 4,336,546,664 | I_kwDOQb6kR88AAAABAnpzaA | 72,854 | https://github.com/openclaw/openclaw/issues/72854 | https://api.github.com/repos/openclaw/openclaw/issues/72854 | [Bug]: readPackageVersion fallback to 'unknown' causes feishu channel crash loop | ## Bug Description
**Version:** OpenClaw 2026.4.24
**OS:** macOS Darwin
## Symptom
Feishu channel enters a crash loop on every gateway startup. Error:
```
[feishu] [default] channel exited: Cannot find module '/Users/oran/.openclaw/plugin-runtime-deps/openclaw-unknown-ddf66de49cc0/dist/extensions/feishu/monitor-C_w... | closed | completed | true | 2 | [] | [] | 2026-04-27T14:36:36Z | 2026-04-30T04:33:21Z | 2026-04-27T20:09:05Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | linglichao | 20,856,711 | MDQ6VXNlcjIwODU2NzEx | User | false |
openclaw/openclaw | 4,336,653,031 | I_kwDOQb6kR88AAAABAnwS5w | 72,860 | https://github.com/openclaw/openclaw/issues/72860 | https://api.github.com/repos/openclaw/openclaw/issues/72860 | Telegram ordinary assistant replies can lose image attachments while text still delivers | ## Summary
Telegram ordinary assistant replies could lose image attachments even though:
- direct outbound media sends worked,
- `routeReply(...)` with `mediaUrls` worked,
- and the final user-visible reply text still arrived.
This made the bug look like a Telegram delivery problem, but it was actually in OpenClaw's ... | open | null | false | 1 | [] | [] | 2026-04-27T14:50:26Z | 2026-05-03T01:38:08Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | cslash-zz | 674,096 | MDQ6VXNlcjY3NDA5Ng== | User | false |
openclaw/openclaw | 4,336,673,515 | I_kwDOQb6kR88AAAABAnxi6w | 72,862 | https://github.com/openclaw/openclaw/issues/72862 | https://api.github.com/repos/openclaw/openclaw/issues/72862 | Windows: Feishu channel exits with ESM loader error 'Received protocol c:' on OpenClaw 2026.4.24 | ## Environment
- OS: Windows 11 / Windows_NT 10.0.26100 x64
- Node.js: v24.14.1
- OpenClaw: 2026.4.24 (cbcfdf6)
- Install: global npm install
## Problem
The gateway stays up, but Feishu channels fail to initialize and keep auto-restarting.
## Observed log
```
[gateway/channels/feishu] [work] channel exited: Only URLs... | closed | completed | true | 2 | [] | [] | 2026-04-27T14:53:19Z | 2026-04-30T04:32:40Z | 2026-04-28T01:00:45Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | One9906 | 274,272,696 | U_kgDOEFkRuA | User | false |
openclaw/openclaw | 4,336,666,746 | I_kwDOQb6kR88AAAABAnxIeg | 72,861 | https://github.com/openclaw/openclaw/issues/72861 | https://api.github.com/repos/openclaw/openclaw/issues/72861 | [Bug]: Control UI / webchat duplicates user message and leaks `Sender (untrusted metadata)` as `openclaw-control-ui` | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Control UI / `webchat` can duplicate a single user message into two visible user turns:
- one normal local/plain user message
- one echoed user message whose content begins with `Sender (untrusted metadata):` and whose send... | closed | completed | true | 5 | [] | [] | 2026-04-27T14:52:20Z | 2026-05-03T04:36:59Z | 2026-04-30T14:32:34Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | cslash-zz | 674,096 | MDQ6VXNlcjY3NDA5Ng== | User | false |
openclaw/openclaw | 4,336,399,068 | I_kwDOQb6kR88AAAABAngy3A | 72,846 | https://github.com/openclaw/openclaw/issues/72846 | https://api.github.com/repos/openclaw/openclaw/issues/72846 | [Bug]: Regression — channel sidecar startup again blocks for ~3 min after `ready` on v2026.4.25 (recurrence of #63450) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The bug fixed in #63450 / PR #63480 ("Gateway channel sidecar startup blocked by chat.history WS request, ~80–110 s delay since v2026.4.8", closed 2026-04-09) has returned in v2026.4.25 with a longer delay (~180 s instead of... | closed | completed | true | 13 | [
"bug",
"gateway",
"regression"
] | [] | 2026-04-27T14:15:00Z | 2026-04-30T04:32:17Z | 2026-04-28T02:57:28Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | RayWoo | 2,201,861 | MDQ6VXNlcjIyMDE4NjE= | User | false |
openclaw/openclaw | 4,336,576,552 | I_kwDOQb6kR88AAAABAnroKA | 72,857 | https://github.com/openclaw/openclaw/issues/72857 | https://api.github.com/repos/openclaw/openclaw/issues/72857 | PR #22280 Regression: scope upgrade still requires pairing on 2026.4.25 (VPS/lan bind) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Despite PR #22280 being merged (which fixes "Local paired-device scope upgrades can require interactive pairing"), the issue persists on OpenClaw 2026.4.25 with a VPS/lan bind configuration.
### Steps to reproduce
1. **Ins... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-27T14:40:24Z | 2026-04-30T04:33:45Z | 2026-04-27T15:45:37Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | JDE1990 | 184,231,923 | U_kgDOCvsn8w | User | false |
openclaw/openclaw | 4,338,865,221 | I_kwDOQb6kR88AAAABAp3URQ | 73,043 | https://github.com/openclaw/openclaw/issues/73043 | https://api.github.com/repos/openclaw/openclaw/issues/73043 | Lossless/pi-ai `completeSimple` fails with `Connection error` for OpenClaw `claude-cli` Sonnet transport | ## Summary
Lossless Claw / pi-ai `completeSimple` fails when using OpenClaw's synthetic `claude-cli/claude-sonnet-4-6` provider path, even though raw Claude CLI Sonnet works and OpenClaw runtime auth can read Claude CLI OAuth credentials from Keychain.
This looks like a transport integration issue between Lossless Cl... | open | null | false | 1 | [] | [] | 2026-04-27T21:02:43Z | 2026-05-03T07:45:25Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | nonzeroclaw | 267,426,379 | U_kgDOD_CaSw | User | false |
openclaw/openclaw | 4,338,936,954 | I_kwDOQb6kR88AAAABAp7seg | 73,049 | https://github.com/openclaw/openclaw/issues/73049 | https://api.github.com/repos/openclaw/openclaw/issues/73049 | v2026.4.25 regression: memory-core breaks agent by using wrong API key for embeddings | ## Bug Description
After upgrading from 2026.4.24 to 2026.4.25, the `memory-core` plugin breaks agent message processing by using the wrong API key for embeddings.
## What changed
Before (2026.4.24): Agent used `deepseek/deepseek-v4-pro` via the deepseek provider at `https://api.deepseek.com`. Memory sync worked fin... | open | null | false | 1 | [] | [] | 2026-04-27T21:14:59Z | 2026-05-03T08:39:46Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | amdhelper | 232,188,533 | U_kgDODdbqdQ | User | false |
openclaw/openclaw | 4,338,501,251 | I_kwDOQb6kR88AAAABAphGgw | 73,010 | https://github.com/openclaw/openclaw/issues/73010 | https://api.github.com/repos/openclaw/openclaw/issues/73010 | [Bug]: BUG_REPORT_OpenClaw_v2026.4.25_Windows_ESM_Path_Regression | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
# BUG: Windows ESM loader fails for bundled plugins after cold registry update in v2026.4.25
- **Version**: v2026.4.25
- **OS**: Windows
- **Component**: Plugin Loader, Bundled Channels (e.g., Telegram, iMessage)
## Summa... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-27T19:55:39Z | 2026-04-30T04:33:04Z | 2026-04-27T20:14:20Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | 2ndNuts | 264,120,928 | U_kgDOD74qYA | User | false |
openclaw/openclaw | 4,338,974,453 | I_kwDOQb6kR88AAAABAp9-9Q | 73,051 | https://github.com/openclaw/openclaw/issues/73051 | https://api.github.com/repos/openclaw/openclaw/issues/73051 | [Feature]: Move EVERYTHING to another LLM | ### Summary
One command to move everything to a different model when rate limited from a provider.
### Problem to solve
When you get rate limited which is increasingly common you need a way to move everything to another provider. You think you changed your main agent - fine, but then you find all your sub agents, cr... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-27T21:23:45Z | 2026-05-03T08:39:47Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | ceramicv | 102,924,988 | U_kgDOBiKCvA | User | false |
openclaw/openclaw | 4,338,888,458 | I_kwDOQb6kR88AAAABAp4vCg | 73,045 | https://github.com/openclaw/openclaw/issues/73045 | https://api.github.com/repos/openclaw/openclaw/issues/73045 | [Bug]: Visible cmd.exe console windows flash on Windows from internal child_process.exec() calls (network discovery loop) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
**Environment**
OS: Windows 10 Pro 22H2 (build 19045.6466), running directly on Windows (not WSL2)
OpenClaw version: 2026.4.2 (d74a122)
Node: v24.14.1
Install method: npm install -g openclaw@latest
Process mana... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-27T21:06:30Z | 2026-04-30T04:33:03Z | 2026-04-27T21:24:04Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | gooseanddragon68-cell | 279,841,908 | U_kgDOEK4MdA | User | false |
openclaw/openclaw | 4,338,875,090 | I_kwDOQb6kR88AAAABAp360g | 73,044 | https://github.com/openclaw/openclaw/issues/73044 | https://api.github.com/repos/openclaw/openclaw/issues/73044 | [Bug]: gateway 1006 disconnect loop on macOS native tied to bonjour/CIAO advertiser (2026.4.24) | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
On macOS native in OpenClaw 2026.4.24, the local gateway repeatedly hits bonjour/CIAO unhandled rejections and clients intermittently disconnect with gateway closed (1006).
### Steps to reproduce
1. Start OpenClaw 2026.4.24 o... | closed | completed | true | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-27T21:04:26Z | 2026-04-30T04:33:02Z | 2026-04-27T21:24:54Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | Noxiffow | 130,965,846 | U_kgDOB85hVg | User | false |
openclaw/openclaw | 4,338,807,422 | I_kwDOQb6kR88AAAABApzyfg | 73,036 | https://github.com/openclaw/openclaw/issues/73036 | https://api.github.com/repos/openclaw/openclaw/issues/73036 | [Bug]: CPU Regression: v2026.4.25 Cold Plugin Registry Refactor Fails to Separate Cold/Hot Paths | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to v2026.4.25, every inbound message causes the openclaw-gateway process to immediately spike to sustained ~1 core (24-core machine) and stays elevated for the full duration of message processing. This did NO... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-04-27T20:51:42Z | 2026-04-30T04:33:01Z | 2026-04-27T21:25:10Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | keaideweiertuode | 59,241,730 | MDQ6VXNlcjU5MjQxNzMw | User | false |
openclaw/openclaw | 4,337,632,896 | I_kwDOQb6kR88AAAABAosGgA | 72,953 | https://github.com/openclaw/openclaw/issues/72953 | https://api.github.com/repos/openclaw/openclaw/issues/72953 | [Bug]:Gateway crashes on macOS when hostname contains apostrophe (Bonjour / ciao) | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
Gateway process crashes during startup when hostname contains an apostrophe, with CIAO PROBING CANCELLED error in logs.
### Steps to reproduce
1. Set macOS hostname to include an apostrophe, e.g. "test's-mac"
2. Install OpenC... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-27T17:25:36Z | 2026-04-30T04:33:00Z | 2026-04-27T21:29:07Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | oldsix-cell | 189,520,719 | U_kgDOC0vbTw | User | false |
openclaw/openclaw | 4,337,130,493 | I_kwDOQb6kR88AAAABAoNb_Q | 72,902 | https://github.com/openclaw/openclaw/issues/72902 | https://api.github.com/repos/openclaw/openclaw/issues/72902 | [Bug]: Bonjour/mdns broadcaster crashes gateway on macOS (4.24) | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
After upgrading from 2026.4.14 to 2026.4.24, the OpenClaw gateway crashes within ~30 seconds of startup on macOS when the Bonjour/mdns broadcaster encounters errors. Gateway logs show repeated CIAO/bonjour errors before process... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-27T16:05:56Z | 2026-04-30T04:32:59Z | 2026-04-27T21:29:09Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | masumbevi | 69,218,764 | MDQ6VXNlcjY5MjE4NzY0 | User | false |
openclaw/openclaw | 4,338,504,645 | I_kwDOQb6kR88AAAABAphTxQ | 73,011 | https://github.com/openclaw/openclaw/issues/73011 | https://api.github.com/repos/openclaw/openclaw/issues/73011 | [Bug]: plugin manifest loader uses JSON5.parse for plain JSON files (~3x slower gateway boot) | ## Bug type
Bug (`loadPluginManifest` parses every plugin manifest with `JSON5.parse`, which is much slower than native `JSON.parse` and the manifests don't use any JSON5-specific syntax)
## Beta release blocker
No
## Summary
Plugin manifests are parsed with `JSON5.parse`. On this Pi that path runs ~10x slower tha... | closed | completed | true | 2 | [] | [] | 2026-04-27T19:56:15Z | 2026-04-30T04:32:39Z | 2026-04-28T00:57:50Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | jasonftl | 175,325,999 | U_kgDOCnNDLw | User | false |
openclaw/openclaw | 4,337,826,285 | I_kwDOQb6kR88AAAABAo357Q | 72,965 | https://github.com/openclaw/openclaw/issues/72965 | https://api.github.com/repos/openclaw/openclaw/issues/72965 | active-memory plugin: timeoutMs=30000 ignored — abort signal doesn't reach HTTP fetch, runs 65s | ## Environment
- **OpenClaw version**: 2026.4.25 (aa36ee6)
- **Node.js**: v22.22.2
- **OS**: Debian 12, Linux 6.1.0-34-amd64
## Problem
The `active-memory` plugin has `timeoutMs: 30000` configured, but the actual elapsed time was **65250ms** — over double the configured limit. The configured timeout is completely in... | closed | completed | true | 3 | [] | [] | 2026-04-27T17:56:28Z | 2026-05-01T04:44:34Z | 2026-04-28T08:15:25Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | rollingshmily | 62,282,098 | MDQ6VXNlcjYyMjgyMDk4 | User | false |
openclaw/openclaw | 4,339,122,551 | I_kwDOQb6kR88AAAABAqHBdw | 73,056 | https://github.com/openclaw/openclaw/issues/73056 | https://api.github.com/repos/openclaw/openclaw/issues/73056 | Outer compaction pre-check misclassifies tool-heavy sessions as idle | ## Symptom
Manual `/compact` on a long, tool-heavy Discord-thread session reports:
```
Compaction skipped: no real conversation messages yet • Context 236k/200k (118%)
```
The session is clearly not idle — context is at 118%.
## Why this is a separate issue from #44138
#44138 was closed as fixed by the `messagesTo... | open | null | false | 1 | [] | [] | 2026-04-27T21:58:19Z | 2026-05-03T11:49:58Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | petesimard | 552,928 | MDQ6VXNlcjU1MjkyOA== | User | false |
openclaw/openclaw | 4,336,840,472 | I_kwDOQb6kR88AAAABAn7vGA | 72,872 | https://github.com/openclaw/openclaw/issues/72872 | https://api.github.com/repos/openclaw/openclaw/issues/72872 | [Bug]: 2026.4.24 schema validator rejects `agents.defaults.agentRuntime` from prior 4.x configs; gates ALL CLI commands including `doctor --fix` | ### Bug type
Regression (worked before, now fails) — config schema rejects a key the prior release wrote.
### Summary
After upgrading from an earlier 2026.4.x release to **2026.4.24** via `npm`, every `openclaw <subcmd>` invocation fails with a config-validation error before the subcommand runs. The validator reject... | closed | completed | true | 2 | [] | [] | 2026-04-27T15:18:44Z | 2026-05-02T04:29:24Z | 2026-04-29T05:09:44Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | 1fanwang | 21,269,208 | MDQ6VXNlcjIxMjY5MjA4 | User | false |
openclaw/openclaw | 4,339,166,262 | I_kwDOQb6kR88AAAABAqJsNg | 73,059 | https://github.com/openclaw/openclaw/issues/73059 | https://api.github.com/repos/openclaw/openclaw/issues/73059 | [Bug]: Windows linked source install can skip Control UI auto-build through junction-launched scripts/ui.js | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On Windows source installs linked through the global npm junction, the first-start Control UI auto-build can report success without building anything because scripts/ui.js does not run main() when launched thro... | open | null | false | 2 | [] | [] | 2026-04-27T22:08:51Z | 2026-05-03T08:39:49Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | LouisGameDev | 1,916,460 | MDQ6VXNlcjE5MTY0NjA= | User | false |
openclaw/openclaw | 4,339,240,422 | I_kwDOQb6kR88AAAABAqON5g | 73,062 | https://github.com/openclaw/openclaw/issues/73062 | https://api.github.com/repos/openclaw/openclaw/issues/73062 | [Feature]: Support for WPS Office Global (wps.com) in the kdocs skill | ### Summary
## Feature Request: Support for WPS Office Global (wps.com) in the kdocs skill
### Summary
The current `kdocs` skill and `kdocs-cli` tool are exclusively integrated with the Chinese ecosystem (`kdocs.cn` / `wps.cn`). Users of WPS Office Global (`wps.com`) have no way to connect OpenClaw to their cloud doc... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-27T22:27:57Z | 2026-05-03T04:36:41Z | 2026-04-30T14:54:38Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | lexcorp | 5,751,288 | MDQ6VXNlcjU3NTEyODg= | User | false |
openclaw/openclaw | 4,338,808,471 | I_kwDOQb6kR88AAAABApz2lw | 73,037 | https://github.com/openclaw/openclaw/issues/73037 | https://api.github.com/repos/openclaw/openclaw/issues/73037 | Source-checkout CLI can use stale compile cache after rebuild | ## Summary
In a live source-checkout install, the `openclaw` CLI can execute stale code after a local rebuild because the bootstrap enables Node compile cache unconditionally.
That makes `openclaw gateway status --json` report misleading probe failures even though the rebuilt source implementation and the running gate... | closed | completed | true | 1 | [] | [] | 2026-04-27T20:51:57Z | 2026-04-30T04:32:56Z | 2026-04-27T22:29:05Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | LouisGameDev | 1,916,460 | MDQ6VXNlcjE5MTY0NjA= | User | false |
openclaw/openclaw | 4,339,255,587 | I_kwDOQb6kR88AAAABAqPJIw | 73,065 | https://github.com/openclaw/openclaw/issues/73065 | https://api.github.com/repos/openclaw/openclaw/issues/73065 | [Bug]: Matrix plugin verify- Self-verification failed: the signing key is missing from the object that signed the message | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
I build version OpenClaw 2026.4.26 (5ccf179) from dev channel.
when I try to verify matrix channel this error happens:
Self-verification failed: the signing key is missing from the object that signed the messag... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-27T22:31:46Z | 2026-05-03T04:35:51Z | 2026-04-30T17:04:49Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | DHosseiny | 9,461,116 | MDQ6VXNlcjk0NjExMTY= | User | false |
openclaw/openclaw | 4,338,698,277 | I_kwDOQb6kR88AAAABAptIJQ | 73,023 | https://github.com/openclaw/openclaw/issues/73023 | https://api.github.com/repos/openclaw/openclaw/issues/73023 | Ollama provider routing fails silently — agent fabricates successful local-model responses | # Ollama Provider Routing Bug
**Environment**: OpenClaw 2026.4.25 on WSL2 (Ubuntu 20.04), Ollama on Windows host.
**Steps to Reproduce**:
1. Configure ollama provider with host/baseUrl pointing to Windows Ollama.
2. Attempt model switch to ollama/qwen3.5:27b.
3. Observe agent claims successful switch and generates re... | closed | completed | true | 1 | [] | [] | 2026-04-27T20:31:27Z | 2026-04-30T04:32:55Z | 2026-04-27T22:33:05Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | pavelyortho-cyber | 279,129,856 | U_kgDOEKMvAA | User | false |
openclaw/openclaw | 4,339,274,916 | I_kwDOQb6kR88AAAABAqQUpA | 73,068 | https://github.com/openclaw/openclaw/issues/73068 | https://api.github.com/repos/openclaw/openclaw/issues/73068 | [Bug]: msteams "JWT Validation Failed" on 2026.4.25+ | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to `2026.4.25`, the msteams channel is broken. The bot doesn't respond to anyone and I see this in the logs when a message is sent. It's also happening on `2026.4.26-beta.1`
I'm aware of #70724 but that lan... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-27T22:36:59Z | 2026-05-01T04:43:50Z | 2026-04-28T10:35:24Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | dseravalli | 1,758,219 | MDQ6VXNlcjE3NTgyMTk= | User | false |
openclaw/openclaw | 4,336,851,432 | I_kwDOQb6kR88AAAABAn8Z6A | 72,873 | https://github.com/openclaw/openclaw/issues/72873 | https://api.github.com/repos/openclaw/openclaw/issues/72873 | Telegram: handle edited_message updates (streaming bot messages appear truncated) | ## Problem
When another bot sends a streaming message in a Telegram group (e.g. editing the same message repeatedly as tokens arrive), OpenClaw only processes the initial `message` update and ignores all subsequent `edited_message` updates. This results in the agent seeing only the first partial chunk of the streamed ... | open | null | false | 2 | [] | [] | 2026-04-27T15:20:29Z | 2026-05-03T03:37:30Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | MapleRecall | 18,360,825 | MDQ6VXNlcjE4MzYwODI1 | User | false |
openclaw/openclaw | 4,339,298,543 | I_kwDOQb6kR88AAAABAqRw7w | 73,071 | https://github.com/openclaw/openclaw/issues/73071 | https://api.github.com/repos/openclaw/openclaw/issues/73071 | [Bug]: ACP turns fail with "Internal error" after upgrade to 2026.4.25 — gateway WS handshake rejected | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading from OpenClaw 2026.4.24 to 2026.4.25, all ACP sessions_spawn calls fail instantly (~2s, 0 tokens) with AcpRuntimeError: Internal error / ACP_TURN_FAILED; gateway logs show a WS client sending sessions.list be... | closed | completed | true | 0 | [
"bug"
] | [] | 2026-04-27T22:44:01Z | 2026-05-01T04:43:34Z | 2026-04-28T12:35:27Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | bobtrahan | 641,552 | MDQ6VXNlcjY0MTU1Mg== | User | false |
openclaw/openclaw | 4,339,252,288 | I_kwDOQb6kR88AAAABAqO8QA | 73,064 | https://github.com/openclaw/openclaw/issues/73064 | https://api.github.com/repos/openclaw/openclaw/issues/73064 | [Bug]: Subagent runs can be stranded or mis-finalized after wait timeout/pending edges, and cleanup can remove the main transcript | ## Bug type
Behavior bug (incorrect output/state without crash)
## Beta release blocker
No
## Summary
Across multiple recent `2026.4.x` releases, subagent runs can fail in two related ways: the child session may be created but remain empty/idle without a useful parent notification, and runs that did make progress can... | closed | not_planned | true | 1 | [] | [] | 2026-04-27T22:30:54Z | 2026-05-03T04:36:40Z | 2026-04-30T14:54:45Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | vrurg | 1,586,992 | MDQ6VXNlcjE1ODY5OTI= | User | false |
openclaw/openclaw | 4,337,526,071 | I_kwDOQb6kR88AAAABAollNw | 72,940 | https://github.com/openclaw/openclaw/issues/72940 | https://api.github.com/repos/openclaw/openclaw/issues/72940 | [Bug]: 2026.4.25 — heartbeat.model override ignored; heartbeat runs on session primary then falls back to gpt-5 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to 2026.4.25, the `heartbeat.model` override is silently ignored. Heartbeat runs on `openai/gpt-5` (the first fallback) instead of the configured `openrouter/anthropic/claude-haiku-4.5`.
**Config (unchanged ... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-27T17:08:20Z | 2026-04-30T04:32:10Z | 2026-04-27T21:47:42Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | shrimpy8 | 642,975 | MDQ6VXNlcjY0Mjk3NQ== | User | false |
openclaw/openclaw | 4,339,067,221 | I_kwDOQb6kR88AAAABAqDpVQ | 73,053 | https://github.com/openclaw/openclaw/issues/73053 | https://api.github.com/repos/openclaw/openclaw/issues/73053 | OpenAI Responses replay still poisons sessions on GPT-5.5/GPT-5.4 in 2026.4.25 | ## Summary
OpenClaw `2026.4.25` still appears to have an OpenAI Responses replay/conversation-state bug that can poison long-lived sessions on current GPT reasoning models.
This is related to the older family of issues around:
- #18282
- #19184
- #19582
- fixed variant in PR #18632
…but this repro appears broader th... | closed | completed | true | 1 | [] | [] | 2026-04-27T21:45:33Z | 2026-04-30T04:32:53Z | 2026-04-27T22:55:09Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | odb36777 | 270,791,998 | U_kgDOECP1Pg | User | false |
openclaw/openclaw | 4,337,456,774 | I_kwDOQb6kR88AAAABAohWhg | 72,933 | https://github.com/openclaw/openclaw/issues/72933 | https://api.github.com/repos/openclaw/openclaw/issues/72933 | [Bug]: plugin mirror rebuilt on every plugin load | ## Bug type
Bug (`mirrorBundledPluginRuntimeRoot` rebuilds plugin tree on every load)
## Beta release blocker
No
## Summary
`mirrorBundledPluginRuntimeRoot()` rebuilds the plugin's mirror directory from scratch on every load, even when the source files have not changed. On my Pi 500 each channel reconnect or model... | closed | completed | true | 1 | [] | [] | 2026-04-27T16:57:07Z | 2026-04-30T04:32:53Z | 2026-04-27T22:57:46Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | jasonftl | 175,325,999 | U_kgDOCnNDLw | User | false |
openclaw/openclaw | 4,339,370,145 | I_kwDOQb6kR88AAAABAqWIoQ | 73,074 | https://github.com/openclaw/openclaw/issues/73074 | https://api.github.com/repos/openclaw/openclaw/issues/73074 | [Bug]:Gateway crash loop: Unhandled promise rejection from Bonjour/ciao mDNS advertiser | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
The Bonjour mDNS advertiser gets stuck in "announcing" state after a network change, causing an unhandled promise rejection that crashes the gateway in an infinite restart loop.
### Steps to reproduce
1. Run OpenClaw gateway ... | closed | completed | true | 1 | [
"bug",
"bug:crash"
] | [
"vincentkoc"
] | 2026-04-27T23:01:23Z | 2026-04-30T04:32:07Z | 2026-04-28T03:44:58Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | tomrob6 | 279,858,554 | U_kgDOEK5Neg | User | false |
openclaw/openclaw | 4,339,372,396 | I_kwDOQb6kR88AAAABAqWRbA | 73,075 | https://github.com/openclaw/openclaw/issues/73075 | https://api.github.com/repos/openclaw/openclaw/issues/73075 | perf: resolvePluginProviders called repeatedly on ARM64, active-memory subagent takes 66-96s | ## Summary
On ARM64 devices (tested: Raspberry Pi 4, 4 GB RAM), the `active-memory` plugin's subagent takes **66–96 seconds** from message receipt to first LLM call. On x86 this is likely masked by faster CPUs, but the underlying cache bugs affect all platforms.
Three root-cause bugs cause `resolvePluginProviders` (~... | closed | completed | true | 0 | [] | [] | 2026-04-27T23:01:52Z | 2026-04-30T04:32:31Z | 2026-04-28T01:38:33Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | jochen | 113,530 | MDQ6VXNlcjExMzUzMA== | User | false |
openclaw/openclaw | 4,339,386,463 | I_kwDOQb6kR88AAAABAqXIXw | 73,078 | https://github.com/openclaw/openclaw/issues/73078 | https://api.github.com/repos/openclaw/openclaw/issues/73078 | sandbox.mode "non-main" semantics: per-session, not per-agent — counter-intuitive for "main-agent-only" use case | ## Summary
`agents.defaults.sandbox.mode = "non-main"` reads as "non-main agents are sandboxed", but the actual logic in `runtime-status.ts` treats it per-**session**, not per-**agent**. Since every agent has its own main session, `"non-main"` mode causes **every** agent's primary conversation (including dynamic peer ... | open | null | false | 1 | [] | [] | 2026-04-27T23:05:37Z | 2026-05-03T09:13:18Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | jimmyag2026-prog | 278,308,846 | U_kgDOEJan7g | User | false |
openclaw/openclaw | 4,339,375,554 | I_kwDOQb6kR88AAAABAqWdwg | 73,077 | https://github.com/openclaw/openclaw/issues/73077 | https://api.github.com/repos/openclaw/openclaw/issues/73077 | [Bug]: parent CLI commands exit 1 when invoked without a subcommand (memory, channels, plugins, approvals, devices, cron, mcp) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Several `openclaw <parent>` commands exit with code 1 when invoked with no subcommand, while printing the same help-style content that `openclaw <parent> --help` prints (which exits 0). This breaks shell `&&` c... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-27T23:02:36Z | 2026-04-30T04:32:30Z | 2026-04-28T01:40:46Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | iot2edge | 188,053,512 | U_kgDOCzV4CA | User | false |
openclaw/openclaw | 4,339,405,039 | I_kwDOQb6kR88AAAABAqYQ7w | 73,082 | https://github.com/openclaw/openclaw/issues/73082 | https://api.github.com/repos/openclaw/openclaw/issues/73082 | Feature Request: auto_approve_scope: readonly for healthcheck/canary skills | # Feature Request: `auto_approve_scope: readonly` for healthcheck/canary skills
## Problem
Canary/healthcheck/status-check skills are designed to run **automatically** to detect anomalies. But every introspection command they run hits the exec approvals gate per-invocation, requiring user approval.
Two compounding i... | open | null | false | 1 | [] | [] | 2026-04-27T23:11:12Z | 2026-05-03T06:50:17Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | etadward | 200,764,463 | U_kgDOC_dsLw | User | false |
openclaw/openclaw | 4,339,011,227 | I_kwDOQb6kR88AAAABAqAOmw | 73,052 | https://github.com/openclaw/openclaw/issues/73052 | https://api.github.com/repos/openclaw/openclaw/issues/73052 | Codex ACP rejects timeout config option during sessions_spawn | When OpenClaw ACP runtime spawns Codex ACP with `runTimeoutSeconds` / `timeoutSeconds`, OpenClaw forwards the runtime timeout as ACP `session/set_config_option` key `timeout`. The current `@zed-industries/codex-acp` adapter rejects that key with `Invalid params (ACP -32602)`, so the child task fails before running.
I ... | closed | completed | true | 3 | [] | [] | 2026-04-27T21:32:01Z | 2026-04-30T04:32:52Z | 2026-04-27T23:14:42Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | richa65 | 40,608,992 | MDQ6VXNlcjQwNjA4OTky | User | false |
openclaw/openclaw | 4,339,108,345 | I_kwDOQb6kR88AAAABAqGJ-Q | 73,055 | https://github.com/openclaw/openclaw/issues/73055 | https://api.github.com/repos/openclaw/openclaw/issues/73055 | Gateway blocks HTTP server startup while reinstalling 63 bundled runtime deps on every restart | ## Problem
When the OpenClaw gateway crashes or is restarted, it spends ~5 minutes reinstalling 63 bundled runtime dependency packages before the HTTP server becomes available on `/health`.
This causes:
- Gateway health checks timeout for ~5 minutes post-restart
- Channels (Telegram etc.) fail to initialize during t... | closed | completed | true | 1 | [] | [] | 2026-04-27T21:54:47Z | 2026-04-30T04:32:51Z | 2026-04-27T23:15:52Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | GCorp2026 | 231,491,002 | U_kgDODcxFug | User | false |
openclaw/openclaw | 4,339,511,644 | I_kwDOQb6kR88AAAABAqexXA | 73,088 | https://github.com/openclaw/openclaw/issues/73088 | https://api.github.com/repos/openclaw/openclaw/issues/73088 | Plugin discovery in infinite loop after upgrade — gateway pegs one CPU core continuously (v2026.4.25) | # Plugin discovery in infinite loop after upgrade — gateway pegs one CPU core continuously (v2026.4.25)
## Summary
After upgrading from `2026.2.21` to `2026.4.25`, the gateway process sustains ~100% CPU on one core indefinitely while idle (no chat activity, no commands). `perf` and `strace` show the gateway is in a t... | closed | completed | true | 2 | [] | [] | 2026-04-27T23:40:01Z | 2026-04-30T04:32:42Z | 2026-04-28T00:43:57Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | sleitor | 14,165,771 | MDQ6VXNlcjE0MTY1Nzcx | User | false |
openclaw/openclaw | 4,339,520,660 | I_kwDOQb6kR88AAAABAqfUlA | 73,090 | https://github.com/openclaw/openclaw/issues/73090 | https://api.github.com/repos/openclaw/openclaw/issues/73090 | Claude CLI-backed sessions show mixed modelProvider metadata; sessions.patch model:null succeeds but does not normalize stale runtime fields | ### Bug type
Behavior bug / session metadata inconsistency
### Summary
After configuring agents to use the Claude CLI runtime with a canonical Anthropic model ref, active sessions can report mixed `modelProvider` values for the same effective setup.
Config uses the documented split:
- model: `anthropic/claude-opus... | closed | completed | true | 1 | [] | [] | 2026-04-27T23:42:23Z | 2026-05-02T04:29:25Z | 2026-04-29T05:08:02Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | vishutdhar | 68,405,187 | MDQ6VXNlcjY4NDA1MTg3 | User | false |
openclaw/openclaw | 4,339,519,900 | I_kwDOQb6kR88AAAABAqfRnA | 73,089 | https://github.com/openclaw/openclaw/issues/73089 | https://api.github.com/repos/openclaw/openclaw/issues/73089 | [Bug]: Changelog lists 2026.4.26 released, but npm stable install target is missing | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The changelog shows 2026.4.26 under released versions, but npm currently reports:
Latest dist-tag: 2026.4.25.
Available versions include 2026.4.26-beta.1 but not stable 2026.4.26
Repro:
npm view openclaw dist... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-27T23:42:12Z | 2026-05-01T04:43:08Z | 2026-04-28T04:52:54Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Joe118 | 12,119,280 | MDQ6VXNlcjEyMTE5Mjgw | User | false |
openclaw/openclaw | 4,339,525,033 | I_kwDOQb6kR88AAAABAqflqQ | 73,091 | https://github.com/openclaw/openclaw/issues/73091 | https://api.github.com/repos/openclaw/openclaw/issues/73091 | [Bug]: Signal voice-note replies skip auto TTS when messages.tts.auto = "inbound" | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On OpenClaw `2026.4.23`, Signal direct-chat voice notes are processed and transcribed successfully, but `messages.tts.auto = "inbound"` does not trigger a voice reply; the assistant replies in plain text and th... | closed | completed | true | 1 | [] | [] | 2026-04-27T23:43:32Z | 2026-05-01T04:42:43Z | 2026-04-28T17:02:27Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | ScientificProgrammer | 4,348,294 | MDQ6VXNlcjQzNDgyOTQ= | User | false |
openclaw/openclaw | 4,338,209,541 | I_kwDOQb6kR88AAAABApPTBQ | 72,993 | https://github.com/openclaw/openclaw/issues/72993 | https://api.github.com/repos/openclaw/openclaw/issues/72993 | [Bug]: 2026.4.25 openclaw status is slow and shows empty Channels table | # [Bug]: 2026.4.25 `openclaw status` is slow and shows empty Channels table while Telegram is connected
## Bug type
Behavior bug (incorrect output/state without crash)
## Beta release blocker
No
## Summary
On OpenClaw 2026.4.25 (aa36ee6), `openclaw status` takes ~52s on a live install, renders an empty Channels tabl... | closed | completed | true | 3 | [] | [] | 2026-04-27T19:03:14Z | 2026-04-30T04:32:50Z | 2026-04-27T23:44:35Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | comick1 | 13,777,442 | MDQ6VXNlcjEzNzc3NDQy | User | false |
openclaw/openclaw | 4,339,128,579 | I_kwDOQb6kR88AAAABAqHZAw | 73,057 | https://github.com/openclaw/openclaw/issues/73057 | https://api.github.com/repos/openclaw/openclaw/issues/73057 | Codex app-server inherits OPENAI_API_KEY from gateway and switches to API key mode | ## Summary
When OpenClaw starts the native Codex app-server via stdio, the child process inherits the full gateway environment, including `OPENAI_API_KEY`. On a mixed setup where OpenClaw needs `OPENAI_API_KEY` for OpenAI embeddings/providers but Codex CLI is logged in with ChatGPT/OAuth, the Codex app-server UI/sessi... | closed | completed | true | 0 | [] | [
"steipete"
] | 2026-04-27T21:59:47Z | 2026-04-30T04:32:48Z | 2026-04-27T23:46:16Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | holgergruenhagen | 60,438,604 | MDQ6VXNlcjYwNDM4NjA0 | User | false |
openclaw/openclaw | 4,339,366,725 | I_kwDOQb6kR88AAAABAqV7RQ | 73,073 | https://github.com/openclaw/openclaw/issues/73073 | https://api.github.com/repos/openclaw/openclaw/issues/73073 | Telegram DM: cliSessionBindings.claude-cli.sessionId not updated after session_expired retry — duplicate cli exec on every turn, mid-conversation amnesia | ## Summary
In OpenClaw 2026.4.21 / 2026.4.24, the per-channel `cliSessionBindings.claude-cli.sessionId` for a Telegram-direct session can fossilize on a session id whose `~/.claude/projects/<encoded-cwd>/<id>.jsonl` no longer exists. Once that happens, every subsequent turn:
1. Tries to resume the dead id (`executePr... | closed | completed | true | 3 | [] | [] | 2026-04-27T23:00:43Z | 2026-05-03T04:36:39Z | 2026-04-30T14:57:38Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | MS-707 | 201,992,406 | U_kgDODAoo1g | User | false |
openclaw/openclaw | 4,339,592,375 | I_kwDOQb6kR88AAAABAqjstw | 73,095 | https://github.com/openclaw/openclaw/issues/73095 | https://api.github.com/repos/openclaw/openclaw/issues/73095 | `prepareContextEngineSubagentSpawn` calls `resolveContextEngine` without initializing the registry — fails 100% on cliBackend-only installs | ## Affected version
OpenClaw `2026.4.24` (build `cbcfdf6`), installed via pnpm. Likely affects any version that bundles the lazy-init pattern shown below.
## Symptom
First (and every) `mcp__openclaw__sessions_spawn` call on an install that routes chat through `cliBackends` (Anthropic CLI backend, OAuth-mode auth pro... | closed | completed | true | 5 | [] | [] | 2026-04-28T00:01:00Z | 2026-05-04T04:34:55Z | 2026-05-01T23:50:26Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | datus1982 | 278,869,131 | U_kgDOEJ80iw | User | false |
openclaw/openclaw | 4,339,592,436 | I_kwDOQb6kR88AAAABAqjs9A | 73,096 | https://github.com/openclaw/openclaw/issues/73096 | https://api.github.com/repos/openclaw/openclaw/issues/73096 | `resolvePinnedAgentHarnessPolicy` hardcodes `fallback: "none"` — config setting is silently ignored, and the safety brake is the only thing preventing silent extra-usage billing | ## Affected version
OpenClaw `2026.4.24` (build `cbcfdf6`).
## Symptom
Setting `agents.defaults.embeddedHarness.fallback: "pi"` in `openclaw.json` does not change the behavior of any code path that goes through the **pinned** harness policy (subagent execution, channel boot, model-fallback chain dispatch when an age... | closed | completed | false | 1 | [] | [] | 2026-04-28T00:01:01Z | 2026-05-03T06:41:43Z | 2026-05-03T06:41:43Z | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | datus1982 | 278,869,131 | U_kgDOEJ80iw | User | false |
openclaw/openclaw | 4,339,592,549 | I_kwDOQb6kR88AAAABAqjtZQ | 73,097 | https://github.com/openclaw/openclaw/issues/73097 | https://api.github.com/repos/openclaw/openclaw/issues/73097 | PI harness ignores `cliBackends` configuration, splitting subagent execution and chat path onto different transports for the same OAuth credential | ## Affected version
OpenClaw `2026.4.24` (build `cbcfdf6`).
## Summary
Anthropic's Claude CLI binary is **exempt** from the "third-party apps now draw from your extra usage, not your plan limits" filter, but Anthropic's direct API treats OAuth-credential-as-bearer requests as "third-party app" usage. OpenClaw chat g... | open | null | false | 1 | [] | [] | 2026-04-28T00:01:02Z | 2026-05-03T09:49:13Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | datus1982 | 278,869,131 | U_kgDOEJ80iw | User | false |
openclaw/openclaw | 4,162,415,456 | I_kwDOQb6kR874GWtg | 56,780 | https://github.com/openclaw/openclaw/issues/56780 | https://api.github.com/repos/openclaw/openclaw/issues/56780 | v2026.3.28: memory-lancedb fails to load (missing dist/package.json + @lancedb/lancedb) and recall broken with embedding context length error | ## Environment
- OpenClaw version: 2026.3.28 (upgraded from 2026.3.24)
- Platform: macOS 26.4 arm64 (Apple Silicon Mac mini)
- Embedding model: mxbai-embed-large via Ollama
## Issues
### 1. memory-lancedb fails to load on startup
After upgrading to 2026.3.28, `memory-lancedb` fails to load with:
```
Error: ENOENT: ... | closed | completed | true | 2 | [] | [] | 2026-03-29T03:58:54Z | 2026-04-30T04:32:47Z | 2026-04-28T00:05:23Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | rungmc357 | 210,088,133 | U_kgDODIWwxQ | User | false |
openclaw/openclaw | 4,336,796,824 | I_kwDOQb6kR88AAAABAn5EmA | 72,867 | https://github.com/openclaw/openclaw/issues/72867 | https://api.github.com/repos/openclaw/openclaw/issues/72867 | Telegram: Agent replies fail in specific session despite healthy connection | **Title:** Telegram: Agent replies fail in specific session despite healthy connection
**Description:**
Agent replies are failing to deliver to Telegram, even though the underlying connection is healthy. Direct messages sent via the `message` tool are delivered successfully, but standard conversational replies from th... | open | null | false | 4 | [
"bug",
"channel: telegram"
] | [] | 2026-04-27T15:11:23Z | 2026-05-03T08:25:52Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | jchilders98 | 258,925,708 | U_kgDOD27kjA | User | false |
openclaw/openclaw | 4,339,670,365 | I_kwDOQb6kR88AAAABAqodXQ | 73,106 | https://github.com/openclaw/openclaw/issues/73106 | https://api.github.com/repos/openclaw/openclaw/issues/73106 | doctor --fix silently archives orphan transcripts, causing session history loss after upgrade | ## Summary
Running `openclaw doctor --fix` after an upgrade can silently archive (rename to `.deleted.*`) transcript files that appear as orphans due to `sessions.json` being regenerated during the upgrade. This causes users to lose recent session history across all channels without any warning or confirmation.
## St... | closed | completed | true | 1 | [] | [] | 2026-04-28T00:22:05Z | 2026-05-01T04:44:36Z | 2026-04-28T08:13:56Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | scottgl9 | 10,623,288 | MDQ6VXNlcjEwNjIzMjg4 | User | false |
openclaw/openclaw | 4,339,685,909 | I_kwDOQb6kR88AAAABAqpaFQ | 73,109 | https://github.com/openclaw/openclaw/issues/73109 | https://api.github.com/repos/openclaw/openclaw/issues/73109 | talk.config still throws on SecretRef apiKey when configured under messages.tts.providers.<id>.apiKey (gap left by #72496) | ## Summary
#72496 fixed `talk.config` RPC throwing `unresolved SecretRef` for `talk.providers.<id>.apiKey`. The same redaction-throws bug still triggers when the SecretRef is on the parallel `messages.tts.providers.<id>.apiKey` site, which `talk.config` also walks when it builds the resolved provider config.
End-to-e... | closed | completed | true | 0 | [
"maintainer"
] | [
"omarshahine"
] | 2026-04-28T00:27:18Z | 2026-04-30T04:31:58Z | 2026-04-28T04:06:30Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | omarshahine | 10,343,873 | MDQ6VXNlcjEwMzQzODcz | User | false |
openclaw/openclaw | 4,337,538,732 | I_kwDOQb6kR88AAAABAomWrA | 72,945 | https://github.com/openclaw/openclaw/issues/72945 | https://api.github.com/repos/openclaw/openclaw/issues/72945 | Gateway crashes 39ms after 'ready' on os.networkInterfaces() failure in restricted sandboxes (NemoClaw, Docker-in-Docker, etc.) | ## Summary
> **In plain English:** start the gateway inside any restricted sandbox that doesn't allow `os.networkInterfaces()` (NemoClaw is one) and the gateway crashes 39ms after `ready` with an unhandled rejection from the `@homebridge/ciao` mDNS sidecar. The crash is silent in the sense that the gateway DID just lo... | closed | completed | true | 2 | [] | [
"vincentkoc"
] | 2026-04-27T17:10:29Z | 2026-04-30T04:32:06Z | 2026-04-28T00:30:45Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | mayank6136 | 16,468,779 | MDQ6VXNlcjE2NDY4Nzc5 | User | false |
openclaw/openclaw | 4,337,322,719 | I_kwDOQb6kR88AAAABAoZK3w | 72,922 | https://github.com/openclaw/openclaw/issues/72922 | https://api.github.com/repos/openclaw/openclaw/issues/72922 | [Bug]: Sluggish response time and unstable Web GUI and CLI on Windows Server 2022 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Summary
OpenClaw appears to be unstable and very slow when installed natively on Windows Server 2022. The Gateway service does not consistently stay running, and the Web GUI either loads very slowly or fails to work.
##... | open | null | false | 4 | [
"bug"
] | [] | 2026-04-27T16:35:22Z | 2026-05-03T08:39:43Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | jondeaf | 194,342,313 | U_kgDOC5VtqQ | User | false |
openclaw/openclaw | 4,339,706,329 | I_kwDOQb6kR88AAAABAqqp2Q | 73,114 | https://github.com/openclaw/openclaw/issues/73114 | https://api.github.com/repos/openclaw/openclaw/issues/73114 | Memory bloat from unbounded session/task record accumulation | ## Problem
Gateway memory usage grows to 2.5GB+ within hours due to **unbounded accumulation of session and task records**.
## Detailed Analysis
### 1. `sessions.json` — 471 entries, 26.8MB on disk, loaded entirely at startup
Every session ever created is stored and reloaded:
| Session Type | Count | Size |
|---|--... | closed | not_planned | true | 1 | [] | [] | 2026-04-28T00:34:15Z | 2026-05-04T04:36:56Z | 2026-05-01T09:57:29Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | ksxh0524 | 40,232,908 | MDQ6VXNlcjQwMjMyOTA4 | User | false |
openclaw/openclaw | 4,339,725,017 | I_kwDOQb6kR88AAAABAqry2Q | 73,117 | https://github.com/openclaw/openclaw/issues/73117 | https://api.github.com/repos/openclaw/openclaw/issues/73117 | [Bug] exec tool emits empty text content block when command produces no stdout | ## Summary
When an exec command produces no stdout (e.g. `grep` with no matches, `mkdir`, `cp`, silent `git` commands), the exec runtime constructs a `tool_result` content block with `text: ""`. Anthropic's API rejects empty text content blocks, causing the session to error out — typically after bulk operations that i... | closed | completed | true | 0 | [] | [] | 2026-04-28T00:39:49Z | 2026-04-30T04:32:13Z | 2026-04-28T03:20:51Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | agentbuilder1999 | 266,951,717 | U_kgDOD-lcJQ | User | false |
openclaw/openclaw | 4,339,749,718 | I_kwDOQb6kR88AAAABAqtTVg | 73,119 | https://github.com/openclaw/openclaw/issues/73119 | https://api.github.com/repos/openclaw/openclaw/issues/73119 | sessions_history triggers transcript-write duplication (cosmetic, not double-billed) | ## Summary
When `sessions_history` is invoked against a running session (the same session reading its own history, and likely any active session), the runtime writes a **duplicate copy of every prior message** back into the session's JSONL transcript as a new branch with fresh message IDs but identical content and ide... | closed | not_planned | true | 2 | [] | [] | 2026-04-28T00:46:44Z | 2026-05-05T04:25:46Z | 2026-05-03T02:04:12Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | jared-rebel | 147,876,109 | U_kgDOCNBpDQ | User | false |
openclaw/openclaw | 4,339,751,945 | I_kwDOQb6kR88AAAABAqtcCQ | 73,121 | https://github.com/openclaw/openclaw/issues/73121 | https://api.github.com/repos/openclaw/openclaw/issues/73121 | Zombie running tasks: CLI runtime tasks that terminate are never auto-transitioned to failed/timed_out | ## Problem
`cli` runtime tasks (subagent/exec-approval) that have had their underlying process terminated remain stuck in `running` state indefinitely. Their task records are never auto-transitioned to `failed` or `timed_out`, causing the task list to accumulate "zombie" entries.
## Reproduction
On 2026-04-28, `open... | open | null | false | 1 | [] | [] | 2026-04-28T00:47:23Z | 2026-05-03T06:50:19Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | ksxh0524 | 40,232,908 | MDQ6VXNlcjQwMjMyOTA4 | User | false |
openclaw/openclaw | 4,163,904,786 | I_kwDOQb6kR874MCUS | 56,968 | https://github.com/openclaw/openclaw/issues/56968 | https://api.github.com/repos/openclaw/openclaw/issues/56968 | Bug: status shows 'Memory enabled · unavailable' for third-party memory plugins (memory-lancedb-pro) | ## Description
When using a third-party memory plugin (e.g., `memory-lancedb-pro`), the `openclaw status` command incorrectly displays:
```
Memory │ enabled (plugin memory-lancedb-pro) · unavailable
```
The memory plugin is **fully functional** (66 memories stored, hybrid retrieval working, smart extraction enabled)... | closed | completed | true | 2 | [] | [] | 2026-03-29T11:36:32Z | 2026-04-30T04:32:42Z | 2026-04-28T00:51:51Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | Tony-ooo | 45,898,990 | MDQ6VXNlcjQ1ODk4OTkw | User | false |
openclaw/openclaw | 4,339,794,059 | I_kwDOQb6kR88AAAABAqwAiw | 73,124 | https://github.com/openclaw/openclaw/issues/73124 | https://api.github.com/repos/openclaw/openclaw/issues/73124 | [Feature Request] Add config option to disable model pricing fetch + support Chinese LLM subscription plans | ## Problem Description
When OpenClaw Gateway starts (or periodically during runtime), it attempts to fetch model pricing data from two external sources:
1. **OpenRouter** - https://openrouter.ai/api/v1/models
2. **LiteLLM** - https://raw.githubusercontent.com/BerriAI/litellm/main/model_prices_and_context_window.json
... | closed | completed | true | 1 | [] | [] | 2026-04-28T00:59:32Z | 2026-05-03T04:36:38Z | 2026-04-30T14:57:47Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | guowenjiao54 | 28,590,729 | MDQ6VXNlcjI4NTkwNzI5 | User | false |
openclaw/openclaw | 4,339,623,116 | I_kwDOQb6kR88AAAABAqlkzA | 73,099 | https://github.com/openclaw/openclaw/issues/73099 | https://api.github.com/repos/openclaw/openclaw/issues/73099 | Silent clean exit (code 0) when OPENCLAW_PLUGIN_STAGE_DIR is on NTFS bind-mounted volume — channels resolve as stopped with no error log | **Environment**
- OpenClaw: v2026.4.25
- Node: 24.14.0
- Docker: Linux container on Windows 11 host (Hyper-V)
- Plugins: whatsapp, telegram, acpx, browser, memory-core, memory-wiki, webhooks
**Symptom**
Gateway boots to ready state (all plugins loaded), then dies silently after:
[gateway/hooks] loaded 4 internal hook ... | closed | completed | true | 1 | [] | [] | 2026-04-28T00:09:06Z | 2026-04-30T04:32:41Z | 2026-04-28T00:59:56Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | balaji1968-kingler | 268,807,278 | U_kgDOEAWsbg | User | false |
openclaw/openclaw | 4,182,642,667 | I_kwDOQb6kR875Tg_r | 58,584 | https://github.com/openclaw/openclaw/issues/58584 | https://api.github.com/repos/openclaw/openclaw/issues/58584 | Agent pre-flight check: verify model provider is reachable before scheduling cron jobs | ## Problem
Agents configured with model providers that aren't running (e.g. Ollama models before Ollama is installed/started) cause constant authentication errors. With 15 agents all hitting a dead provider, this creates:
- Continuous CPU churn from failed requests
- Auth error spam in logs (contributing to log bloat)... | closed | completed | true | 1 | [] | [] | 2026-03-31T22:01:21Z | 2026-04-30T04:32:39Z | 2026-04-28T01:12:48Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | jpeghead | 7,033,691 | MDQ6VXNlcjcwMzM2OTE= | User | false |
openclaw/openclaw | 4,339,884,694 | I_kwDOQb6kR88AAAABAq1ilg | 73,137 | https://github.com/openclaw/openclaw/issues/73137 | https://api.github.com/repos/openclaw/openclaw/issues/73137 | [Bug]: Slack text reset trigger leaks into the new session prompt | ## Bug type
Behavior bug (incorrect output/state without crash)
## Beta release blocker
No
## Summary
Slack text reset triggers such as `new session` can start a fresh OpenClaw session but still be restored into the model prompt as ordinary user input during the same turn.
## Steps to reproduce
1. Configure Open... | closed | completed | true | 1 | [] | [] | 2026-04-28T01:20:32Z | 2026-05-01T04:44:05Z | 2026-04-28T09:24:44Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | neeravmakwana | 261,249,544 | U_kgDOD5JaCA | User | false |
openclaw/openclaw | 4,339,843,720 | I_kwDOQb6kR88AAAABAqzCiA | 73,131 | https://github.com/openclaw/openclaw/issues/73131 | https://api.github.com/repos/openclaw/openclaw/issues/73131 | [Bug] One incompatible plugin in plugins.entries blocks entire config reload — should skip+warn instead | ## Description
When a single plugin in \`plugins.entries\` requires a newer OpenClaw version than the running host, the entire config reload is rejected with \"Invalid config\" — instead of skipping just that plugin and loading everything else.
## Evidence from gateway.err.log
```
[reload] config reload skipped (inv... | closed | completed | true | 1 | [] | [] | 2026-04-28T01:10:25Z | 2026-04-30T04:32:34Z | 2026-04-28T01:25:47Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | Adam-Researchh | 261,827,313 | U_kgDOD5sq8Q | User | false |
openclaw/openclaw | 4,339,843,755 | I_kwDOQb6kR88AAAABAqzCqw | 73,132 | https://github.com/openclaw/openclaw/issues/73132 | https://api.github.com/repos/openclaw/openclaw/issues/73132 | [Bug] openclaw gateway stop / launchctl bootout doesn't terminate underlying node process — wedged binary persists across stop | ## Description
\`openclaw gateway stop\` returns successfully (with the message \"Stopped LaunchAgent (degraded)\") but the underlying \`node\` process hosting the gateway continues to run. This wedged process keeps the gateway port bound and prevents a clean restart.
## Evidence
After running \`openclaw gateway sto... | open | null | false | 2 | [] | [] | 2026-04-28T01:10:26Z | 2026-05-03T07:05:23Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Adam-Researchh | 261,827,313 | U_kgDOD5sq8Q | User | false |
openclaw/openclaw | 4,337,538,902 | I_kwDOQb6kR88AAAABAomXVg | 72,947 | https://github.com/openclaw/openclaw/issues/72947 | https://api.github.com/repos/openclaw/openclaw/issues/72947 | Plugin loader silently ignores async register() return value — race conditions on first-traffic | ## Summary
> **In plain English:** if a plugin's `register()` is `async`, the gateway's plugin loader logs "plugin register returned a promise; async registration is ignored" — and then proceeds as if registration is complete, ignoring whatever async work the plugin is still doing. This creates a small window where ho... | closed | completed | true | 1 | [] | [] | 2026-04-27T17:10:31Z | 2026-04-30T04:30:20Z | 2026-04-28T01:29:43Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | mayank6136 | 16,468,779 | MDQ6VXNlcjE2NDY4Nzc5 | User | false |
openclaw/openclaw | 4,337,547,336 | I_kwDOQb6kR88AAAABAom4SA | 72,950 | https://github.com/openclaw/openclaw/issues/72950 | https://api.github.com/repos/openclaw/openclaw/issues/72950 | Plugin config (plugins.entries.<id>.config) has no env-var or writable-overlay path — every change requires image rebuild in policy-locked sandboxes | ## Summary
> **In plain English:** plugin config (`plugins.entries.<id>.config`) is sourced from `openclaw.json` at startup — and that file is in a directory that NemoClaw makes read-only by Landlock policy at runtime. So a plugin like `<observability-plugin>` that needs per-deployment config (OTLP endpoint, service n... | open | null | false | 4 | [] | [] | 2026-04-27T17:11:58Z | 2026-05-03T01:38:09Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | mayank6136 | 16,468,779 | MDQ6VXNlcjE2NDY4Nzc5 | User | false |
openclaw/openclaw | 4,339,936,517 | I_kwDOQb6kR88AAAABAq4tBQ | 73,143 | https://github.com/openclaw/openclaw/issues/73143 | https://api.github.com/repos/openclaw/openclaw/issues/73143 | openclaw-codex-app-server plugin fails to load on 2026.4.25 - missing root-alias.cjs/discord module | ## Issue
openclaw-codex-app-server plugin (v0.5.0) fails to load on OpenClaw 2026.4.25 with a missing module error.
## Error
```
[plugins] openclaw-codex-app-server failed to load from /Users/mrod/.openclaw/extensions/openclaw-codex-app-server/index.ts: Error: Cannot find module '/usr/local/lib/node_modules/openclaw/d... | closed | completed | true | 1 | [
"r: third-party-extension"
] | [] | 2026-04-28T01:33:31Z | 2026-05-01T04:45:18Z | 2026-04-28T05:29:17Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Axiomai007 | 266,307,263 | U_kgDOD9-Gvw | User | false |
openclaw/openclaw | 4,339,707,241 | I_kwDOQb6kR88AAAABAqqtaQ | 73,115 | https://github.com/openclaw/openclaw/issues/73115 | https://api.github.com/repos/openclaw/openclaw/issues/73115 | 2026.4.25: Telegram polling broken + 30s-per-candidate model fallback (regression vs 2026.4.23) | # Bug Report: openclaw 2026.4.25 — Telegram polling broken + model fallback unusable
**Repo:** https://github.com/openclaw/openclaw
**Affected versions:** `2026.4.24`, `2026.4.25` (both regressions appear after `2026.4.23`)
**Last known good:** `2026.4.23`
**Workaround:** `npm i -g openclaw@2026.4.23`
**Platform:** Wi... | closed | completed | true | 3 | [] | [] | 2026-04-28T00:34:36Z | 2026-04-30T04:32:33Z | 2026-04-28T01:33:59Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | vdruts | 20,264,016 | MDQ6VXNlcjIwMjY0MDE2 | User | false |
openclaw/openclaw | 4,339,940,349 | I_kwDOQb6kR88AAAABAq47_Q | 73,144 | https://github.com/openclaw/openclaw/issues/73144 | https://api.github.com/repos/openclaw/openclaw/issues/73144 | Model switch experience: 5 issues when switching from qwen3.6-plus to deepseek-v4-pro | ## Summary
I recently switched my default model from `bailian/qwen3.6-plus` to `deepseek/deepseek-v4-pro`. The process exposed several UX issues in config management. Documenting here for visibility.
### Issue 1: `config.patch` cannot add a new provider
```
Error: gateway config.patch cannot change protected config ... | open | null | false | 1 | [] | [] | 2026-04-28T01:34:24Z | 2026-05-03T09:49:14Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | shaolin-cloud | 268,214,353 | U_kgDOD_ygUQ | User | false |
openclaw/openclaw | 4,339,941,870 | I_kwDOQb6kR88AAAABAq5B7g | 73,145 | https://github.com/openclaw/openclaw/issues/73145 | https://api.github.com/repos/openclaw/openclaw/issues/73145 | [Bug]: Feishu TTS voice replies not auto-transcoded to Ogg/Opus — delivery fails with MP3-only output (v2026.4.25) | ## Description
Feishu TTS voice replies are **not being auto-transcoded to Ogg/Opus** despite the feishu plugin declaring `transcodesAudio: true` in v2026.4.25. The TTS synthesis produces MP3 files successfully, but the transcoding step to Ogg/Opus (required for native Feishu voice messages) either does not execute or... | closed | completed | true | 1 | [] | [] | 2026-04-28T01:34:48Z | 2026-05-03T04:36:38Z | 2026-04-30T14:57:55Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Bowerai | 17,424,365 | MDQ6VXNlcjE3NDI0MzY1 | User | false |
openclaw/openclaw | 4,339,840,741 | I_kwDOQb6kR88AAAABAqy25Q | 73,128 | https://github.com/openclaw/openclaw/issues/73128 | https://api.github.com/repos/openclaw/openclaw/issues/73128 | [Bug] 2026.4.25 packaging: browser plugin module missing from published tarball — control-auth-DMfkKbls.js | ## Description
OpenClaw 2026.4.25 ships with a broken browser plugin: `dist/control-auth-DMfkKbls.js` is referenced by `dist/server-context-COugkIR3.js` in the plugin-runtime-deps tree but is not actually present in the published npm package.
## Steps to Reproduce
1. Install 2026.4.25: `npm i -g openclaw@2026.4.25 -... | closed | completed | true | 1 | [] | [] | 2026-04-28T01:09:44Z | 2026-04-30T04:32:32Z | 2026-04-28T01:37:16Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | Adam-Researchh | 261,827,313 | U_kgDOD5sq8Q | User | false |
openclaw/openclaw | 4,339,855,498 | I_kwDOQb6kR88AAAABAqzwig | 73,133 | https://github.com/openclaw/openclaw/issues/73133 | https://api.github.com/repos/openclaw/openclaw/issues/73133 | fix(discord): fetchChannel called before reactionMode check causes 30-73s slow listeners on gateway reconnect | ## Summary
In \`handleDiscordReactionEvent\`, \`client.fetchChannel(data.channel_id)\` is called unconditionally at line 475 — before any \`reactionMode\` check. When the Discord gateway disconnects and reconnects, queued \`MESSAGE_REACTION_ADD\` events fire simultaneously, all calling \`fetchChannel\` at once. This h... | closed | completed | true | 1 | [] | [] | 2026-04-28T01:13:02Z | 2026-04-30T04:32:31Z | 2026-04-28T01:37:18Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | isaacsummers | 13,228,201 | MDQ6VXNlcjEzMjI4MjAx | User | false |
openclaw/openclaw | 4,339,951,821 | I_kwDOQb6kR88AAAABAq5ozQ | 73,146 | https://github.com/openclaw/openclaw/issues/73146 | https://api.github.com/repos/openclaw/openclaw/issues/73146 | Cron `payload.model` route can be rejected before provider/proxy is called, causing fallback to agent default | ## Summary
Cron jobs configured with a logical provider/model such as `litellm/cron` were rejected before OpenClaw called the configured provider/proxy. The job then fell back to the agent default model instead of preserving the provider/proxy routing chain.
In our setup, `litellm/cron` was supposed to route cron work... | closed | completed | true | 1 | [] | [] | 2026-04-28T01:37:45Z | 2026-04-30T04:32:14Z | 2026-04-28T03:17:42Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | oneandrewwang | 261,546,085 | U_kgDOD5bgZQ | User | false |
openclaw/openclaw | 4,339,956,239 | I_kwDOQb6kR88AAAABAq56Dw | 73,148 | https://github.com/openclaw/openclaw/issues/73148 | https://api.github.com/repos/openclaw/openclaw/issues/73148 | Image tool: opaque "Failed to optimize image" when sharp is not installed | ## Summary
The `image` tool fails with the opaque error `Failed to optimize image` on **every** image in environments where the optional `sharp` package is not installed. The vision pipeline has no fallback and no clear indication that a missing native dependency is the cause.
## Environment
- **OpenClaw:** 2026.4.2... | open | null | false | 10 | [] | [] | 2026-04-28T01:38:50Z | 2026-05-05T04:02:11Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | azgardtek | 219,332,981 | U_kgDODRLBdQ | User | false |
openclaw/openclaw | 4,339,959,265 | I_kwDOQb6kR88AAAABAq6F4Q | 73,149 | https://github.com/openclaw/openclaw/issues/73149 | https://api.github.com/repos/openclaw/openclaw/issues/73149 | Cron/heartbeat no-op runs can leak visible messages despite silence/no-op semantics | ## Summary
We saw repeated visible messages from background/cron/heartbeat workflows that were supposed to remain silent on no-op results. Examples include `HEARTBEAT_OK`, watchdog summaries, and earlier voice-transcription `NO_NEW_AUDIO` handling. In some cases the same no-op/silent response appeared duplicated around... | open | null | false | 2 | [] | [] | 2026-04-28T01:39:42Z | 2026-05-03T09:49:15Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | oneandrewwang | 261,546,085 | U_kgDOD5bgZQ | User | false |
openclaw/openclaw | 4,339,965,112 | I_kwDOQb6kR88AAAABAq6cuA | 73,152 | https://github.com/openclaw/openclaw/issues/73152 | https://api.github.com/repos/openclaw/openclaw/issues/73152 | Docs/doctor request: clarify gateway reachability for OrbStack/WSL/VM/Tailscale setups | ## Summary
During a Mac mini OrbStack + Windows/WSL2 era, it was difficult to distinguish OpenClaw gateway bind behavior from VM/NAT/Tailscale reachability constraints. We used reverse SSH/Tailscale workarounds and at times misdiagnosed environment networking as an OpenClaw LAN binding problem.
This is probably not a ... | open | null | false | 1 | [] | [] | 2026-04-28T01:41:20Z | 2026-05-03T06:50:21Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | oneandrewwang | 261,546,085 | U_kgDOD5bgZQ | User | false |
openclaw/openclaw | 4,339,964,556 | I_kwDOQb6kR88AAAABAq6ajA | 73,151 | https://github.com/openclaw/openclaw/issues/73151 | https://api.github.com/repos/openclaw/openclaw/issues/73151 | [Bug]: pnpm build does not write runtime-postbuild stamp; first CLI invocation re-syncs | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
After pnpm build completes successfully, the next pnpm openclaw <anything> invocation prints [openclaw] Syncing runtime artifacts (missing_runtime_postbuild_stamp - runtime postbuild stamp missing). and re-runs... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-28T01:41:08Z | 2026-05-01T04:44:41Z | 2026-04-28T07:33:55Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | bittoby | 218,712,309 | U_kgDODQlI9Q | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.