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 | 4,056,740,716 | I_kwDOQb6kR87xzPNs | 42,982 | https://github.com/openclaw/openclaw/issues/42982 | https://api.github.com/repos/openclaw/openclaw/issues/42982 | [Bug]: redact logging not working properly | ### Bug type
### Summary
redact logging not working with regex : user **sensitive** info will be **exposed** to admin or who can access the session log.
### Steps to reproduce
1. config to redact email address
```
"logging": {
"redactSensitive": "tools",
"redactPatterns": [
"([\\w]|[-.])+@([\\w]|[... | closed | completed | true | 2 | [
"bug",
"regression"
] | [
"vincentkoc"
] | 2026-03-11T08:36:39Z | 2026-04-29T04:34:14Z | 2026-04-26T19:14:53Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | panpan0000 | 14,049,268 | MDQ6VXNlcjE0MDQ5MjY4 | User | false |
openclaw/openclaw | 4,331,750,102 | I_kwDOQb6kR88AAAABAjFC1g | 72,338 | https://github.com/openclaw/openclaw/issues/72338 | https://api.github.com/repos/openclaw/openclaw/issues/72338 | Gateway CPU spin causes Telegram replies to stall and status probe to time out | ### Summary
The OpenClaw gateway can enter a high-CPU state where Telegram polling still appears to ingest updates, but the agent reply pipeline stalls and `openclaw channels status --deep` times out against the local loopback gateway. Restarting `openclaw-gateway.service` temporarily restores Telegram replies, but th... | closed | completed | true | 25 | [] | [] | 2026-04-26T19:16:13Z | 2026-05-04T04:33:38Z | 2026-05-01T11:49:11Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | midhunmonachan | 70,493,664 | MDQ6VXNlcjcwNDkzNjY0 | User | false |
openclaw/openclaw | 4,331,760,846 | I_kwDOQb6kR88AAAABAjFszg | 72,340 | https://github.com/openclaw/openclaw/issues/72340 | https://api.github.com/repos/openclaw/openclaw/issues/72340 | Slack: native streaming gate excludes DMs even though chat.startStream accepts recipient_user_id | ## Summary
When `channels.slack.streaming = { mode: "partial", nativeTransport: true }` (default), DMs to the bot never use Slack's native streaming API (`chat.startStream` / `chat.appendStream` / `chat.stopStream`). They always hit the non-streaming fallback path, which posts each text-between-tools assistant block a... | closed | not_planned | true | 1 | [] | [] | 2026-04-26T19:21:30Z | 2026-04-29T04:33:48Z | 2026-04-26T21:35:22Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jason-alvarez-git | 188,690,207 | U_kgDOCz8vHw | User | false |
openclaw/openclaw | 4,331,615,103 | I_kwDOQb6kR88AAAABAi8zfw | 72,315 | https://github.com/openclaw/openclaw/issues/72315 | https://api.github.com/repos/openclaw/openclaw/issues/72315 | [Bug]: npm grammyjs npm anthropic frequent disturbance shell popup in windows every 15 seconds | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
npm grammyjs npm anthropic frequent disturbance shell popup in windows every 15 seconds
### Steps to reproduce
OpenClaw 2026.4.24 (cbcfdf6)
configure telegram
run
### Expected behavior
no issue
### Actual behavior
... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-04-26T18:17:55Z | 2026-04-30T04:36:02Z | 2026-04-27T07:31:19Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | athuljayaram | 141,337,513 | U_kgDOCGyjqQ | User | false |
openclaw/openclaw | 4,331,761,865 | I_kwDOQb6kR88AAAABAjFwyQ | 72,341 | https://github.com/openclaw/openclaw/issues/72341 | https://api.github.com/repos/openclaw/openclaw/issues/72341 | Control UI / WebChat: assistant text-between-tools blocks render as cumulative duplicates instead of fresh cards | ## Summary
In Control UI WebChat, when an agent loop produces multiple `text → tool_call → tool_result → text → tool_call → ...` cycles, each new assistant text block is rendered as a *separate* visible message card containing the cumulative text from all prior assistant text blocks in the same turn. The gateway-side ... | open | null | false | 4 | [] | [] | 2026-04-26T19:22:01Z | 2026-05-03T13:21:35Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | jason-alvarez-git | 188,690,207 | U_kgDOCz8vHw | User | false |
openclaw/openclaw | 4,331,810,268 | I_kwDOQb6kR88AAAABAjIt3A | 72,344 | https://github.com/openclaw/openclaw/issues/72344 | https://api.github.com/repos/openclaw/openclaw/issues/72344 | Bonjour/ciao mDNS causes repeated gateway crashes on headless Linux (CIAO PROBING CANCELLED) | ### Bug Report: Bonjour/ciao mDNS plugin causes repeated gateway crashes on headless Linux VPS ("CIAO PROBING CANCELLED")
**Version:** 2026.4.24 (and previous weeks)
**Environment:**
- Ubuntu 24.04 VPS (Contabo)
- Headless
- `gateway.mode: local`, `bind: loopback`
- Both Avahi daemon (`avahi-daemon`) and OpenClaw's b... | closed | completed | true | 1 | [] | [] | 2026-04-26T19:44:52Z | 2026-04-29T04:33:47Z | 2026-04-26T21:35:27Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | oc20260214 | 264,373,579 | U_kgDOD8IFSw | User | false |
openclaw/openclaw | 4,331,687,323 | I_kwDOQb6kR88AAAABAjBNmw | 72,323 | https://github.com/openclaw/openclaw/issues/72323 | https://api.github.com/repos/openclaw/openclaw/issues/72323 | [Bug]: The gateway keeps restarting endlessly | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
The gateway keeps restarting endlessly
### Steps to reproduce
1. Use `openclaw update`, version 4.23 -> 4.24
2. Exec `systemctl --user restart openclaw-gateway`
### Expected behavior
NOT_ENOUGH_INFO
### Actual behavior
<i... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-04-26T18:50:57Z | 2026-04-29T04:31:04Z | 2026-04-26T21:34:52Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | soladxy | 51,918,729 | MDQ6VXNlcjUxOTE4NzI5 | User | false |
openclaw/openclaw | 4,041,790,665 | I_kwDOQb6kR87w6NTJ | 40,125 | https://github.com/openclaw/openclaw/issues/40125 | https://api.github.com/repos/openclaw/openclaw/issues/40125 | [Feature]: Proactively warn before model_context_window_exceeded and provide /reset or /new guidance | ### Summary
This change improves OpenClaw’s behavior when long-running sessions approach the active model’s context limit.
### Problem to solve
Problem
Persistent chat channels like Telegram DMs can accumulate enough history and tool output to exceed the model context window. Today, users usually encounter this on... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-08T17:37:12Z | 2026-05-04T09:32:58Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | sruckh | 29,985,975 | MDQ6VXNlcjI5OTg1OTc1 | User | false |
openclaw/openclaw | 4,331,824,729 | I_kwDOQb6kR88AAAABAjJmWQ | 72,345 | https://github.com/openclaw/openclaw/issues/72345 | https://api.github.com/repos/openclaw/openclaw/issues/72345 | voice-call: CLI `voicecall call` cannot run when gateway is owning the webhook port | ## Bug / Missing Feature
When the OpenClaw gateway daemon is running and the voice-call plugin is enabled, the gateway opens its own webhook listener (default `127.0.0.1:3334`).
If you then run `openclaw voicecall call ...` from the CLI, it fails with:
```
[openclaw] Failed to start CLI: Error: listen EADDRINUSE: ad... | closed | completed | true | 4 | [] | [] | 2026-04-26T19:51:35Z | 2026-05-04T04:37:13Z | 2026-05-01T08:12:33Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | serrurco | 264,545,475 | U_kgDOD8Skww | User | false |
openclaw/openclaw | 4,331,832,652 | I_kwDOQb6kR88AAAABAjKFTA | 72,346 | https://github.com/openclaw/openclaw/issues/72346 | https://api.github.com/repos/openclaw/openclaw/issues/72346 | [Bug]: bonjour plugin: unhandled rejection on probe cancellation crashes process in 2026.4.24 (worked in 2026.4.23) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## TL;DR
Upgrading from `2026.4.23-arm64` to `2026.4.24-arm64` introduced an unhandled
promise rejection from the bonjour plugin's watchdog-triggered advertiser
restart. The rejection takes the openclaw process down. On env... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-26T19:55:21Z | 2026-04-29T04:33:47Z | 2026-04-26T21:35:31Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | gshockbu | 12,898,775 | MDQ6VXNlcjEyODk4Nzc1 | User | false |
openclaw/openclaw | 4,331,857,067 | I_kwDOQb6kR88AAAABAjLkqw | 72,349 | https://github.com/openclaw/openclaw/issues/72349 | https://api.github.com/repos/openclaw/openclaw/issues/72349 | [Bug]: Subagent spawn fails with "No API key found for bedrock" when using IAM Roles Anywhere (AWS_PROFILE + credential_process) | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
`sessions_spawn` subagent runs to Bedrock-backed agents fail with `FailoverError: No API key found for bedrock` even though the same agent, same model, and same AWS credentials work perfectly from the interactive (main session)... | closed | completed | true | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-26T20:06:32Z | 2026-04-29T04:33:45Z | 2026-04-26T21:36:55Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | sudheermanubolu | 44,275,113 | MDQ6VXNlcjQ0Mjc1MTEz | User | false |
openclaw/openclaw | 4,043,136,342 | I_kwDOQb6kR87w_V1W | 40,536 | https://github.com/openclaw/openclaw/issues/40536 | https://api.github.com/repos/openclaw/openclaw/issues/40536 | [Feature]: Per-Skill Model Override | ### Summary
Add model override support for individual skills
### Problem to solve
Currently, OpenClaw allows configuring models at:
-Global default `agents.defaults.model`
-Per-agent `agents.list[].model`
-Per-channel `channels.modelByChannel`
But there's no way to specify a model for a particular skill. Some ski... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-09T03:37:21Z | 2026-05-01T04:41:03Z | 2026-04-29T00:08:18Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | ShuyunJia | 25,626,486 | MDQ6VXNlcjI1NjI2NDg2 | User | false |
openclaw/openclaw | 4,043,757,195 | I_kwDOQb6kR87xBtaL | 40,712 | https://github.com/openclaw/openclaw/issues/40712 | https://api.github.com/repos/openclaw/openclaw/issues/40712 | [Feature]: Resume interrupted message generation after gateway restart | ### Summary
## Problem
When the gateway restarts (via `openclaw gateway restart`, updates, or crashes) during active message generation, the in-progress LLM response is permanently lost. The user receives an incomplete message and must manually ask the agent to continue, losing context about what was being generated.... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-09T07:02:17Z | 2026-05-04T10:51:41Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | wufei-png | 63,766,429 | MDQ6VXNlcjYzNzY2NDI5 | User | false |
openclaw/openclaw | 4,331,853,146 | I_kwDOQb6kR88AAAABAjLVWg | 72,347 | https://github.com/openclaw/openclaw/issues/72347 | https://api.github.com/repos/openclaw/openclaw/issues/72347 | Active Memory: timeoutMs unenforceable due to multi-turn agent loop; need single-shot mode | ## Active Memory: structural latency from multi-turn agent loop makes timeoutMs effectively unenforceable
### Summary
Active Memory is documented and configured as a "fast preflight" with a `timeoutMs` budget (defaulting to 3000ms, our config 5000ms), but in practice it consistently runs 12–22s regardless of which mo... | closed | completed | true | 2 | [] | [] | 2026-04-26T20:04:50Z | 2026-04-29T04:33:46Z | 2026-04-26T21:36:52Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | thecolormaroun | 60,901,592 | MDQ6VXNlcjYwOTAxNTky | User | false |
openclaw/openclaw | 4,331,758,414 | I_kwDOQb6kR88AAAABAjFjTg | 72,339 | https://github.com/openclaw/openclaw/issues/72339 | https://api.github.com/repos/openclaw/openclaw/issues/72339 | [Bug]: Gateway crashes after update to 2026.4.24 (CIAO PROBING CANCELLED) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
📋 Description
After updating OpenClaw from 2026.4.23 to 2026.4.24, the gateway no longer stays running. It appears to start successfully, but then crashes due to a Bonjour/ciao probing issue.
### Steps to reproduce
🔁 St... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-26T19:20:12Z | 2026-04-29T04:33:49Z | 2026-04-26T21:35:17Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | DanThe3r | 66,535,701 | MDQ6VXNlcjY2NTM1NzAx | User | false |
openclaw/openclaw | 4,331,946,087 | I_kwDOQb6kR88AAAABAjRAZw | 72,353 | https://github.com/openclaw/openclaw/issues/72353 | https://api.github.com/repos/openclaw/openclaw/issues/72353 | Gateway prepends custom-provider prefix to /api/chat — follow-up to #67435 | ## Summary
Follow-up to #67435 (closed by #67457) — the fix in #67457 strips the literal `"ollama/"` prefix from model IDs in `buildOllamaChatRequest`, but the same bug class survives for **custom-named Ollama-compatible providers** (e.g. `ollama-spark`, `ollama-studio2`, `ollama-remote`). The custom provider name is ... | closed | completed | true | 2 | [] | [] | 2026-04-26T20:49:47Z | 2026-04-29T04:33:19Z | 2026-04-27T00:03:17Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | maximus-dss | 263,210,249 | U_kgDOD7BFCQ | User | false |
openclaw/openclaw | 4,331,958,353 | I_kwDOQb6kR88AAAABAjRwUQ | 72,355 | https://github.com/openclaw/openclaw/issues/72355 | https://api.github.com/repos/openclaw/openclaw/issues/72355 | [Bug]: Bonjour/mDNS plugin hardcoded to 'openclaw.local' causes crash loop | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
The Bonjour plugin hardcodes the mDNS hostname to openclaw.local instead of using the system hostname, causing CIAO library "PROBING CANCELED" errors and gateway crash loops.
### Steps to reproduce
1. Install OpenClaw 2026.4.... | closed | completed | true | 6 | [
"bug",
"bug:crash"
] | [] | 2026-04-26T20:54:55Z | 2026-04-30T04:34:53Z | 2026-04-27T10:35:33Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | mscheuerlein-bot | 229,647,208 | U_kgDODbAjaA | User | false |
openclaw/openclaw | 4,331,961,165 | I_kwDOQb6kR88AAAABAjR7TQ | 72,356 | https://github.com/openclaw/openclaw/issues/72356 | https://api.github.com/repos/openclaw/openclaw/issues/72356 | [Bug]: Double Messages | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Bug: Double Messages in WebChat (2026.4.24 Regression)
After updating from 2026.4.11 → 2026.4.24, both user messages AND agent responses appear as duplicates in webchat. Every message is sent 2-3x. Discord/Telegram work fin... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-26T20:56:21Z | 2026-04-29T04:33:43Z | 2026-04-26T21:37:01Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | topshlf | 262,241,403 | U_kgDOD6F8ew | User | false |
openclaw/openclaw | 4,331,947,771 | I_kwDOQb6kR88AAAABAjRG-w | 72,354 | https://github.com/openclaw/openclaw/issues/72354 | https://api.github.com/repos/openclaw/openclaw/issues/72354 | Subagent result exists but requester reports "aborted/no result" due to completion announce timeout (Telegram) | ## Summary
A subagent run finishes successfully (its session is marked done and contains an assistant reply), but the requesting agent/UI reports the run as "aborted/without result".
## Version
- OpenClaw tag: `v2026.4.25-beta.9`
- Channel: Telegram
## Expected
Requester reliably receives the subagent completion resu... | closed | completed | true | 2 | [] | [] | 2026-04-26T20:50:24Z | 2026-04-29T04:33:44Z | 2026-04-26T21:36:58Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | BomBastikDE | 48,631,963 | MDQ6VXNlcjQ4NjMxOTYz | User | false |
openclaw/openclaw | 4,331,967,579 | I_kwDOQb6kR88AAAABAjSUWw | 72,357 | https://github.com/openclaw/openclaw/issues/72357 | https://api.github.com/repos/openclaw/openclaw/issues/72357 | [Bug]: Bundled LSP tsserver processes can remain after Gateway stop | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Bundled LSP cleanup could leave nested `tsserver` child processes running after the OpenClaw Gateway stopped.
### Steps to reproduce
1. Run OpenClaw with a bundled plugin that declares a stdio LSP server such... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-26T20:59:40Z | 2026-04-30T04:33:37Z | 2026-04-27T18:10:58Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | bittoby | 218,712,309 | U_kgDODQlI9Q | User | false |
openclaw/openclaw | 4,331,978,307 | I_kwDOQb6kR88AAAABAjS-Qw | 72,359 | https://github.com/openclaw/openclaw/issues/72359 | https://api.github.com/repos/openclaw/openclaw/issues/72359 | Active Memory: add single-shot mode (no embedded agent loop) for low-latency preflight injection | ## Summary
Request a new `plugins.entries.active-memory.config.mode: "single-shot"` (or similar) that bypasses `runEmbeddedPiAgent` and produces the memory summary in **one** LLM round-trip + one qmd query, instead of an N-turn agent loop.
This is the second of the two fixes proposed in #72347. That issue addressed t... | open | null | false | 1 | [] | [] | 2026-04-26T21:04:43Z | 2026-05-03T13:21:36Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | thecolormaroun | 60,901,592 | MDQ6VXNlcjYwOTAxNTky | User | false |
openclaw/openclaw | 4,331,983,244 | I_kwDOQb6kR88AAAABAjTRjA | 72,360 | https://github.com/openclaw/openclaw/issues/72360 | https://api.github.com/repos/openclaw/openclaw/issues/72360 | [Bug]: /allowlist --store bypasses channel configWrites policy | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 7.6 / 10.0 | 7.2 / 10.0 |
| **Severity** | High | High |
| **Vector** | CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:U/C:L/I:H/A:L | CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:L/VI:H/VA:L/SC:N/SI:N/SA:N |
| **Calculator** | [CVSS v... | open | null | false | 1 | [] | [] | 2026-04-26T21:07:02Z | 2026-05-03T09:13:01Z | null | CONTRIBUTOR | null | 20260503T120038Z | 2026-05-03T12:00:38Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 4,331,988,059 | I_kwDOQb6kR88AAAABAjTkWw | 72,363 | https://github.com/openclaw/openclaw/issues/72363 | https://api.github.com/repos/openclaw/openclaw/issues/72363 | Reasoning/thinking tokens and tool call intermediates leak into Telegram chat (GLM, MiniMax) | ## Bug Description
On Telegram (non-streaming mode), model reasoning/thinking tokens and tool call intermediates are being sent as visible messages before being cleaned up when the final response is ready. This occurs approximately 3 out of 5 messages with GLM and MiniMax models.
## Observed Behavior
1. **Reasoning/... | closed | completed | true | 3 | [] | [] | 2026-04-26T21:09:28Z | 2026-05-01T04:42:10Z | 2026-04-28T19:28:26Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | SweetSophia | 44,297,511 | MDQ6VXNlcjQ0Mjk3NTEx | User | false |
openclaw/openclaw | 4,331,993,898 | I_kwDOQb6kR88AAAABAjT7Kg | 72,365 | https://github.com/openclaw/openclaw/issues/72365 | https://api.github.com/repos/openclaw/openclaw/issues/72365 | [Bug]: Regression: Control UI / gateway becomes unresponsive for 30-45s during chat runs after v2026.4.14 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Hi, I’m seeing a regression after upgrading from `v2026.4.14` to the newer `v2026.4.24 beta/main` line.
Before the upgrade, the Docker gateway was usable after ~5s, chat messages appeared immediately in Control UI, and simp... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-04-26T21:12:19Z | 2026-04-30T04:34:45Z | 2026-04-27T10:51:49Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | Gabiii2398 | 279,440,179 | U_kgDOEKfrMw | User | false |
openclaw/openclaw | 4,331,986,541 | I_kwDOQb6kR88AAAABAjTebQ | 72,362 | https://github.com/openclaw/openclaw/issues/72362 | https://api.github.com/repos/openclaw/openclaw/issues/72362 | [Bug]: `fs-safe` writes silently fail in Docker with misleading "path is not a regular file under root" when `python3` is not in PATH | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
writeFileWithinRoot and openFileWithinRoot in dist/fs-safe-*.js require python3 to be available in PATH because they spawn a Python subprocess for atomic file writes (using os.open with O_NOFOLLOW, dir_fd-based... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-26T21:08:44Z | 2026-05-03T04:33:57Z | 2026-05-01T04:28:47Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | mutterdavid-debug | 260,285,947 | U_kgDOD4Ol-w | User | false |
openclaw/openclaw | 4,332,045,832 | I_kwDOQb6kR88AAAABAjXGCA | 72,370 | https://github.com/openclaw/openclaw/issues/72370 | https://api.github.com/repos/openclaw/openclaw/issues/72370 | Workspace hooks rejected as "cannot override openclaw-managed hook code" and replaced with empty managed versions (2026.4.23+) | ## Summary
On openclaw 2026.4.23 and 2026.4.24 (macOS), **all workspace-defined hooks are silently rejected at gateway startup** with the message:
```
[hooks/workspace] Ignoring openclaw-workspace hook "<name>" because it cannot override openclaw-managed hook code
```
openclaw then loads its own internal versions of... | open | null | false | 1 | [] | [] | 2026-04-26T21:36:36Z | 2026-05-03T13:21:38Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | joeywrightphoto | 114,463,115 | U_kgDOBtKRiw | User | false |
openclaw/openclaw | 4,332,050,444 | I_kwDOQb6kR88AAAABAjXYDA | 72,371 | https://github.com/openclaw/openclaw/issues/72371 | https://api.github.com/repos/openclaw/openclaw/issues/72371 | Beta blocker: google-meet chrome-node realtime bridges can orphan and stop on transient input failure | ## Summary
Chrome-node Google Meet realtime sessions can leave the node-side command-pair audio bridge alive after the gateway/plugin-side session registry loses ownership. A later join for the same Meet URL can start a second node bridge while the stale SoX/BlackHole pair keeps running. Separately, one transient `pul... | closed | completed | true | 0 | [
"beta-blocker"
] | [] | 2026-04-26T21:38:54Z | 2026-04-30T04:36:22Z | 2026-04-27T06:37:20Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | BsnizND | 199,837,910 | U_kgDOC-lI1g | User | false |
openclaw/openclaw | 4,332,044,131 | I_kwDOQb6kR88AAAABAjW_Yw | 72,369 | https://github.com/openclaw/openclaw/issues/72369 | https://api.github.com/repos/openclaw/openclaw/issues/72369 | [webchat] Hide startup trigger message from chat history | ## Summary
When a session resets via \`/new\` or \`/reset\` in the Control UI WebChat, the startup trigger message is stored in the session transcript and rendered as a visible user message in the chat thread. It looks like a user sent it, but it is an internal system instruction emitted by \`openclaw-control-ui\`.
T... | closed | completed | true | 1 | [] | [] | 2026-04-26T21:35:42Z | 2026-04-30T04:34:49Z | 2026-04-27T10:41:34Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | collynes | 17,289,375 | MDQ6VXNlcjE3Mjg5Mzc1 | User | false |
openclaw/openclaw | 4,088,925,593 | I_kwDOQb6kR87zuA2Z | 49,060 | https://github.com/openclaw/openclaw/issues/49060 | https://api.github.com/repos/openclaw/openclaw/issues/49060 | ACP server stdout corruption when logging subsystem writes to stdout | ## Bug Description
The ACP (Agent Client Protocol) server writes JSON-RPC messages to stdout, but doesn't route logging output to stderr. Any logs written through the logging subsystem end up on stdout, corrupting the JSON-RPC stream.
This is the same class of bug previously fixed for the RPC loop in:
- d66a05dc4 ("R... | closed | completed | true | 3 | [] | [] | 2026-03-17T14:25:57Z | 2026-04-29T04:33:43Z | 2026-04-26T21:44:58Z | CONTRIBUTOR | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jb510 | 801,241 | MDQ6VXNlcjgwMTI0MQ== | User | false |
openclaw/openclaw | 4,167,179,452 | I_kwDOQb6kR874Yhy8 | 57,471 | https://github.com/openclaw/openclaw/issues/57471 | https://api.github.com/repos/openclaw/openclaw/issues/57471 | runWithModelFallback() treats LiveSessionModelSwitchError as candidate_failed, causing unnecessary fallback chain | ## Summary
When a live session model switch is active (e.g., session locked to `openai-codex/gpt-5.4`), every request that initially targets a different model (e.g., `minimax-cn/MiniMax-M2.7-highspeed` for compaction) triggers the **entire fallback chain** of 12+ candidates before finally reaching the locked model. Th... | closed | completed | true | 0 | [] | [
"vincentkoc"
] | 2026-03-30T04:23:44Z | 2026-04-29T04:33:42Z | 2026-04-26T21:45:04Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | yuxiaoyang2007-prog | 258,562,502 | U_kgDOD2lZxg | User | false |
openclaw/openclaw | 4,332,142,178 | I_kwDOQb6kR88AAAABAjc-Yg | 72,390 | https://github.com/openclaw/openclaw/issues/72390 | https://api.github.com/repos/openclaw/openclaw/issues/72390 | [Feature]: support SecretRef for env.vars | ### Summary
Include SecretRef support for openclaw.json `env` block to allow a workflow where secrets stored as environment variables can be resolved at runtime (e.g. from 1Password)
### Problem to solve
Currently, secretRef is not supported for the env block in openclaw.json. This makes setting/passing environment ... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-26T22:19:06Z | 2026-05-03T09:13:03Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | edlevin6612 | 17,462,280 | MDQ6VXNlcjE3NDYyMjgw | User | false |
openclaw/openclaw | 4,332,096,005 | I_kwDOQb6kR88AAAABAjaKBQ | 72,380 | https://github.com/openclaw/openclaw/issues/72380 | https://api.github.com/repos/openclaw/openclaw/issues/72380 | [v2026.4.24] doctor --fix installs bundled deps to wrong dir; ENOTEMPTY in plugin-sdk alias crashes node host | ## Summary
On a fresh upgrade to OpenClaw 2026.4.24 (npm global install), two distinct
regressions in the bundled-plugin-runtime-deps machinery prevent the agent
from responding on Telegram (and likely other channels). Both are reproducible
and have local workarounds, but they should be fixed upstream.
## Environment... | closed | completed | true | 1 | [] | [] | 2026-04-26T22:00:32Z | 2026-04-29T04:33:32Z | 2026-04-26T22:19:53Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | MrRRRabbit | 11,497,595 | MDQ6VXNlcjExNDk3NTk1 | User | false |
openclaw/openclaw | 4,332,154,421 | I_kwDOQb6kR88AAAABAjduNQ | 72,394 | https://github.com/openclaw/openclaw/issues/72394 | https://api.github.com/repos/openclaw/openclaw/issues/72394 | Transfer pending system event entries across isolated heartbeat session boundary | ## Summary
When `heartbeat.isolatedSession: true` is enabled, the heartbeat runner uses a separate `:heartbeat` session key for the model run. Pending system event entries enqueued on the base direct session are not transferred to the isolated session.
## Asymmetry with cron path
The cron path already transfers the ... | open | null | false | 1 | [] | [] | 2026-04-26T22:24:42Z | 2026-05-03T09:13:04Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | keeminlee | 67,605,380 | MDQ6VXNlcjY3NjA1Mzgw | User | false |
openclaw/openclaw | 4,332,156,784 | I_kwDOQb6kR88AAAABAjd3cA | 72,396 | https://github.com/openclaw/openclaw/issues/72396 | https://api.github.com/repos/openclaw/openclaw/issues/72396 | Plugin runtime deps: support layered stage dirs for preinstalled baseline deps | ### Summary
Support layered plugin runtime dependency stage dirs, so OpenClaw can resolve preinstalled baseline deps from one location and install new deps into another writable location.
This helps avoid repeated slow, heavy installs while keeping new plugin deps persistent.
### Problem
Installing baseline plugin ... | closed | completed | true | 4 | [] | [] | 2026-04-26T22:25:56Z | 2026-04-30T04:35:43Z | 2026-04-27T08:21:49Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | liorb-mountapps | 205,664,440 | U_kgDODEIwuA | User | false |
openclaw/openclaw | 4,332,156,441 | I_kwDOQb6kR88AAAABAjd2GQ | 72,395 | https://github.com/openclaw/openclaw/issues/72395 | https://api.github.com/repos/openclaw/openclaw/issues/72395 | [Bug]: [Windows] ERR_UNSUPPORTED_ESM_URL_SCHEME — Discord channel and browser plugin fail to load on v2026.4.24 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
On Windows, version 2026.4.24 fails to load the Discord channel plugin and browser plugin due to ERR_UNSUPPORTED_ESM_URL_SCHEME. Node's ESM loader receives raw Windows paths (C:\...) instead of file:// URLs, causing both plu... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-26T22:25:45Z | 2026-04-29T04:33:10Z | 2026-04-27T01:05:44Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Nomar-Garcia | 272,658,558 | U_kgDOEEBwfg | User | false |
openclaw/openclaw | 4,332,209,830 | I_kwDOQb6kR88AAAABAjhGpg | 72,401 | https://github.com/openclaw/openclaw/issues/72401 | https://api.github.com/repos/openclaw/openclaw/issues/72401 | [Bug]: models struggle to use discord tool/skill for sending discord DM's because of target requiring "user:" prefix | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Many models struggle and have to retry several times to send a DM to a discord user via the message tool
### Steps to reproduce
1, A virgin OpenClaw 2026.4.22 (using only prepackaged AGENTS, TOOLS, etc)
2. ... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-26T22:52:41Z | 2026-04-30T04:33:50Z | 2026-04-27T14:03:56Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | garyd9 | 1,250,182 | MDQ6VXNlcjEyNTAxODI= | User | false |
openclaw/openclaw | 4,332,210,732 | I_kwDOQb6kR88AAAABAjhKLA | 72,402 | https://github.com/openclaw/openclaw/issues/72402 | https://api.github.com/repos/openclaw/openclaw/issues/72402 | Ollama Cloud models should receive native think: max instead of mapped think: high | ## Summary
The fix in ff570f3 maps `/think max` to `think: "high"` for **all** Ollama models.
However, Ollama Cloud endpoints (models with `:cloud` suffix) accept the native
`think: "max"` value without rejection.
## Evidence
Live test against `ollama/deepseek-v4-flash:cloud` (local proxy to Ollama Cloud):
```bash
... | closed | not_planned | true | 1 | [] | [] | 2026-04-26T22:53:07Z | 2026-04-29T04:32:38Z | 2026-04-27T02:13:51Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | g0st1n | 279,233,793 | U_kgDOEKTFAQ | User | false |
openclaw/openclaw | 4,332,228,603 | I_kwDOQb6kR88AAAABAjiP-w | 72,407 | https://github.com/openclaw/openclaw/issues/72407 | https://api.github.com/repos/openclaw/openclaw/issues/72407 | [Bug]: Control UI thinking dropdown label ignores agents.defaults.thinkingDefault and per-model params.thinking | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Control UI thinking dropdown shows "Default (off)" even when both `agents.defaults.thinkingDefault: "adaptive"` and per-model `agents.defaults.models[<key>].params.thinking: "adaptive"` are configured for non-C... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-26T23:03:15Z | 2026-04-29T04:32:37Z | 2026-04-27T02:22:19Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | NotecAG | 149,318,311 | U_kgDOCOZqpw | User | false |
openclaw/openclaw | 4,067,398,338 | I_kwDOQb6kR87yb5LC | 44,395 | https://github.com/openclaw/openclaw/issues/44395 | https://api.github.com/repos/openclaw/openclaw/issues/44395 | feat: heading-aware chunking + entity extraction for memory search | ## Problem
OpenClaw's memory chunker (`chunkMarkdown` in `src/memory/internal.ts`) uses fixed-size character-count accumulation (default 400 tokens / ~1600 chars). It walks line-by-line and flushes when a chunk exceeds `maxChars`, with optional overlap.
This means semantically distinct sections (e.g., two different p... | open | null | false | 3 | [] | [] | 2026-03-12T21:33:33Z | 2026-05-04T12:47:18Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | tymrtn | 4,305,058 | MDQ6VXNlcjQzMDUwNTg= | User | false |
openclaw/openclaw | 4,332,245,294 | I_kwDOQb6kR88AAAABAjjRLg | 72,415 | https://github.com/openclaw/openclaw/issues/72415 | https://api.github.com/repos/openclaw/openclaw/issues/72415 | [Bug]: ERR_MODULE_NOT_FOUND 'openclaw' still happening on 2026.4.22 (Telegram channel + scheduled cron) — same root cause as #71484, fix appears incomplete on upgrade path | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Same root cause as #71484 (closed 2026-04-25 as completed) — the bundled-deps installer for `~/.openclaw/plugin-runtime-deps/openclaw-<VER>-<HASH>/node_modules/` populates ~250 transitive packages but does **not** install or... | closed | completed | true | 1 | [] | [] | 2026-04-26T23:11:44Z | 2026-04-29T04:33:23Z | 2026-04-27T00:02:29Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ortisiano | 73,437,003 | MDQ6VXNlcjczNDM3MDAz | User | false |
openclaw/openclaw | 4,332,265,381 | I_kwDOQb6kR88AAAABAjkfpQ | 72,418 | https://github.com/openclaw/openclaw/issues/72418 | https://api.github.com/repos/openclaw/openclaw/issues/72418 | [Bug]: shouldSkipBackendSelfPairing allows loopback clients to self-declare GATEWAY_CLIENT identity and bypass device pairing | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 8.7 / 10.0 | 9.3 / 10.0 |
| **Severity** | High | Critical |
| **Vector** | CVSS:3.1/AV:A/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:N | CVSS:4.0/AV:A/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:N/SC:H/SI:H/SA:N |
| **Calculator** | [CV... | open | null | false | 1 | [] | [] | 2026-04-26T23:23:19Z | 2026-05-03T06:07:36Z | null | CONTRIBUTOR | null | 20260503T090142Z | 2026-05-03T09:01:42Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 4,332,268,224 | I_kwDOQb6kR88AAAABAjkqwA | 72,420 | https://github.com/openclaw/openclaw/issues/72420 | https://api.github.com/repos/openclaw/openclaw/issues/72420 | Bonjour plugin crash-loops gateway on hosts without working multicast (2026.4.24) | ### Environment
- OpenClaw: 2026.4.24 (cbcfdf6), upgraded from 2026.4.5
- Host: Ubuntu 24.04.4 LTS on a Hostinger VPS (no real LAN multicast)
- Node: v22.22.1
- Install method: npm global
### Symptom
Gateway enters a restart loop with restart counter climbing every ~45 seconds.
Stability bundle written for each crash;... | closed | completed | true | 1 | [] | [] | 2026-04-26T23:25:04Z | 2026-04-29T04:33:22Z | 2026-04-27T00:02:34Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | modafoca | 70,458,567 | MDQ6VXNlcjcwNDU4NTY3 | User | false |
openclaw/openclaw | 4,332,015,553 | I_kwDOQb6kR88AAAABAjVPwQ | 72,366 | https://github.com/openclaw/openclaw/issues/72366 | https://api.github.com/repos/openclaw/openclaw/issues/72366 | [Bug]: Gateway crash-loop on startup when bonjour/mDNS plugin hits "CIAO PROBING CANCELED" | Bug type
Crash (process/app exits or hangs)
Beta release blocker
No
Summary
Gateway crash-loops on startup when the bonjour/mDNS plugin hits CIAO PROBING CANCELLED, and disabling bonjour stops the failure.
Steps to reproduce
Install OpenClaw on a Linux host with the bonjour plugin enabled (default)
Start the gateway... | closed | completed | true | 7 | [
"bug",
"bug:crash"
] | [] | 2026-04-26T21:23:18Z | 2026-04-29T04:33:03Z | 2026-04-27T01:18:40Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | zachzion762 | 152,152,819 | U_kgDOCRGq8w | User | false |
openclaw/openclaw | 4,332,277,424 | I_kwDOQb6kR88AAAABAjlOsA | 72,421 | https://github.com/openclaw/openclaw/issues/72421 | https://api.github.com/repos/openclaw/openclaw/issues/72421 | Discord channel lanes intermittently held after run completes — "lane wait exceeded" with stale per-session jsonl mtime | ## Symptom
User-sent Discord messages stop being processed in a channel even though the gateway is healthy and other channels work. Diagnostic log emits repeatedly:
```
lane wait exceeded: lane=session:agent:main:discord:channel:<id> waitedMs=<>200000 queueAhead=0|1
```
with no concurrent LLM activity (per-session `... | closed | not_planned | true | 1 | [] | [] | 2026-04-26T23:30:40Z | 2026-05-05T04:26:47Z | 2026-05-02T16:26:27Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | jessewunderlich | 83,950,240 | MDQ6VXNlcjgzOTUwMjQw | User | false |
openclaw/openclaw | 4,331,990,325 | I_kwDOQb6kR88AAAABAjTtNQ | 72,364 | https://github.com/openclaw/openclaw/issues/72364 | https://api.github.com/repos/openclaw/openclaw/issues/72364 | Treat auto-compaction as a first-class agent task (route through /compact) to eliminate mid-conversation transcript mutations | ## Summary
Auto-compaction in 2026.4.24 fires invisibly inside the next-turn prompt-build path. When `agents.defaults.compaction.truncateAfterCompaction: true` is set, this produces visibly disruptive behavior in long-running channel sessions (especially WebChat-viewed Telegram/Slack DMs): messages render and then dis... | closed | completed | true | 4 | [] | [] | 2026-04-26T21:10:33Z | 2026-04-29T04:33:04Z | 2026-04-27T01:18:33Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jason-alvarez-git | 188,690,207 | U_kgDOCz8vHw | User | false |
openclaw/openclaw | 4,332,058,570 | I_kwDOQb6kR88AAAABAjX3yg | 72,373 | https://github.com/openclaw/openclaw/issues/72373 | https://api.github.com/repos/openclaw/openclaw/issues/72373 | OpenShell should fail fast on malformed generated commands and hard-abort repeated tool loops | ## Summary
OpenShell currently accepts clearly malformed/generated shell commands and hands them to `/bin/sh`; repeated identical failed tool calls are only warned on rather than hard-aborted. In a chat-surface session this can create a long-running “typing/processing” wedge when a model repeatedly emits malformed com... | open | null | false | 1 | [] | [] | 2026-04-26T21:42:54Z | 2026-05-03T10:19:37Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | Kaspnov | 207,042,562 | U_kgDODFc4Ag | User | false |
openclaw/openclaw | 4,332,290,261 | I_kwDOQb6kR88AAAABAjmA1Q | 72,425 | https://github.com/openclaw/openclaw/issues/72425 | https://api.github.com/repos/openclaw/openclaw/issues/72425 | Google Live tool responses omit function names | ## Summary
Google Live realtime tool responses omit the original function call `name`, so Gemini Live tool calls can fail when OpenClaw submits the result back through the official `@google/genai` SDK.
This shows up most visibly in Google Meet realtime sessions using `gemini-3.1-flash-live-preview`: the model greets ... | closed | completed | true | 1 | [] | [] | 2026-04-26T23:37:46Z | 2026-04-30T04:35:35Z | 2026-04-27T08:58:26Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | BsnizND | 199,837,910 | U_kgDOC-lI1g | User | false |
openclaw/openclaw | 4,332,294,464 | I_kwDOQb6kR88AAAABAjmRQA | 72,427 | https://github.com/openclaw/openclaw/issues/72427 | https://api.github.com/repos/openclaw/openclaw/issues/72427 | [Discord] Duplicate messages sent to Discord DM (同一消息出现两次) | ## Bug Description
Discord DM 频道，bot 发送的消息在 Discord 中重复出现（同一消息出现两次）。
**环境：**
- OpenClaw: 2026.4.24
- macOS: 26.4.1
- Node: 25.8.1
- Channel: discord (direct DM)
**复现步骤：**
1. 在 Discord DM 中发送任意消息给 bot
2. Bot 回复正常，但同一条回复在 Discord 中出现两次
**尝试过的修复：**
1. 清除 plugin-runtime-deps 缓存：`rm -rf /Users/jeyo/.openclaw/plugin-runt... | closed | completed | true | 0 | [] | [] | 2026-04-26T23:39:48Z | 2026-04-29T04:33:29Z | 2026-04-26T23:40:39Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | JeyoWang | 268,744,013 | U_kgDOEAS1TQ | User | false |
openclaw/openclaw | 4,332,298,484 | I_kwDOQb6kR88AAAABAjmg9A | 72,429 | https://github.com/openclaw/openclaw/issues/72429 | https://api.github.com/repos/openclaw/openclaw/issues/72429 | [Bug]: Duplication of messages with minmax 2.7 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
"When using MiniMax-M2.7 directly, the model immediately loops, hallucinating both questions and answers. Switching to gemini fixes it instantly, so it seems like OpenClaw is missing the proper stop token... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-26T23:41:41Z | 2026-04-29T04:33:09Z | 2026-04-27T01:05:51Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | busaeedw | 238,569,950 | U_kgDODjhJ3g | User | false |
openclaw/openclaw | 4,332,299,672 | I_kwDOQb6kR88AAAABAjmlmA | 72,430 | https://github.com/openclaw/openclaw/issues/72430 | https://api.github.com/repos/openclaw/openclaw/issues/72430 | [Bug]: Windows: Slack channel fails to start with ESM URL error (`c:` protocol) in v2026.4.24 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Windows: Slack channel fails to start with ESM URL error (`c:` protocol) in v2026.4.24
**Version:** 2026.4.24
**Platform:** Windows 11
**Node.js:** 24.14.0
### Summary
After updating to v2026.4.24, the Slack ... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-26T23:42:23Z | 2026-04-29T04:33:08Z | 2026-04-27T01:05:57Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jamaspoli | 259,723,170 | U_kgDOD3sPog | User | false |
openclaw/openclaw | 4,332,107,498 | I_kwDOQb6kR88AAAABAja26g | 72,382 | https://github.com/openclaw/openclaw/issues/72382 | https://api.github.com/repos/openclaw/openclaw/issues/72382 | [Bug]: General install | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Just trying to install. Used an old Macbook (2015). Got mired in dependencies not met with MacOS. Assumed old Macbook, installed Ubuntu 26.04.
First try: ERROR: Node.js v22+ required but could not be activated ... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-26T22:05:03Z | 2026-05-03T09:13:02Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | jbob762 | 53,946,060 | MDQ6VXNlcjUzOTQ2MDYw | User | false |
openclaw/openclaw | 4,332,059,069 | I_kwDOQb6kR88AAAABAjX5vQ | 72,374 | https://github.com/openclaw/openclaw/issues/72374 | https://api.github.com/repos/openclaw/openclaw/issues/72374 | Bundled runtime dependency staging should create an isolated package root before npm install | ## Summary
OpenClaw's bundled runtime dependency staging can run `npm install` from a staging directory that has no root `package.json`. npm then walks upward, finds an unrelated parent `package.json`, and installs that project’s dependencies instead of an isolated minimal runtime dependency set.
In one local inciden... | closed | completed | true | 1 | [] | [] | 2026-04-26T21:43:09Z | 2026-04-29T04:33:29Z | 2026-04-26T23:50:31Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Kaspnov | 207,042,562 | U_kgDODFc4Ag | User | false |
openclaw/openclaw | 4,332,131,924 | I_kwDOQb6kR88AAAABAjcWVA | 72,386 | https://github.com/openclaw/openclaw/issues/72386 | https://api.github.com/repos/openclaw/openclaw/issues/72386 | [Bug] v2026.4.25-beta.4: new runtime-context message text echoed verbatim by model into visible reply | ## Summary
**v2026.4.25-beta.4 regression vs v2026.4.24.** A new `openclaw.runtime-context` custom message is injected into the model's conversation history on every Telegram inbound. The injected message starts with the literal preface:
> ```
> OpenClaw runtime context for the immediately preceding user message.
> T... | closed | completed | true | 4 | [] | [] | 2026-04-26T22:15:00Z | 2026-04-30T04:33:12Z | 2026-04-27T20:28:10Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | jhsmith409 | 198,252,919 | U_kgDOC9EZdw | User | false |
openclaw/openclaw | 4,332,329,367 | I_kwDOQb6kR88AAAABAjoZlw | 72,434 | https://github.com/openclaw/openclaw/issues/72434 | https://api.github.com/repos/openclaw/openclaw/issues/72434 | [Bug]: Regression in 2026.4.24: agent harness "claude-cli" is not registered, all gateway requests fail | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
After upgrading from 2026.4.23 to 2026.4.24, every gateway request fails with "Requested agent harness 'claude-cli' is not registered and PI fallback is disabled" — affecting all Claude model fallbacks and comp... | closed | completed | true | 8 | [
"bug",
"bug:behavior"
] | [] | 2026-04-26T23:57:10Z | 2026-04-29T04:31:03Z | 2026-04-27T03:21:29Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | KMMossy | 270,393,824 | U_kgDOEB3h4A | User | false |
openclaw/openclaw | 4,332,365,451 | I_kwDOQb6kR88AAAABAjqmiw | 72,438 | https://github.com/openclaw/openclaw/issues/72438 | https://api.github.com/repos/openclaw/openclaw/issues/72438 | Gateway WS handshake timeout + plugin re-stage cache miss in 2026.4.24 (regression vs 4.23) | ## Summary
Upgrading from `2026.4.23` → `2026.4.24` introduces two distinct event-loop hazards that make the gateway and embedded modes effectively unusable for real workloads. Rolling back to `2026.4.23` restores both paths to working order with all the same agent/plugin config.
## Environment
- macOS 15 (Darwin 25... | closed | completed | true | 1 | [] | [] | 2026-04-27T00:14:45Z | 2026-04-30T04:36:03Z | 2026-04-27T07:28:36Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | WattyouSay | 269,988,426 | U_kgDOEBeySg | User | false |
openclaw/openclaw | 4,332,383,464 | I_kwDOQb6kR88AAAABAjrs6A | 72,440 | https://github.com/openclaw/openclaw/issues/72440 | https://api.github.com/repos/openclaw/openclaw/issues/72440 | Google Meet tool-created realtime sessions can lose gateway ownership | ## Summary
When the `google_meet` agent tool creates or joins a realtime Google Meet room, the live session can be owned only by the tool-turn plugin runtime instead of the gateway-owned `googlemeet.*` runtime. After the agent turn completes, `googlemeet.status`, `googlemeet.leave`, and `googlemeet.speak` can no longe... | closed | completed | true | 1 | [] | [] | 2026-04-27T00:22:12Z | 2026-04-30T04:35:41Z | 2026-04-27T08:28:16Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | BsnizND | 199,837,910 | U_kgDOC-lI1g | User | false |
openclaw/openclaw | 4,084,375,630 | I_kwDOQb6kR87zcqBO | 48,457 | https://github.com/openclaw/openclaw/issues/48457 | https://api.github.com/repos/openclaw/openclaw/issues/48457 | [Bug]: nodes.run (and node exec) fails for interpreter one-liners (python3 -c, bash -c) with SYSTEM_RUN_DENIED: approval cannot safely bind even when approvals are security=full, ask=off | ### Bug type
Regression (worked before, now fails)
### Summary
After updating from v2026.3.8 → v2026.3.13, nodes.run on a macOS node host fails for interpreter one‑liners like `python3 -c …` and `bash -c …` with:
`INVALID_REQUEST: SYSTEM_RUN_DENIED: approval cannot safely bind this interpreter/runtime command`
Thi... | closed | completed | true | 8 | [
"bug",
"security",
"regression"
] | [] | 2026-03-16T19:29:47Z | 2026-04-29T04:33:18Z | 2026-04-27T00:29:07Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ajtran303 | 31,839,316 | MDQ6VXNlcjMxODM5MzE2 | User | false |
openclaw/openclaw | 4,332,410,820 | I_kwDOQb6kR88AAAABAjtXxA | 72,447 | https://github.com/openclaw/openclaw/issues/72447 | https://api.github.com/repos/openclaw/openclaw/issues/72447 | Feature: Typing indicator during agent active processing | ## Problem
Currently OpenClaw sends a typing indicator only during heartbeats (~30 min interval), not during active agent processing. Users see no feedback while the agent is thinking or working.
## Expected Behavior
When the agent receives a message and is actively processing (model inference, tool execution, etc.)... | closed | completed | true | 1 | [] | [] | 2026-04-27T00:33:20Z | 2026-04-30T04:35:47Z | 2026-04-27T08:01:18Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | urrielu | 213,940,341 | U_kgDODMB4dQ | User | false |
openclaw/openclaw | 4,232,790,220 | I_kwDOQb6kR878S0DM | 63,775 | https://github.com/openclaw/openclaw/issues/63775 | https://api.github.com/repos/openclaw/openclaw/issues/63775 | [Bug]: Multi-ollama provider routing still broken after #61776 — requests silently go to first provider | ## Follow-up to #61678 (locked)
#61678 was closed by @steipete after PR #61776 merged, but the fix is incomplete. The issue is locked so I cannot comment there.
## Bug: Multi-ollama provider routing still broken
After upgrading to **OpenClaw 2026.4.8** (which includes PR #61776), all requests to `ollama2` are still ... | closed | completed | true | 1 | [] | [] | 2026-04-09T14:01:38Z | 2026-04-29T04:33:16Z | 2026-04-27T00:34:43Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | adamai26 | 271,567,870 | U_kgDOEC_L_g | User | false |
openclaw/openclaw | 4,220,958,959 | I_kwDOQb6kR877lrjv | 62,735 | https://github.com/openclaw/openclaw/issues/62735 | https://api.github.com/repos/openclaw/openclaw/issues/62735 | Regression: `api: "ollama"` rejected for custom-URL providers; no working path for `num_ctx` injection through a local proxy | ## Summary
As of the 2026.4.x release, OpenClaw refuses to use the `ollama` API type for any provider that isn't the built-in Ollama URL. Custom provider entries in `openclaw.json` with `"api": "ollama"` now fail immediately with:
```
No API provider registered for api: ollama
```
This worked in prior releases (conf... | closed | completed | true | 1 | [] | [] | 2026-04-07T22:32:52Z | 2026-04-29T04:33:15Z | 2026-04-27T00:34:45Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | STaveras | 6,901,694 | MDQ6VXNlcjY5MDE2OTQ= | User | false |
openclaw/openclaw | 4,332,419,371 | I_kwDOQb6kR88AAAABAjt5Kw | 72,450 | https://github.com/openclaw/openclaw/issues/72450 | https://api.github.com/repos/openclaw/openclaw/issues/72450 | [Bug]: web ui日志界面无信息 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
升级到2026.4.24后web ui日志无内容
<img width="706" height="1105" alt="Image" src="https://github.com/user-attachments/assets/85364ecd-95d2-43db-9c3f-ed6dd5e705f7" />
### Steps to reproduce
升级到2026.4.24
### Expected behavior
有日志显... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-27T00:36:45Z | 2026-05-03T04:37:26Z | 2026-04-30T13:50:11Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | sobergou | 50,661,423 | MDQ6VXNlcjUwNjYxNDIz | User | false |
openclaw/openclaw | 4,332,435,276 | I_kwDOQb6kR88AAAABAju3TA | 72,459 | https://github.com/openclaw/openclaw/issues/72459 | https://api.github.com/repos/openclaw/openclaw/issues/72459 | [Bug]: Plugin loader silently drops register* on missing required fields — should hard-fail or throw | ### Bug type
Behavior bug — silent failure during plugin load.
### Beta release blocker
No
### Summary
The plugin loader silently drops `register*` calls that fail acceptance gates. The plugin still loads, the surface never fires, and the only signal is a `level=warn` diagnostic in `gateway.log`. Plugins ship gree... | closed | completed | true | 0 | [] | [] | 2026-04-27T00:42:49Z | 2026-04-30T04:36:37Z | 2026-04-27T05:46:52Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | 1fanwang | 21,269,208 | MDQ6VXNlcjIxMjY5MjA4 | User | false |
openclaw/openclaw | 4,197,832,508 | I_kwDOQb6kR876Ndc8 | 60,000 | https://github.com/openclaw/openclaw/issues/60000 | https://api.github.com/repos/openclaw/openclaw/issues/60000 | [Bug]: control-ui writes exec approval entries with unsupported `source` field, causing repeated INVALID_REQUEST errors | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Every time the OpenClaw control-ui (webchat) connects to the gateway, it attempts to sync exec approval data via `exec.approvals.set`. The entries it writes include a `source` field (e.g., `"source": "allow-alw... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-03T02:01:59Z | 2026-04-29T04:33:13Z | 2026-04-27T00:46:52Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | sd1471123 | 5,255,486 | MDQ6VXNlcjUyNTU0ODY= | User | false |
openclaw/openclaw | 4,332,453,914 | I_kwDOQb6kR88AAAABAjwAGg | 72,463 | https://github.com/openclaw/openclaw/issues/72463 | https://api.github.com/repos/openclaw/openclaw/issues/72463 | [Bug]: Chat Channels: Both QQBot and Feishu are currently unusable. I am using the Windows version. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The error has persisted across versions 4.22 through 4.24.
<img width="2846" height="1659" alt="Image" src="https://github.com/user-attachments/assets/0afeabee-8771-4284-a77a-f14b96e7afb1" />
When using OpenClaw, QQBot, an... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-27T00:50:35Z | 2026-04-30T04:35:53Z | 2026-04-27T07:47:21Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | 1362chenbai | 142,016,729 | U_kgDOCHcA2Q | User | false |
openclaw/openclaw | 4,332,489,905 | I_kwDOQb6kR88AAAABAjyMsQ | 72,470 | https://github.com/openclaw/openclaw/issues/72470 | https://api.github.com/repos/openclaw/openclaw/issues/72470 | [Bug]: Gateway ERR_MODULE_NOT_FOUND after update + repeated unhandled rejections (empty snapshots) | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
Gateway startup fails with `ERR_MODULE_NOT_FOUND` after update, accompanied by repeated unhandled rejections with empty event snapshots
### Steps to reproduce
1. Trigger an OpenClaw update on 2026.4.25-beta.1 (63803d7)
2. The... | closed | completed | true | 0 | [] | [] | 2026-04-27T01:05:51Z | 2026-04-29T04:33:08Z | 2026-04-27T01:09:00Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | yu200512 | 177,332,642 | U_kgDOCpHhog | User | false |
openclaw/openclaw | 4,076,765,556 | I_kwDOQb6kR87y_oF0 | 46,584 | https://github.com/openclaw/openclaw/issues/46584 | https://api.github.com/repos/openclaw/openclaw/issues/46584 | Bug: Memory search ollama API key false warning | ## Bug: Memory search ollama API key 误报
### 问题描述
当 memorySearch.provider 设置为 "ollama"（本地模型）时，openclaw doctor 会显示警告：
> "Memory search provider is set to 'ollama' but the API key was not found in the CLI environment."
但实际上第二句显示：
> "The running gateway reports memory embeddings are ready for the default agent."
说明 embe... | closed | completed | true | 2 | [] | [] | 2026-03-14T21:21:37Z | 2026-04-29T04:33:07Z | 2026-04-27T01:11:35Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | fengly78 | 70,963,857 | MDQ6VXNlcjcwOTYzODU3 | User | false |
openclaw/openclaw | 4,163,674,492 | I_kwDOQb6kR874LKF8 | 56,939 | https://github.com/openclaw/openclaw/issues/56939 | https://api.github.com/repos/openclaw/openclaw/issues/56939 | Unexpected local Ollama probing to 127.0.0.1:11434 on 2026.3.28 when Ollama is not explicitly configured | ## Summary
After upgrading to OpenClaw `2026.3.28`, I started seeing repeated local connection attempts from `openclaw-gateway` (`node`) to `127.0.0.1:11434`, even though I was not intentionally using Ollama and did not have an explicit Ollama provider configured in `openclaw.json`.
This behavior was visible both in ... | closed | completed | true | 1 | [] | [] | 2026-03-29T10:13:16Z | 2026-04-29T04:33:05Z | 2026-04-27T01:11:41Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | IanxDev | 7,808,071 | MDQ6VXNlcjc4MDgwNzE= | User | false |
openclaw/openclaw | 4,332,463,955 | I_kwDOQb6kR88AAAABAjwnUw | 72,465 | https://github.com/openclaw/openclaw/issues/72465 | https://api.github.com/repos/openclaw/openclaw/issues/72465 | [Bug]: openclaw status fails with "PlatformAdapter not registered" from bundled qqbot extension on CLI context | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Environment
OpenClaw version: 2026.4.24 (cbcfdf6)
OS: Windows 11 (10.0.26200.8246)
Node.js: installed via npm global
Installation: npm install -g openclaw
Description
Running openclaw status fails immediately with Platfor... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-27T00:54:56Z | 2026-04-30T04:36:24Z | 2026-04-27T06:33:12Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | aleneves | 62,732,177 | MDQ6VXNlcjYyNzMyMTc3 | User | false |
openclaw/openclaw | 4,332,548,488 | I_kwDOQb6kR88AAAABAj1xiA | 72,477 | https://github.com/openclaw/openclaw/issues/72477 | https://api.github.com/repos/openclaw/openclaw/issues/72477 | `models.providers.<id>.api` enum mismatch crashes gateway on startup — no recovery without manual config edit | ---
title: "`models.providers.<id>.api` enum mismatch crashes gateway on startup — no recovery without manual config edit"
labels: bug
---
## Summary
When the `models.providers.<id>.api` enum values are refined across versions (or when a provider is configured with a value that is later removed from the enum), the ga... | closed | completed | true | 0 | [] | [] | 2026-04-27T01:24:45Z | 2026-04-29T04:32:09Z | 2026-04-27T03:37:33Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | JooyoungChoi14 | 63,181,822 | MDQ6VXNlcjYzMTgxODIy | User | false |
openclaw/openclaw | 4,332,551,182 | I_kwDOQb6kR88AAAABAj18Dg | 72,478 | https://github.com/openclaw/openclaw/issues/72478 | https://api.github.com/repos/openclaw/openclaw/issues/72478 | Google Meet: support reliable headless agent join with audio/transcription health checks | ## Summary
We tested the OpenClaw Google Meet integration on macOS with OpenClaw 2026.4.24 and were able to get partial/visual meeting presence, but not a reliable realtime audio participant. This issue is both a bug report from the failed test and a product request: please move the Google Meet integration toward a su... | closed | completed | true | 3 | [] | [] | 2026-04-27T01:25:42Z | 2026-05-05T04:28:23Z | 2026-05-02T07:07:46Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | DougButdorf | 68,029,180 | MDQ6VXNlcjY4MDI5MTgw | User | false |
openclaw/openclaw | 4,332,553,022 | I_kwDOQb6kR88AAAABAj2DPg | 72,479 | https://github.com/openclaw/openclaw/issues/72479 | https://api.github.com/repos/openclaw/openclaw/issues/72479 | [Bug]: QQbot initialize failure after upgrading to 2026.04.24 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading from 2026.4.6 to 2026.4.24, qqbot fails to initialize.
The actual error is:
> qqbot failed during register from C:\Users\nemom\AppData\Roaming\npm\node_modules\openclaw\dist\extensions\qqbot\index.js: Error:... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-27T01:26:14Z | 2026-04-30T04:35:52Z | 2026-04-27T07:47:23Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | amarillonmc | 7,876,442 | MDQ6VXNlcjc4NzY0NDI= | User | false |
openclaw/openclaw | 4,332,554,676 | I_kwDOQb6kR88AAAABAj2JtA | 72,481 | https://github.com/openclaw/openclaw/issues/72481 | https://api.github.com/repos/openclaw/openclaw/issues/72481 | [Bug]: bonjour: registerUnhandledRejectionHandler not plumbed in; watchdog teardown crashes gateway | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
## Environment
- OpenClaw version: 2026.4.24 (cbcfdf6)
- Node: 22.22.0
- OS: macOS Tahoe 26.x (Apple Silicon, ARM64)
- Install method: `npm install -g openclaw` (Homebrew npm)
## Summary
`@openclaw/bonjour` already has the r... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-04-27T01:26:44Z | 2026-04-29T04:32:09Z | 2026-04-27T03:44:27Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | alfredo-torrijos | 262,460,514 | U_kgDOD6TUYg | User | false |
openclaw/openclaw | 4,332,566,839 | I_kwDOQb6kR88AAAABAj25Nw | 72,484 | https://github.com/openclaw/openclaw/issues/72484 | https://api.github.com/repos/openclaw/openclaw/issues/72484 | QMD memory search fans out one subprocess per collection | ## Summary
QMD-backed memory search currently fans out one `qmd search`/`qmd query` subprocess per managed collection when more than one collection is configured. For agents with many QMD memory collections, a single `memory_search` tool call can spend most of its latency serially starting QMD processes.
This is espe... | closed | completed | true | 2 | [] | [] | 2026-04-27T01:30:43Z | 2026-04-30T04:33:55Z | 2026-04-27T13:37:36Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | BsnizND | 199,837,910 | U_kgDOC-lI1g | User | false |
openclaw/openclaw | 4,332,567,896 | I_kwDOQb6kR88AAAABAj29WA | 72,486 | https://github.com/openclaw/openclaw/issues/72486 | https://api.github.com/repos/openclaw/openclaw/issues/72486 | [Feature]: Add Vietnamese (vi) language support to Control UI | ### Summary
Add Vietnamese as a supported language option in the OpenClaw Control UI language selector.
### Problem to solve
Vietnamese users (a growing OpenClaw community) cannot use the Control UI in their native language. Currently only English is available as a practical option. Vietnam has a large and active se... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-27T01:31:11Z | 2026-05-02T04:27:28Z | 2026-04-29T19:01:25Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | lohichill | 217,164,394 | U_kgDODPGqag | User | false |
openclaw/openclaw | 4,075,425,486 | I_kwDOQb6kR87y6g7O | 45,982 | https://github.com/openclaw/openclaw/issues/45982 | https://api.github.com/repos/openclaw/openclaw/issues/45982 | memory-lancedb: Ollama embedding returns wrong dimensions due to base64 encoding | ## Bug Description
When using Ollama as the embedding provider for `memory-lancedb`, vector search fails with dimension mismatch error:
```
Error: Failed to execute query stream: GenericFailure, Invalid input,
No vector column found to match with the query vector dimension: 192
```
The expected dimension i... | closed | completed | true | 1 | [] | [] | 2026-03-14T10:24:27Z | 2026-04-29T04:33:02Z | 2026-04-27T01:43:51Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | thiswind | 23,764 | MDQ6VXNlcjIzNzY0 | User | false |
openclaw/openclaw | 4,052,070,063 | I_kwDOQb6kR87xha6v | 42,213 | https://github.com/openclaw/openclaw/issues/42213 | https://api.github.com/repos/openclaw/openclaw/issues/42213 | [Bug]: Control UI chat can stay stuck busy and stop flushing queued messages after a long run | ## Summary
Control UI chat can get stuck in a busy state after a long-running `chat.send` run. When that happens, the next user message is queued in the UI but never flushes automatically. Sending another message later can cause the previous queued work to resume.
## Symptoms
- A long task appears to stop progressin... | open | null | false | 1 | [] | [] | 2026-03-10T14:09:04Z | 2026-05-04T10:38:00Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | siewcapital | 264,786,654 | U_kgDOD8hS3g | User | false |
openclaw/openclaw | 4,055,850,904 | I_kwDOQb6kR87xv1-Y | 42,819 | https://github.com/openclaw/openclaw/issues/42819 | https://api.github.com/repos/openclaw/openclaw/issues/42819 | [Bug]: browser in docker hidden in cache folder | ### Bug type
Regression (worked before, now fails)
### Summary
(1) Currently, we have to get chrome by path like /home/node/.cache/ms-playwright/chromium-1208/chrome-linux64/chrome
but when its version may be upgraded after a new build (> 12.08), the absolute path will change.
so a softlink is a better way.
(2) us... | open | null | false | 4 | [
"bug",
"regression"
] | [] | 2026-03-11T05:05:28Z | 2026-05-04T20:13:50Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | panpan0000 | 14,049,268 | MDQ6VXNlcjE0MDQ5MjY4 | User | false |
openclaw/openclaw | 4,105,477,587 | I_kwDOQb6kR870tJ3T | 50,799 | https://github.com/openclaw/openclaw/issues/50799 | https://api.github.com/repos/openclaw/openclaw/issues/50799 | macOS GUI attachment picker only allows image files; documents are visible but disabled | In the OpenClaw macOS GUI, clicking the paperclip attachment button opens a file picker where non-image files are visible but dimmed/unselectable. Image files such as `.jpg` and `.png` are selectable.
### Steps to reproduce
1. Open OpenClaw on macOS
2. Click the paperclip attachment button
3. Navigate to a folder cont... | open | null | false | 1 | [] | [] | 2026-03-20T03:25:37Z | 2026-05-04T18:23:59Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | A1fred-AI | 269,621,014 | U_kgDOEBIXFg | User | false |
openclaw/openclaw | 4,202,722,614 | I_kwDOQb6kR876gHU2 | 60,602 | https://github.com/openclaw/openclaw/issues/60602 | https://api.github.com/repos/openclaw/openclaw/issues/60602 | Feature Request: Per-Agent Bedrock requestMetadata Injection for Multi-Agent Cost Attribution | ## Summary
When running multiple agents via OpenClaw's multi-agent architecture, there's no way to attribute Bedrock API costs to individual agents. All calls appear as the same IAM role with no distinguishing metadata.
## Problem
- OpenClaw supports multiple agents (e.g., 6 agents sharing one EC2 instance)
- All Be... | open | null | false | 1 | [] | [] | 2026-04-04T00:21:26Z | 2026-05-03T10:18:58Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | maxatv | 215,872,058 | U_kgDODN3yOg | User | false |
openclaw/openclaw | 4,222,465,866 | I_kwDOQb6kR877rbdK | 62,968 | https://github.com/openclaw/openclaw/issues/62968 | https://api.github.com/repos/openclaw/openclaw/issues/62968 | bug: subagent announce direct-send error silently swallowed when queue fallback recovers | ## Problem
When `sendSubagentAnnounceDirectly` fails but the queue fallback succeeds in `runSubagentAnnounceDispatch`, the direct-send error is silently swallowed — never logged.
### Code path
**`src/agents/subagent-announce-delivery.ts`** — `sendSubagentAnnounceDirectly` catches errors and returns them in the resul... | open | null | false | 1 | [] | [] | 2026-04-08T06:06:02Z | 2026-05-02T18:31:11Z | null | NONE | null | 20260502T210030Z | 2026-05-02T21:00:30Z | alexey-pelykh | 345,102 | MDQ6VXNlcjM0NTEwMg== | User | false |
openclaw/openclaw | 4,222,523,570 | I_kwDOQb6kR877rpiy | 62,974 | https://github.com/openclaw/openclaw/issues/62974 | https://api.github.com/repos/openclaw/openclaw/issues/62974 | Feature: cache-aware sticky fallback to prevent prompt cache bouncing across providers | ## Summary
When the primary provider (e.g. Anthropic direct) intermittently times out and falls back to a secondary provider (e.g. OpenRouter/Bedrock), the per-request fallback design causes **prompt cache bouncing**: each successful primary request warms cache on Provider A, while each fallback request needs a full c... | closed | completed | true | 1 | [] | [] | 2026-04-08T06:20:06Z | 2026-04-30T04:35:00Z | 2026-04-27T10:22:30Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | liu51115 | 31,350,941 | MDQ6VXNlcjMxMzUwOTQx | User | false |
openclaw/openclaw | 4,223,102,766 | I_kwDOQb6kR877t28u | 63,024 | https://github.com/openclaw/openclaw/issues/63024 | https://api.github.com/repos/openclaw/openclaw/issues/63024 | [Bug]: openclaw onboard --install-daemon error "Did you mean to import "tslog/cjs/index.js"?" | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
root@i-69c243f6aa1c6e1bccc751dc:/ai# openclaw onboard --install-daemon
node:internal/modules/run_main:107
triggerUncaughtException(
^
Error: Cannot find package '/root/.local/share/pnpm/global/5/.pnpm/openclaw@2026.... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-08T08:16:53Z | 2026-04-30T04:34:59Z | 2026-04-27T10:25:25Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | hu19940121 | 26,077,620 | MDQ6VXNlcjI2MDc3NjIw | User | false |
openclaw/openclaw | 4,223,456,008 | I_kwDOQb6kR877vNMI | 63,069 | https://github.com/openclaw/openclaw/issues/63069 | https://api.github.com/repos/openclaw/openclaw/issues/63069 | [Bug]: hell completion writes to wrong profile path when ZDOTDIR or XDG_CONFIG_HOME is set | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Shell completion install writes to hardcoded `~/.zshrc`, `~/.config/fish/config.fish`, etc. instead of respecting `$ZDOTDIR`, `$XDG_CONFIG_HOME`, and bash `.bash_profile` fallback, causing the completion source... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-08T09:21:42Z | 2026-05-02T20:57:11Z | null | NONE | null | 20260502T210030Z | 2026-05-02T21:00:30Z | zh1C | 41,350,407 | MDQ6VXNlcjQxMzUwNDA3 | User | false |
openclaw/openclaw | 4,223,922,523 | I_kwDOQb6kR877w_Fb | 63,106 | https://github.com/openclaw/openclaw/issues/63106 | https://api.github.com/repos/openclaw/openclaw/issues/63106 | cron: ghost runs recorded as ok when gateway is down (durationMs < 50ms) | ## Problem
When the OpenClaw gateway goes down (e.g. due to a config error or model routing failure), cron jobs configured with `sessionTarget: "main"` continue to fire on schedule and record `status: "ok"` in the run log. The runs complete in 1–8 ms (`durationMs: 1`) because `executeMainSessionCronJob` returns immedi... | open | null | false | 1 | [] | [] | 2026-04-08T10:43:42Z | 2026-05-02T19:15:42Z | null | NONE | null | 20260502T210030Z | 2026-05-02T21:00:30Z | liaoandi | 31,417,269 | MDQ6VXNlcjMxNDE3MjY5 | User | false |
openclaw/openclaw | 4,223,924,498 | I_kwDOQb6kR877w_kS | 63,107 | https://github.com/openclaw/openclaw/issues/63107 | https://api.github.com/repos/openclaw/openclaw/issues/63107 | cron: systemEvent on main session silently ignores shell commands in payload text | ## Problem
Cron jobs configured with `sessionTarget: "main"` + `payload.kind: "systemEvent"` + `wakeMode: "next-heartbeat"` fire successfully (`status: ok`, `durationMs: 1–8ms`) but **never actually execute shell commands** contained in the payload text.
For example, a job with `payload.text: "run: python3 /path/to/s... | open | null | false | 1 | [] | [] | 2026-04-08T10:44:01Z | 2026-05-02T18:31:12Z | null | NONE | null | 20260502T210030Z | 2026-05-02T21:00:30Z | liaoandi | 31,417,269 | MDQ6VXNlcjMxNDE3MjY5 | User | false |
openclaw/openclaw | 4,224,011,664 | I_kwDOQb6kR877xU2Q | 63,118 | https://github.com/openclaw/openclaw/issues/63118 | https://api.github.com/repos/openclaw/openclaw/issues/63118 | [Feature]: Missing Session Management Commands in OpenClaw | ### Summary
# [Bug/Feature Request] Missing Session Management Commands in OpenClaw
## 🐛 Defect Type: Critical Functionality Missing
**Severity**: High
**Impact**: All OpenClaw Users
**Priority**: P1 (Urgent)
## 🎯 Expected Behavior
OpenClaw should provide complete session lifecycle management:
## 📈 Busin... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-08T11:01:47Z | 2026-05-03T06:49:31Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | zhuyong74 | 36,598,124 | MDQ6VXNlcjM2NTk4MTI0 | User | false |
openclaw/openclaw | 4,224,381,640 | I_kwDOQb6kR877yvLI | 63,135 | https://github.com/openclaw/openclaw/issues/63135 | https://api.github.com/repos/openclaw/openclaw/issues/63135 | [Bug]: Agents respond they are working on a request but then fail to perform any actions. | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
## Summary
All agents intermittently (now consistently) return a single “working on it” style message, then do no real work.
No active sessions/tasks are created, no files/posts are produced, and no completion... | open | null | false | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-08T12:10:31Z | 2026-05-03T10:19:05Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | Strenky | 138,918,676 | U_kgDOCEe7FA | User | false |
openclaw/openclaw | 4,224,440,751 | I_kwDOQb6kR877y9mv | 63,141 | https://github.com/openclaw/openclaw/issues/63141 | https://api.github.com/repos/openclaw/openclaw/issues/63141 | [Feature] Auto-tune or warn on session.parentForkMaxTokens instead of silent skip | ## Problem
`session.parentForkMaxTokens` defaults to `100000`. When a parent session exceeds this, `session-init` silently skips the fork with a WARN-level log line:
```
[session-init] skipping parent fork (parent too large):
parentKey=agent:main:discord:channel:1234
→ sessionKey=agent:main:discord:channel:5678
... | open | null | false | 1 | [] | [] | 2026-04-08T12:21:19Z | 2026-05-02T23:42:32Z | null | NONE | null | 20260503T000641Z | 2026-05-03T00:06:41Z | Hugo0 | 12,943,235 | MDQ6VXNlcjEyOTQzMjM1 | User | false |
openclaw/openclaw | 4,224,492,793 | I_kwDOQb6kR877zKT5 | 63,145 | https://github.com/openclaw/openclaw/issues/63145 | https://api.github.com/repos/openclaw/openclaw/issues/63145 | [Feature]: openclaw models status --probe --all — per-model health check across all configured models | ### Summary
Add a --all flag to models status --probe that sends a real completion request to every configured model (primary + fallbacks) and reports per-model health, not just per-provider auth.
### Problem to solve
`models status --probe` currently validates **provider authentication**, not individual **model IDs... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-08T12:29:40Z | 2026-05-02T17:31:06Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | ashok-2003 | 123,005,635 | U_kgDOB1Tqww | User | false |
openclaw/openclaw | 4,224,539,777 | I_kwDOQb6kR877zVyB | 63,149 | https://github.com/openclaw/openclaw/issues/63149 | https://api.github.com/repos/openclaw/openclaw/issues/63149 | [Bug]: Gateway CPU stuck at 100% causing service degradation under high load | ## Bug type
Crash (process hangs under load)
## Beta release blocker
No
## Summary
Gateway process continuously occupies a single CPU core at 100% under high load, causing subagent notifications to fail and heartbeat queues to severely delay.
## Steps to reproduce
1. Configure 5 Agents (orchestrator/monitor/coder/te... | open | null | false | 1 | [] | [] | 2026-04-08T12:37:07Z | 2026-05-03T10:19:07Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | chenlong220192 | 25,541,841 | MDQ6VXNlcjI1NTQxODQx | User | false |
openclaw/openclaw | 4,225,194,998 | I_kwDOQb6kR87711v2 | 63,181 | https://github.com/openclaw/openclaw/issues/63181 | https://api.github.com/repos/openclaw/openclaw/issues/63181 | Misleading prompt section can cause incorrect date reasoning even when timezone-only behavior is intended | Title: Misleading `Current Date & Time` prompt section can cause incorrect date reasoning even when timezone-only behavior is intended
## Summary
OpenClaw’s current docs say the system prompt now includes only the user timezone, not the live date/time, in order to preserve prompt cache stability. However, the prompt ... | open | null | false | 1 | [] | [] | 2026-04-08T14:14:37Z | 2026-05-02T20:57:13Z | null | NONE | null | 20260502T210030Z | 2026-05-02T21:00:30Z | matthiasroder | 9,212,056 | MDQ6VXNlcjkyMTIwNTY= | User | false |
openclaw/openclaw | 4,332,637,356 | I_kwDOQb6kR88AAAABAj7MrA | 72,496 | https://github.com/openclaw/openclaw/issues/72496 | https://api.github.com/repos/openclaw/openclaw/issues/72496 | talk.config RPC fails with "unresolved SecretRef" for talk.providers.*.apiKey, breaks Talk Mode discovery | ## Summary
`talk.config` (the Talk-mode discovery RPC that iOS / macOS / Control UI Talk overlays call on startup to learn the configured TTS provider) throws `unresolved SecretRef` when `talk.providers.<id>.apiKey` is configured as a SecretRef object. The credential itself resolves correctly — `talk.realtime.session`... | closed | completed | true | 0 | [
"maintainer"
] | [
"omarshahine"
] | 2026-04-27T01:56:57Z | 2026-04-30T04:33:43Z | 2026-04-27T15:59:14Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | omarshahine | 10,343,873 | MDQ6VXNlcjEwMzQzODcz | User | false |
openclaw/openclaw | 4,332,638,444 | I_kwDOQb6kR88AAAABAj7Q7A | 72,497 | https://github.com/openclaw/openclaw/issues/72497 | https://api.github.com/repos/openclaw/openclaw/issues/72497 | [Bug]: Tool calls invisible in Telegram DM despite verbose: on — user cannot see agent activity | ## Bug Report: Tool calls invisible in Telegram DM despite reasoning: on + verbose: on
### Summary
When `/reasoning on` and `/verbose on` are active in a Telegram DM session, tool calls (e.g. `exec`, `web_fetch`, `web_search`) are **never** rendered as visible messages in Telegram. Only the final assistant text reply ... | closed | completed | true | 1 | [] | [] | 2026-04-27T01:57:21Z | 2026-04-30T04:34:46Z | 2026-04-27T10:50:22Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | Johannes0402 | 207,033,718 | U_kgDODFcVdg | User | false |
openclaw/openclaw | 4,122,367,883 | I_kwDOQb6kR871tleL | 53,032 | https://github.com/openclaw/openclaw/issues/53032 | https://api.github.com/repos/openclaw/openclaw/issues/53032 | [Feature]: Native Conversation Management & Sidebar in Control UI | ### Summary
Add a sidebar and session management to the Control UI for persistent, organized conversations.
### Problem to solve
The current Control UI (Web Dashboard) feels like a single, endless chat. Users cannot easily separate different projects (e.g., "Coding Project A" vs. "Marketing Research B") without usin... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-23T17:10:03Z | 2026-04-29T04:32:59Z | 2026-04-27T02:00:10Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | beuch26 | 142,102,050 | U_kgDOCHhOIg | User | false |
openclaw/openclaw | 4,131,796,440 | I_kwDOQb6kR872RjXY | 54,150 | https://github.com/openclaw/openclaw/issues/54150 | https://api.github.com/repos/openclaw/openclaw/issues/54150 | pdf-native-providers bypasses HTTP_PROXY — direct fetch calls don't use environment proxy | ## Problem
`src/agents/tools/pdf-native-providers.ts` uses bare `fetch()` calls to Anthropic and Google APIs (lines 64 and 144) without calling `ensureGlobalUndiciEnvProxyDispatcher()` first.
In Node.js, the global `fetch` (from undici) does **not** automatically honor `HTTP_PROXY`/`HTTPS_PROXY` environment variables... | closed | completed | true | 1 | [] | [] | 2026-03-25T01:44:40Z | 2026-04-29T04:32:54Z | 2026-04-27T02:00:52Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jinduwang1001-max | 251,753,350 | U_kgDODwFzhg | User | false |
openclaw/openclaw | 4,132,658,175 | I_kwDOQb6kR872U1v_ | 54,254 | https://github.com/openclaw/openclaw/issues/54254 | https://api.github.com/repos/openclaw/openclaw/issues/54254 | Telegram: bot gets stuck in 'typing' during long model inference, /stop unresponsive | ## Problem
When chatting in a Telegram channel/chat with a model that has a long thinking/inference phase (e.g. reasoning-heavy queries), the bot sometimes gets stuck indefinitely showing **typing…** with no response.
During this state:
- **/stop does not work** — the command is sent but the bot does not acknowledge ... | closed | completed | true | 3 | [] | [] | 2026-03-25T05:30:51Z | 2026-04-29T04:32:53Z | 2026-04-27T02:00:57Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | cloudpaw5216 | 267,859,441 | U_kgDOD_c18Q | User | false |
openclaw/openclaw | 4,133,824,228 | I_kwDOQb6kR872ZSbk | 54,381 | https://github.com/openclaw/openclaw/issues/54381 | https://api.github.com/repos/openclaw/openclaw/issues/54381 | [Bug]: Subsystem logger caches file path on first call, logs written to wrong date file after midnight | ### Bug type
Regression (worked before, now fails)
### Summary
Description
When OpenClaw Gateway runs as a long-running service, the subsystem logger ( createSubsystemLogger ) caches the file logger on first call. After midnight, new logs continue to be written to the previous day's log file instead of creating/usi... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-03-25T09:10:54Z | 2026-04-29T04:32:51Z | 2026-04-27T02:01:05Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | alwayshope25 | 222,406,278 | U_kgDODUGmhg | User | false |
openclaw/openclaw | 4,134,309,847 | I_kwDOQb6kR872bI_X | 54,413 | https://github.com/openclaw/openclaw/issues/54413 | https://api.github.com/repos/openclaw/openclaw/issues/54413 | Discord delivery silently drops messages after response termination + retry | ## Summary
When an LLM response is terminated mid-execution (e.g., due to timeout or interruption) and automatically retried, the retry may complete successfully (tool calls execute, text response generated) but the final text message fails to deliver to Discord silently — no error log, no retry, user never receives i... | closed | completed | true | 1 | [] | [] | 2026-03-25T10:29:54Z | 2026-04-29T04:32:51Z | 2026-04-27T02:01:09Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jxdbaby | 249,119,425 | U_kgDODtlCwQ | User | false |
openclaw/openclaw | 4,136,315,644 | I_kwDOQb6kR872iyr8 | 54,535 | https://github.com/openclaw/openclaw/issues/54535 | https://api.github.com/repos/openclaw/openclaw/issues/54535 | [Bug] openclaw browser CLI crashes Chrome 146 via Playwright connectOverCDP (EXC_BREAKPOINT/SIGTRAP) | ## Environment
- **OpenClaw:** 2026.3.24 (updated from 2026.3.23-2 — still persists)
- **Chrome:** 146.0.7680.165
- **macOS:** 26.4 (25E246) arm64
## Description
The `openclaw browser navigate` CLI command (and all browser CLI subcommands) causes Chrome to crash with `EXC_BREAKPOINT (SIGTRAP)` when Playwright's `conne... | closed | completed | true | 3 | [] | [] | 2026-03-25T15:27:39Z | 2026-04-29T04:32:50Z | 2026-04-27T02:01:14Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | malshaalan-ai | 260,099,975 | U_kgDOD4DPhw | User | false |
openclaw/openclaw | 4,138,004,021 | I_kwDOQb6kR872pO41 | 54,658 | https://github.com/openclaw/openclaw/issues/54658 | https://api.github.com/repos/openclaw/openclaw/issues/54658 | Silent message delivery failure: AI response generated but never delivered to Discord | ## Summary
AI-generated responses are occasionally lost silently — the AI completes its turn (tool calls execute successfully), but the final response never reaches Discord. No error is logged, no retry is attempted, and the user sees nothing.
## Incident Timeline (2026-03-25, UTC)
| Time | Event |
|------|-------|
... | closed | completed | true | 1 | [] | [] | 2026-03-25T19:35:31Z | 2026-04-29T04:32:49Z | 2026-04-27T02:01:19Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | i-trytoohard | 193,449,657 | U_kgDOC4fOuQ | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.