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 | 4,145,401,375 | I_kwDOQb6kR873Fc4f | 55,269 | https://github.com/openclaw/openclaw/issues/55269 | https://api.github.com/repos/openclaw/openclaw/issues/55269 | [Feature]: allow command exec to specify shell profile | ### Summary
Add an option to specify profile for shell commands
### Problem to solve
Users could not specify a custom shell profile file for command execution, causing commands to run without user-defined environment variables, aliases, and functions.
### Proposed solution
https://github.com/openclaw/openclaw/pull... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-26T16:17:15Z | 2026-04-29T04:41:17Z | 2026-04-26T08:35:27Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | easyteacher | 23,738,961 | MDQ6VXNlcjIzNzM4OTYx | User | false |
openclaw/openclaw | 4,146,178,469 | I_kwDOQb6kR873Iaml | 55,310 | https://github.com/openclaw/openclaw/issues/55310 | https://api.github.com/repos/openclaw/openclaw/issues/55310 | [Feature]: Add a parameter to docker-setup.sh to skip onboarding. | ### Summary
Add a parameter to docker-setup.sh to skip onboarding.
### Problem to solve
In older versions, the `.openclaw` directory within the container was mounted to the host machine's directory when `docker run` was executed. In newer versions, executing `docker-setup.sh` does not need to re-run `onboard`; the c... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-03-26T17:59:02Z | 2026-04-29T04:41:16Z | 2026-04-26T08:36:56Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ygnzhcool | 51,191,534 | MDQ6VXNlcjUxMTkxNTM0 | User | false |
openclaw/openclaw | 4,147,435,464 | I_kwDOQb6kR873NNfI | 55,378 | https://github.com/openclaw/openclaw/issues/55378 | https://api.github.com/repos/openclaw/openclaw/issues/55378 | [Bug]: Telegram forum topics show Claude Opus 4.6 pre-tool narration / work-note text as visible replies | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
In a **Telegram forum topic** session, `anthropic/claude-opus-4-6` is sending **visible pre-tool narration / step-by-step work notes** into the user chat instead of only sending the final answer.
This is **not just hidden thinking being sh... | closed | not_planned | true | 2 | [] | [] | 2026-03-26T21:27:14Z | 2026-04-29T04:41:15Z | 2026-04-26T08:37:01Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | sene1337 | 261,265,124 | U_kgDOD5KW5A | User | false |
openclaw/openclaw | 4,147,834,136 | I_kwDOQb6kR873Ou0Y | 55,405 | https://github.com/openclaw/openclaw/issues/55405 | https://api.github.com/repos/openclaw/openclaw/issues/55405 | Feature Request: visionOS / Apple Vision Pro node | ## Overview
Building a native visionOS node for OpenClaw as part of LOAM STUDIO's work on spatial AI. Opening this issue to signal intent, surface protocol nuances early, and confirm the contribution path before opening a PR.
## Capabilities Planned
- **WebSocket handshake** — `role: node` with Ed25519 keypair, prot... | closed | not_planned | true | 2 | [] | [] | 2026-03-26T22:51:51Z | 2026-04-29T04:41:14Z | 2026-04-26T08:37:05Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Zera78 | 270,192,173 | U_kgDOEBrOLQ | User | false |
openclaw/openclaw | 4,148,414,610 | I_kwDOQb6kR873Q8iS | 55,451 | https://github.com/openclaw/openclaw/issues/55451 | https://api.github.com/repos/openclaw/openclaw/issues/55451 | Surface raw provider errors in UI/logs when OpenClaw rewrites assistant error text | # Draft GitHub Issue — Surface raw provider errors in UI/logs when OpenClaw rewrites assistant error text
## Summary
When a provider error occurs, OpenClaw can preserve the **raw provider message** in the session transcript while showing a much less useful message in the Control UI / operator-facing surfaces.
In toda... | closed | completed | true | 1 | [] | [] | 2026-03-27T01:23:55Z | 2026-04-29T04:41:14Z | 2026-04-26T08:37:08Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | kevinheneveld | 1,192,102 | MDQ6VXNlcjExOTIxMDI= | User | false |
openclaw/openclaw | 4,148,625,298 | I_kwDOQb6kR873Rv-S | 55,489 | https://github.com/openclaw/openclaw/issues/55489 | https://api.github.com/repos/openclaw/openclaw/issues/55489 | Bug: skillsSnapshot never refreshes after gateway restart (version counter resets to 0) | ## Summary
Long-lived sessions (e.g., the main Telegram session) retain stale skill entries indefinitely — even across gateway restarts. Skills that were renamed or deleted days ago still appear in the `available_skills` prompt, and new/replacement skills are missing.
## Root Cause
The skills snapshot version counte... | closed | not_planned | true | 2 | [] | [] | 2026-03-27T02:20:31Z | 2026-04-29T04:41:13Z | 2026-04-26T08:37:13Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | skadauke | 10,904,538 | MDQ6VXNlcjEwOTA0NTM4 | User | false |
openclaw/openclaw | 4,148,830,673 | I_kwDOQb6kR873SiHR | 55,516 | https://github.com/openclaw/openclaw/issues/55516 | https://api.github.com/repos/openclaw/openclaw/issues/55516 | feat: expose hooks.maxConcurrentRuns config for parallel webhook agent execution | ## Problem
All hook-dispatched agent runs (`action: "agent"` in `hooks.mappings`) serialize on a shared `"nested"` global lane with `maxConcurrent: 1`. Even when each webhook event has a unique `sessionKey` (via transform overrides), only one hook agent can execute at a time — the rest queue behind it.
This is a sign... | closed | not_planned | true | 1 | [] | [] | 2026-03-27T03:04:18Z | 2026-04-29T04:41:12Z | 2026-04-26T08:37:17Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | swiser-nexa | 269,324,650 | U_kgDOEA2Rag | User | false |
openclaw/openclaw | 4,148,874,385 | I_kwDOQb6kR873SsyR | 55,522 | https://github.com/openclaw/openclaw/issues/55522 | https://api.github.com/repos/openclaw/openclaw/issues/55522 | [Feature]: drop the alpha ipa | ### Summary
drop the alpha ipa We have sideloaders like Altstore and many more
### Problem to solve
drop the alpha ipa We have sideloaders like Altstore and many more
### Proposed solution
drop the alpha ipa We have sideloaders like Altstore and many more
### Alternatives considered
drop the alpha ipa We have... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-27T03:13:18Z | 2026-04-29T04:41:10Z | 2026-04-26T08:37:21Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | erencodesguys | 69,877,348 | MDQ6VXNlcjY5ODc3MzQ4 | User | false |
openclaw/openclaw | 4,148,985,152 | I_kwDOQb6kR873TH1A | 55,531 | https://github.com/openclaw/openclaw/issues/55531 | https://api.github.com/repos/openclaw/openclaw/issues/55531 | Feishu WebSocket: No exponential backoff on reconnect, causes rate limit amplification | ## Bug Description
When Feishu WebSocket returns `system busy` (code 1000040345), OpenClaw reconnects at a fixed ~2min interval with no exponential backoff — creating a feedback loop that prolongs the outage.
Error in logs:
```
Cannot read properties of undefined (reading 'PingInterval')
```
## Expected Behavior
1.... | closed | not_planned | true | 1 | [] | [] | 2026-03-27T03:38:12Z | 2026-04-29T04:41:10Z | 2026-04-26T08:37:25Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | sirfengyu | 16,450,114 | MDQ6VXNlcjE2NDUwMTE0 | User | false |
openclaw/openclaw | 4,149,502,928 | I_kwDOQb6kR873VGPQ | 55,572 | https://github.com/openclaw/openclaw/issues/55572 | https://api.github.com/repos/openclaw/openclaw/issues/55572 | [Feature]: Add google-vertex as a supported api type for Vertex AI Express (API key auth) | ## Summary
Support google-vertex as a valid api type in MODEL_APIS to enable Vertex AI Express authentication via API key.
## Problem to solve
The pi-ai library already ships a `google-vertex` provider with `streamGoogleVertex` and `createClientWithApiKey`, which supports Vertex AI Express using a simple API key (no... | closed | not_planned | true | 1 | [] | [] | 2026-03-27T05:22:49Z | 2026-04-29T04:41:09Z | 2026-04-26T08:37:33Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ysnock404 | 82,345,402 | MDQ6VXNlcjgyMzQ1NDAy | User | false |
openclaw/openclaw | 4,149,505,675 | I_kwDOQb6kR873VG6L | 55,574 | https://github.com/openclaw/openclaw/issues/55574 | https://api.github.com/repos/openclaw/openclaw/issues/55574 | feat(acp): support streamTo="thread" to relay ACP progress directly to Discord/slack threads | ## Problem
Currently `sessions_spawn` with `streamTo: "parent"` relays ACP progress as system events to the parent session. This works internally but does not surface real-time progress to users on chat platforms (Discord, Slack, etc.) — the progress events stay invisible in the channel.
## Proposed Solution
Add `"t... | closed | not_planned | true | 1 | [] | [] | 2026-03-27T05:23:21Z | 2026-04-29T04:41:08Z | 2026-04-26T08:37:37Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ricoyudog | 73,219,750 | MDQ6VXNlcjczMjE5NzUw | User | false |
openclaw/openclaw | 4,149,672,296 | I_kwDOQb6kR873Vvlo | 55,590 | https://github.com/openclaw/openclaw/issues/55590 | https://api.github.com/repos/openclaw/openclaw/issues/55590 | [Feature Request] Multi-Model Support for Memory Search Embeddings with Fallback | ## Problem Statement
Currently, `memorySearch` only supports a single embedding model configuration without fallback options:
```json
"memorySearch": {
"provider": "openai",
"model": "BAAI/bge-m3",
"remote": {
"apiKey": "sk-xxx",
"baseUrl": "https://api.siliconflow.cn/v1"
}
}
```
Unlike the main LLM ... | closed | not_planned | true | 2 | [] | [] | 2026-03-27T05:53:27Z | 2026-04-29T04:41:07Z | 2026-04-26T08:37:41Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | mangcai1986 | 268,647,120 | U_kgDOEAM60A | User | false |
openclaw/openclaw | 4,330,433,465 | I_kwDOQb6kR88AAAABAh0ruQ | 72,062 | https://github.com/openclaw/openclaw/issues/72062 | https://api.github.com/repos/openclaw/openclaw/issues/72062 | Feature: agents.list[].tools.deny_wrappers — per-agent denial of bash wrapper paths | # Feature: `agents.list[].tools.deny_wrappers` — per-agent denial of bash wrapper paths
## Summary
`agents.list[].tools.deny` (manual line 33635, codebase `isToolAllowedByPolicyName`) is excellent for filtering MCP tools by `tool.name` (e.g. `github__*`, `supabase__*`). However, **bash wrappers invoked through `tools... | closed | not_planned | true | 1 | [] | [] | 2026-04-26T08:37:44Z | 2026-05-04T04:36:28Z | 2026-05-01T12:35:46Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | David-CKS | 238,664,549 | U_kgDODjm7ZQ | User | false |
openclaw/openclaw | 4,150,483,336 | I_kwDOQb6kR873Y1mI | 55,651 | https://github.com/openclaw/openclaw/issues/55651 | https://api.github.com/repos/openclaw/openclaw/issues/55651 | Discord multi-instance slash commands can be denied by the wrong instance | ## Summary
When the same Discord bot token is connected from multiple OpenClaw instances, slash commands can be handled by the wrong instance and rejected with `This channel is not allowed.` even though the channel is allowed on the intended instance.
## Repro
Environment:
- instance A (`~/.openclaw`) allows channel... | closed | not_planned | true | 1 | [] | [] | 2026-03-27T08:12:27Z | 2026-04-29T04:41:06Z | 2026-04-26T08:37:50Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | seasonhuang | 1,913,836 | MDQ6VXNlcjE5MTM4MzY= | User | false |
openclaw/openclaw | 4,150,697,094 | I_kwDOQb6kR873ZpyG | 55,677 | https://github.com/openclaw/openclaw/issues/55677 | https://api.github.com/repos/openclaw/openclaw/issues/55677 | [Bug]: 3.24版本、mac、操作浏览器无法执行脚本代码 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
使用openclaw沙箱浏览器，出现报错：
"timed out. Restart the OpenClaw gateway ([OpenClaw.app](http://openclaw.app/) menubar, or `openclaw gateway`). Do NOT retry the browser tool — it will keep failing."
### Steps to reproduce
1. opencla... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-03-27T08:45:18Z | 2026-04-29T04:41:06Z | 2026-04-26T08:37:58Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | skyler-developer | 151,726,273 | U_kgDOCQsowQ | User | false |
openclaw/openclaw | 4,151,039,726 | I_kwDOQb6kR873a9bu | 55,704 | https://github.com/openclaw/openclaw/issues/55704 | https://api.github.com/repos/openclaw/openclaw/issues/55704 | [Bug]: Shared hook fallback can surface silent hook output into the main session | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On stock OpenClaw 2026.3.24, shared hook runs can still inject fallback `Hook ...` system events into the main session when the hook lane was intended to stay silent.
### Steps to reproduce
1. Run stock OpenC... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-03-27T09:37:59Z | 2026-04-29T04:41:05Z | 2026-04-26T08:38:02Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | EffortlessSteven | 15,812,269 | MDQ6VXNlcjE1ODEyMjY5 | User | false |
openclaw/openclaw | 4,151,328,048 | I_kwDOQb6kR873cD0w | 55,727 | https://github.com/openclaw/openclaw/issues/55727 | https://api.github.com/repos/openclaw/openclaw/issues/55727 | [Bug]: | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Telegram channel configured correctly but gateway never initiates polling to api.telegram.org.
### Steps to reproduce
1. Configure `channels.telegram` with a valid bot token (`openclaw channels add --channel telegram --tok... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-03-27T10:21:07Z | 2026-04-29T04:41:04Z | 2026-04-26T08:38:11Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | leokuku111-ops | 271,486,625 | U_kgDOEC6OoQ | User | false |
openclaw/openclaw | 4,153,330,349 | I_kwDOQb6kR873jsqt | 55,842 | https://github.com/openclaw/openclaw/issues/55842 | https://api.github.com/repos/openclaw/openclaw/issues/55842 | v2026.3.24: Gateway crash loop — RegExpCompiler OOM on startup (upgrade from v2026.3.13) | ## Environment
- **Platform:** Linux 6.12 (Debian, x86_64)
- **Node:** v22.22.0 (V8 12.4.254.21)
- **Upgrade path:** v2026.3.13 → v2026.3.24
- **Install method:** `npm install -g openclaw`
- **Service management:** systemd (user unit)
## Summary
After upgrading from v2026.3.13 to v2026.3.24 via `npm update -g opencl... | closed | completed | true | 1 | [] | [] | 2026-03-27T14:41:20Z | 2026-04-29T04:41:01Z | 2026-04-26T08:39:48Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | tw3akercc | 7,076,453 | MDQ6VXNlcjcwNzY0NTM= | User | false |
openclaw/openclaw | 4,153,396,812 | I_kwDOQb6kR873j85M | 55,849 | https://github.com/openclaw/openclaw/issues/55849 | https://api.github.com/repos/openclaw/openclaw/issues/55849 | Session transcript headers should include sessionKey for recovery/audit | ## Problem
Session transcript `.jsonl` files do not store the `sessionKey` (e.g. `agent:main:my-project`) in their header. The session header currently only contains:
```json
{"type": "session", "version": 3, "id": "22052c59-...", "timestamp": "...", "cwd": "..."}
```
The only mapping between `sessionKey` and the tr... | closed | not_planned | true | 1 | [] | [] | 2026-03-27T14:49:54Z | 2026-04-29T04:41:00Z | 2026-04-26T08:39:56Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Clausinho | 5,921,903 | MDQ6VXNlcjU5MjE5MDM= | User | false |
openclaw/openclaw | 4,154,934,970 | I_kwDOQb6kR873p0a6 | 55,963 | https://github.com/openclaw/openclaw/issues/55963 | https://api.github.com/repos/openclaw/openclaw/issues/55963 | session-memory hook: UTC date in filename + subagent_ended not handled | ## Bug 1: Memory file uses UTC date instead of local date
**File:** `dist/bundled/session-memory/handler.js` line ~290
**Problem:** `now.toISOString().split("T")[0]` always returns UTC date, causing memory filenames like `2026-03-27-xxxx.md` when local date is actually 2026-03-28.
**Fix:** Replace with `now.toLocale... | closed | not_planned | true | 2 | [] | [] | 2026-03-27T18:14:36Z | 2026-04-29T04:40:59Z | 2026-04-26T08:40:00Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Tiantian225650 | 117,279,838 | U_kgDOBv2MXg | User | false |
openclaw/openclaw | 4,155,431,338 | I_kwDOQb6kR873rtmq | 55,993 | https://github.com/openclaw/openclaw/issues/55993 | https://api.github.com/repos/openclaw/openclaw/issues/55993 | Include sessionKey and metadata in session log file header | ## Problem
The session JSONL log file header (type: "session") only includes `id`, `version`, `timestamp`, and `cwd`. Key metadata like `sessionKey`, `channel`, `agentId`, and session `kind` (main/subagent/cron) are gateway runtime concepts that never get persisted.
This means any tooling reading session logs (dashbo... | closed | not_planned | true | 1 | [] | [] | 2026-03-27T19:34:39Z | 2026-04-29T04:40:59Z | 2026-04-26T08:40:04Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | phoebebeauchamp | 262,689,415 | U_kgDOD6hShw | User | false |
openclaw/openclaw | 4,155,458,405 | I_kwDOQb6kR873r0Nl | 55,994 | https://github.com/openclaw/openclaw/issues/55994 | https://api.github.com/repos/openclaw/openclaw/issues/55994 | message action=send 被误判为 poll（openclaw-weixin 通道，文本/图片均失败） | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
环境信息
- OpenClaw: 2026.3.13 (61d171a)
- OS: macOS 14.3.1 (x64)
- Node: 24.11.1
- Channel: openclaw-weixin
- accountId: 185d35c6c10f-im-bot
- openclaw status: openclaw-weixin ON / OK / accounts 1/1
问题描述
调用 message 工具发送消息时，即使 ... | closed | not_planned | true | 2 | [
"bug",
"regression"
] | [] | 2026-03-27T19:38:46Z | 2026-04-29T04:40:58Z | 2026-04-26T08:40:08Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | 360245033-beep | 271,614,469 | U_kgDOEDCCBQ | User | false |
openclaw/openclaw | 4,156,063,478 | I_kwDOQb6kR873uH72 | 56,045 | https://github.com/openclaw/openclaw/issues/56045 | https://api.github.com/repos/openclaw/openclaw/issues/56045 | Cron job output continues to be delivered after jobs are deleted; persists with zero active jobs | ## Description
Three cron jobs that had been running correctly for several weeks suddenly began firing repeatedly in rapid succession — one job that normally fires once daily was delivered approximately 10 times in a short window.
## Timeline
1. Jobs had been running normally for weeks on `sessionTarget: "main"` wit... | closed | completed | true | 1 | [] | [] | 2026-03-27T21:37:36Z | 2026-04-29T04:40:57Z | 2026-04-26T08:40:12Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | iamvmac | 261,374,807 | U_kgDOD5RDVw | User | false |
openclaw/openclaw | 4,157,036,457 | I_kwDOQb6kR873x1ep | 56,126 | https://github.com/openclaw/openclaw/issues/56126 | https://api.github.com/repos/openclaw/openclaw/issues/56126 | Feature: Batch API support for non-urgent cron jobs (50% cost reduction) | ## Summary
Allow cron jobs with `agentTurn` payloads to optionally route through Anthropic's Batch API instead of the standard Messages API. The Batch API offers a 50% discount on both input and output tokens with a 24-hour completion window — perfect for scheduled, non-interactive tasks.
## Use Case
I run several r... | closed | not_planned | true | 1 | [] | [] | 2026-03-28T02:34:43Z | 2026-04-29T04:40:56Z | 2026-04-26T08:40:24Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jaysturm | 1,745,678 | MDQ6VXNlcjE3NDU2Nzg= | User | false |
openclaw/openclaw | 4,330,266,987 | I_kwDOQb6kR88AAAABAhqhaw | 72,019 | https://github.com/openclaw/openclaw/issues/72019 | https://api.github.com/repos/openclaw/openclaw/issues/72019 | [Bug]: session_spawn duplicates subagent task prompt in both system and user roles, doubling token cost | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When an orchestration agent spawns a subagent via `session_spawn`, the full subagent task prompt is sent **twice** in the API request — once embedded at the end of the `system` prompt and again as the entire `u... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-26T07:02:28Z | 2026-04-30T04:34:13Z | 2026-04-27T12:11:07Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | Wizongod | 9,720,020 | MDQ6VXNlcjk3MjAwMjA= | User | false |
openclaw/openclaw | 4,330,423,481 | I_kwDOQb6kR88AAAABAh0EuQ | 72,061 | https://github.com/openclaw/openclaw/issues/72061 | https://api.github.com/repos/openclaw/openclaw/issues/72061 | [Bug]: Skill toggle triggers full gateway restart with plugins.entries.*.config in reload diff | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On 2026.4.24, toggling a single skill in WebUI causes a full gateway process restart. The reload evaluator's diff includes `plugins.entries.{browser,xai,openai,openrouter}.config` even though the user only chan... | closed | not_planned | true | 2 | [
"bug"
] | [] | 2026-04-26T08:32:07Z | 2026-04-29T04:35:03Z | 2026-04-26T16:17:27Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | lendover89 | 247,469,929 | U_kgDODsAXaQ | User | false |
openclaw/openclaw | 4,157,633,317 | I_kwDOQb6kR8730HMl | 56,188 | https://github.com/openclaw/openclaw/issues/56188 | https://api.github.com/repos/openclaw/openclaw/issues/56188 | /skill command not transformed when no model aliases are configured | ## Describe the bug
When a user sends `/skill <name> <input>`, the message body should be transformed into the skill-injection format before being passed to the agent:
```
Use the "<name>" skill for this request.
User input:
<input>
```
However, this transformation is silently skipped when no model aliases are conf... | closed | not_planned | true | 1 | [] | [] | 2026-03-28T04:59:03Z | 2026-04-29T04:40:56Z | 2026-04-26T08:40:36Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | BrilliantWang | 120,538,413 | U_kgDOBy9FLQ | User | false |
openclaw/openclaw | 4,157,870,582 | I_kwDOQb6kR8731BH2 | 56,205 | https://github.com/openclaw/openclaw/issues/56205 | https://api.github.com/repos/openclaw/openclaw/issues/56205 | Write failed 错误提示频繁出现 - 文件实际写入成功但返回错误 | ## 问题描述
在使用 OpenClaw 进行文件写入操作时，频繁出现 **"Write failed"** 错误提示，但实际文件已经成功写入。
## 复现场景
- 使用 `write` 工具创建或编辑文件
- 特别是在编辑 `MEMORY.md` 等大型文件时
- 错误出现时间不固定，偶发性
## 根因分析
经过深入调查，发现问题可能由以下原因导致：
1. **工具响应超时或显示延迟** - 文件实际写入成功，但工具响应返回错误提示
2. **编码/行尾不统一** - UTF-8 BOM vs without BOM，CRLF vs LF 导致匹配失败
3. **验证机制缺失** - 写入后没有立即验证文件是否真正存在... | closed | not_planned | true | 1 | [] | [] | 2026-03-28T05:43:41Z | 2026-04-29T04:40:54Z | 2026-04-26T08:40:44Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | hnfjf4289-sketch | 267,009,132 | U_kgDOD-o8bA | User | false |
openclaw/openclaw | 4,330,438,216 | I_kwDOQb6kR88AAAABAh0-SA | 72,064 | https://github.com/openclaw/openclaw/issues/72064 | https://api.github.com/repos/openclaw/openclaw/issues/72064 | Discord ACP sessions fail with ACP_TURN_FAILED under default acpx permissions while direct acpx succeeds | ## Summary
We hit a reproducible Discord -> OpenClaw -> ACPX/Codex failure where Discord-launched ACP coding sessions failed almost immediately with `AcpRuntimeError: Internal error` / `code=ACP_TURN_FAILED`, while the underlying `acpx` + Codex adapter worked directly on the same host.
The practical fix was to overri... | closed | not_planned | true | 2 | [] | [] | 2026-04-26T08:40:20Z | 2026-04-29T04:37:09Z | 2026-04-26T10:48:29Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | felirami | 6,752,178 | MDQ6VXNlcjY3NTIxNzg= | User | false |
openclaw/openclaw | 4,330,378,320 | I_kwDOQb6kR88AAAABAhxUUA | 72,049 | https://github.com/openclaw/openclaw/issues/72049 | https://api.github.com/repos/openclaw/openclaw/issues/72049 | [Feature]: Feishu TTS Audio Replies Send File Path Instead of Playable Audio | ### Summary
When the Feishu channel receives a voice message and OpenClaw's built-in TTS auto mechanism generates an audio reply, the file path (e.g., `C:\Users\...\media\outbound\voice-xxx.mp3`) is sent as plain text instead of using Feishu's `msg_type: "audio"` API to deliver a directly-playable audio message.
###... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-26T08:06:57Z | 2026-04-29T04:40:50Z | 2026-04-26T08:41:46Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | osirisjaret | 4,230,888 | MDQ6VXNlcjQyMzA4ODg= | User | false |
openclaw/openclaw | 4,330,340,735 | I_kwDOQb6kR88AAAABAhvBfw | 72,035 | https://github.com/openclaw/openclaw/issues/72035 | https://api.github.com/repos/openclaw/openclaw/issues/72035 | tasks audit emits large numbers of false inconsistent_timestamps warnings when startedAt < createdAt by a few ms | ## Summary
`openclaw tasks audit` reports large numbers of `inconsistent_timestamps` warnings where `startedAt` is earlier than `createdAt`, even though the tasks appear to have completed normally. On this host the warning count was 954, with 953 warnings of this form and no corresponding operational failure.
## Versi... | closed | completed | true | 1 | [] | [] | 2026-04-26T07:44:38Z | 2026-04-29T04:40:51Z | 2026-04-26T08:41:46Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | AVeryCrudeMan | 72,014,797 | MDQ6VXNlcjcyMDE0Nzk3 | User | false |
openclaw/openclaw | 4,330,198,194 | I_kwDOQb6kR88AAAABAhmUsg | 72,006 | https://github.com/openclaw/openclaw/issues/72006 | https://api.github.com/repos/openclaw/openclaw/issues/72006 | [Bug]: Device pairing scope escalation via repeated re-request before operator approval | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 7.9 / 10.0 | 8.4 / 10.0 |
| **Severity** | High | High |
| **Vector** | CVSS:3.1/AV:L/AC:L/PR:L/UI:R/S:C/C:H/I:H/A:N | CVSS:4.0/AV:L/AC:L/AT:N/PR:L/UI:P/VC:H/VI:H/VA:N/SC:H/SI:H/SA:N |
| **Calculator** | [CVSS v... | closed | completed | true | 1 | [] | [] | 2026-04-26T06:24:14Z | 2026-04-29T04:40:52Z | 2026-04-26T08:41:46Z | CONTRIBUTOR | null | 20260429T060025Z | 2026-04-29T06:00:25Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 4,157,176,040 | I_kwDOQb6kR873yXjo | 56,142 | https://github.com/openclaw/openclaw/issues/56142 | https://api.github.com/repos/openclaw/openclaw/issues/56142 | Slack block_actions not dispatched as inbound messages — buttons never wake session | ## Summary
Slack button clicks (`block_actions`) are handled via `enqueueSystemEvent()` but never wake the agent session. The event sits in the queue until the next human message arrives, making interactive buttons effectively dead.
This was previously reported in #21221 (auto-closed by stale bot — the bug is still p... | closed | not_planned | true | 1 | [] | [] | 2026-03-28T03:23:57Z | 2026-04-29T04:40:50Z | 2026-04-26T08:42:51Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ChickenEggRoll | 181,599,005 | U_kgDOCtL7HQ | User | false |
openclaw/openclaw | 4,158,701,343 | I_kwDOQb6kR8734L8f | 56,290 | https://github.com/openclaw/openclaw/issues/56290 | https://api.github.com/repos/openclaw/openclaw/issues/56290 | Bug: Recursive loadConfig → console.warn → getLogger loop causes high CPU in v2026.3.24 | ## Bug: Recursive `loadConfig → console.warn → getLogger → loadConfig` loop causes high CPU
### Version
OpenClaw 2026.3.24 (cff6dc9) on macOS (Darwin arm64, Node v24.14.0 / v25.8.1)
### Description
After upgrading from v2026.3.13 to v2026.3.24, the gateway process enters a recursive loop that causes sustained high CP... | closed | completed | true | 2 | [] | [] | 2026-03-28T08:49:16Z | 2026-04-29T04:40:49Z | 2026-04-26T08:42:54Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | silvesterxm | 268,212,401 | U_kgDOD_yYsQ | User | false |
openclaw/openclaw | 4,158,728,965 | I_kwDOQb6kR8734SsF | 56,293 | https://github.com/openclaw/openclaw/issues/56293 | https://api.github.com/repos/openclaw/openclaw/issues/56293 | [Feature]: [TUI] Add optional message timestamps to chat view | ### Summary
When using `openclaw tui` over SSH (e.g., via PuTTY), there is no way to see the timestamp of each individual message. This makes it hard to track when conversations happened, especially in longer sessions.
### Problem to solve
Add an optional timestamp prefix to each message in the TUI chat view, displ... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-28T08:55:19Z | 2026-04-29T04:40:48Z | 2026-04-26T08:42:59Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | MikeAizman | 104,719,361 | U_kgDOBj3kAQ | User | false |
openclaw/openclaw | 4,159,600,521 | I_kwDOQb6kR8737neJ | 56,375 | https://github.com/openclaw/openclaw/issues/56375 | https://api.github.com/repos/openclaw/openclaw/issues/56375 | [Feature]: Web UI interface file upload | ### Summary
Web UI interface file upload.
### Problem to solve
Unable to upload after selecting the file. No effect. Is the file upload button invalid?
### Proposed solution
Is this button invalid
### Alternatives conside... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-03-28T12:04:45Z | 2026-04-29T04:40:47Z | 2026-04-26T08:43:12Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | triggergun | 103,357,879 | U_kgDOBikdtw | User | false |
openclaw/openclaw | 4,160,130,823 | I_kwDOQb6kR8739o8H | 56,427 | https://github.com/openclaw/openclaw/issues/56427 | https://api.github.com/repos/openclaw/openclaw/issues/56427 | memory index --force fails with 'fetch failed' on large workspace (1400+ .md files) | ## Bug Report
### Description
`openclaw memory index --force` consistently fails with `Memory index failed (main): fetch failed` on a workspace with 1400+ `.md` files. After 2 failures, batch mode is permanently disabled (`Batch: disabled (failures 0/2)`), making re-indexing impossible without restarting gateway.
###... | closed | not_planned | true | 2 | [] | [] | 2026-03-28T14:12:42Z | 2026-04-29T04:40:46Z | 2026-04-26T08:43:17Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | runrelay | 260,778,404 | U_kgDOD4sppA | User | false |
openclaw/openclaw | 4,160,357,582 | I_kwDOQb6kR873-gTO | 56,462 | https://github.com/openclaw/openclaw/issues/56462 | https://api.github.com/repos/openclaw/openclaw/issues/56462 | [Bug]: Windows exec tool garbled Chinese characters (confirmed in 2026.3.24) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The exec tool on Windows produces garbled Chinese characters (乱码). This is the same issue as #50519, but I'm confirming it still exists in version 2026.3.24.
### Steps to reproduce
1. Use OpenClaw exec tool o... | closed | not_planned | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-03-28T15:19:45Z | 2026-04-29T04:40:45Z | 2026-04-26T08:43:21Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | knightplat-blip | 271,784,033 | U_kgDOEDMYYQ | User | false |
openclaw/openclaw | 4,160,878,253 | I_kwDOQb6kR874Afat | 56,521 | https://github.com/openclaw/openclaw/issues/56521 | https://api.github.com/repos/openclaw/openclaw/issues/56521 | Feature: Route cron failure alerts as agent-triggering events to main session | ## Problem
When a cron job fails, the `failureAlert` config sends a Telegram notification to the DM chat. However, this notification appears as a **system message** — it does NOT trigger a new agent turn in the main session.
This means:
1. The human sees the error in chat
2. The agent (main session) is unaware of it
... | closed | not_planned | true | 1 | [] | [] | 2026-03-28T17:06:37Z | 2026-04-29T04:40:45Z | 2026-04-26T08:43:40Z | CONTRIBUTOR | null | 20260429T060025Z | 2026-04-29T06:00:25Z | andyk-ms | 91,510,251 | U_kgDOBXRV6w | User | false |
openclaw/openclaw | 4,160,919,500 | I_kwDOQb6kR874ApfM | 56,531 | https://github.com/openclaw/openclaw/issues/56531 | https://api.github.com/repos/openclaw/openclaw/issues/56531 | [Bug]: xAI model normalization silently rewrites grok-4-1-fast-reasoning to grok-4-1-fast | ## Summary
OpenClaw currently rewrites the explicit xAI model id `grok-4-1-fast-reasoning` to `grok-4-1-fast` during model normalization.
This makes a user-configured default model resolve to a different effective model than the one declared in config, and it shows up immediately in `openclaw models status`.
## Repr... | closed | not_planned | true | 1 | [] | [] | 2026-03-28T17:15:11Z | 2026-04-29T04:40:44Z | 2026-04-26T08:43:45Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | whoisheartbreak | 165,145,263 | U_kgDOCdfqrw | User | false |
openclaw/openclaw | 4,330,441,166 | I_kwDOQb6kR88AAAABAh1Jzg | 72,065 | https://github.com/openclaw/openclaw/issues/72065 | https://api.github.com/repos/openclaw/openclaw/issues/72065 | v2026.4.24 bundled channel runtime deps resolve to openclaw-unknown and stall Telegram/WhatsApp startup | ## Summary
After upgrading a Docker-based gateway from `2026.4.20` to `2026.4.24`, Telegram and WhatsApp channels do not start promptly after gateway readiness. The gateway is HTTP healthy and reports sidecar startup, but channel provider handoff is delayed for ~19 minutes and appears to happen only after the health m... | closed | completed | true | 4 | [] | [] | 2026-04-26T08:42:02Z | 2026-04-29T04:36:30Z | 2026-04-26T10:48:33Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | balaji1968-kingler | 268,807,278 | U_kgDOEAWsbg | User | false |
openclaw/openclaw | 4,330,333,223 | I_kwDOQb6kR88AAAABAhukJw | 72,031 | https://github.com/openclaw/openclaw/issues/72031 | https://api.github.com/repos/openclaw/openclaw/issues/72031 | [Bug]: `image` tool fails for Bedrock with `auth mode: aws-sdk` — `requireApiKey` throws even when AWS SDK creds are available | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The `image` tool (media-understanding image-description path) always requires an explicit `apiKey` when resolving provider auth. For `amazon-bedrock` using IAM role / instance profile / `AWS_PROFILE` credentials (i.e. `Resol... | open | null | false | 2 | [
"bug"
] | [] | 2026-04-26T07:40:01Z | 2026-05-03T07:13:42Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | GunnarHelliesen | 22,202,631 | MDQ6VXNlcjIyMjAyNjMx | User | false |
openclaw/openclaw | 4,162,109,089 | I_kwDOQb6kR874FL6h | 56,689 | https://github.com/openclaw/openclaw/issues/56689 | https://api.github.com/repos/openclaw/openclaw/issues/56689 | [Feature]: LaTeX Math Formula Rendering Support in OpenClaw Dashboard Chat | ### Summary
# Feature Request: LaTeX Math Formula Rendering Support in OpenClaw Dashboard Chat
## Feature Description
I would like to request support for **LaTeX math formula rendering** in the OpenClaw Dashboard chat interface.
## Use Case
When using OpenClaw for mathematics, physics, engineering, or algorithm-rela... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-29T01:01:42Z | 2026-04-29T04:40:42Z | 2026-04-26T08:45:35Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ScriptForge666 | 221,077,483 | U_kgDODS1f6w | User | false |
openclaw/openclaw | 4,162,406,220 | I_kwDOQb6kR874GUdM | 56,777 | https://github.com/openclaw/openclaw/issues/56777 | https://api.github.com/repos/openclaw/openclaw/issues/56777 | Slash commands bypass topic-level groupPolicy: disabled in forum groups | ## Summary
In a Telegram forum group, slash commands (e.g. `/new`, `/status`) are not subject to per-topic `groupPolicy` overrides. When a group is configured with `groupPolicy: "disabled"` and only specific topics are whitelisted via per-topic `groupPolicy: "allowlist"`, the bot still responds to slash commands in no... | closed | not_planned | true | 1 | [] | [] | 2026-03-29T03:55:05Z | 2026-04-29T04:40:41Z | 2026-04-26T08:46:03Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Git-on-my-level | 9,387,252 | MDQ6VXNlcjkzODcyNTI= | User | false |
openclaw/openclaw | 4,162,576,437 | I_kwDOQb6kR874G-A1 | 56,805 | https://github.com/openclaw/openclaw/issues/56805 | https://api.github.com/repos/openclaw/openclaw/issues/56805 | Expose knownLongFlags in safeBinProfiles config schema | ## Summary
The `safeBinProfiles` config schema does not expose the `knownLongFlags` field, even though it is used internally by `validateSafeBinArgv` for GNU long-option abbreviation resolution. Attempting to set `knownLongFlags` in a profile causes a Zod `.strict()` validation crash: **"Unrecognized key(s) in object:... | closed | not_planned | true | 1 | [] | [] | 2026-03-29T04:50:09Z | 2026-04-29T04:40:41Z | 2026-04-26T08:46:12Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | spencer2211 | 28,957,500 | MDQ6VXNlcjI4OTU3NTAw | User | false |
openclaw/openclaw | 4,163,562,783 | I_kwDOQb6kR874Ku0f | 56,919 | https://github.com/openclaw/openclaw/issues/56919 | https://api.github.com/repos/openclaw/openclaw/issues/56919 | Skills: repeated 'Skipping skill path that resolves outside its configured root' warnings on 2026.3.28 | # OpenClaw 2026.3.28 skill path warning spam
Date: 2026-03-29
Status: open / deferred
Severity: low functional impact, high log noise
## Symptom
On startup and during skill checks, OpenClaw emits repeated warnings:
`[skills] Skipping skill path that resolves outside its configured root.`
Observed frequency:
- 18 wa... | closed | completed | true | 1 | [] | [] | 2026-03-29T09:36:45Z | 2026-04-29T04:40:40Z | 2026-04-26T08:46:22Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jakobmygind | 8,362,263 | MDQ6VXNlcjgzNjIyNjM= | User | false |
openclaw/openclaw | 4,163,724,690 | I_kwDOQb6kR874LWWS | 56,942 | https://github.com/openclaw/openclaw/issues/56942 | https://api.github.com/repos/openclaw/openclaw/issues/56942 | [Bug]: | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading from 2026.3.23-2 to 2026.3.28, Gateway start failed,
Error: Config validation failed: channels.feishu.accounts.main: invalid config: must NOT have additional properties
### Steps to reproduce
1. use opencl... | closed | not_planned | true | 4 | [
"bug",
"regression"
] | [] | 2026-03-29T10:34:01Z | 2026-04-29T04:40:39Z | 2026-04-26T08:46:31Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | x2x4com | 52,311 | MDQ6VXNlcjUyMzEx | User | false |
openclaw/openclaw | 4,163,803,672 | I_kwDOQb6kR874LpoY | 56,951 | https://github.com/openclaw/openclaw/issues/56951 | https://api.github.com/repos/openclaw/openclaw/issues/56951 | Slack inline buttons should be enabled by default | ## Original Request
> inline button可以file一下、就说默认应该开启毕竟没啥害处。然后直接fix（应该一两行）然后pr发出去就好了
> **English Translation:** File an issue for inline buttons — they should be enabled by default since there is no downside. Then fix it directly (should be 1-2 lines) and send a PR.
## Agent's Two Cents (could be wrong)
> Everything... | closed | completed | true | 1 | [] | [] | 2026-03-29T10:58:46Z | 2026-04-29T04:40:38Z | 2026-04-26T08:46:35Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | xinbenlv | 640,325 | MDQ6VXNlcjY0MDMyNQ== | User | false |
openclaw/openclaw | 4,164,227,526 | I_kwDOQb6kR874NRHG | 57,054 | https://github.com/openclaw/openclaw/issues/57054 | https://api.github.com/repos/openclaw/openclaw/issues/57054 | Request: collaborator access for lsw2408 | ## Request
Requesting write (push) access to the repository for `@lsw2408`.
## Context
Currently contributing to local tooling and configuration for the project. Recent work includes:
- `.gitignore` updates for local development packages
- Cron job configuration fixes (Telegram delivery)
## Current Permissions
- `... | closed | not_planned | true | 1 | [] | [] | 2026-03-29T13:10:17Z | 2026-04-29T04:40:37Z | 2026-04-26T08:46:44Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | lsw2408 | 192,283,228 | U_kgDOC3YCXA | User | false |
openclaw/openclaw | 4,164,902,113 | I_kwDOQb6kR874P1zh | 57,141 | https://github.com/openclaw/openclaw/issues/57141 | https://api.github.com/repos/openclaw/openclaw/issues/57141 | [Bug]: Telegram DM topic ACP bind fails on next message with missing ACP metadata | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On `v2026.3.28-beta.1`, binding a fresh Telegram DM topic to a Codex ACP session with `/acp spawn codex --bind here` can fail on the next inbound message with `ACP_SESSION_INIT_FAILED` because the routed `:thre... | closed | completed | true | 2 | [] | [] | 2026-03-29T16:07:17Z | 2026-04-29T04:40:36Z | 2026-04-26T08:46:58Z | CONTRIBUTOR | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Jackten | 2,895,479 | MDQ6VXNlcjI4OTU0Nzk= | User | false |
openclaw/openclaw | 4,330,451,564 | I_kwDOQb6kR88AAAABAh1ybA | 72,071 | https://github.com/openclaw/openclaw/issues/72071 | https://api.github.com/repos/openclaw/openclaw/issues/72071 | Tool calls and chain-of-thought leaked to user in Telegram output (kimi-k2.6:cloud) | ## Bug Description
When using `ollama/kimi-k2.6:cloud` as the model, assistant responses leak raw tool call blocks and chain-of-thought text to the user in Telegram.
## What Happened
During a conversation about researching a gymnastics club for a child, the assistant (powered by kimi-k2.6:cloud) produced a response ... | closed | completed | true | 1 | [] | [] | 2026-04-26T08:48:03Z | 2026-04-29T04:35:02Z | 2026-04-26T16:17:31Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | davidvv | 11,057,344 | MDQ6VXNlcjExMDU3MzQ0 | User | false |
openclaw/openclaw | 4,166,685,355 | I_kwDOQb6kR874WpKr | 57,424 | https://github.com/openclaw/openclaw/issues/57424 | https://api.github.com/repos/openclaw/openclaw/issues/57424 | [Feature]: 设备列表看不出具体是哪台设备，多个同类型条目无法区分 | ### Summary
- 问题：设备列表看不出具体是哪台设备，多个同类型条目无法区分
- 建议：增加设备名称、浏览器、IP、OS 版本、最后活跃时间
- 影响命令：devices list / devices remove
### Problem to solve
RT
### Proposed solution
RT
### Alternatives considered
_No response_
### Impact
RT
### Evidence/examples
_No response_
### Additional information
_No response_ | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-30T02:49:42Z | 2026-04-29T04:40:35Z | 2026-04-26T08:48:40Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | davedavehong | 10,491,999 | MDQ6VXNlcjEwNDkxOTk5 | User | false |
openclaw/openclaw | 4,330,452,767 | I_kwDOQb6kR88AAAABAh13Hw | 72,072 | https://github.com/openclaw/openclaw/issues/72072 | https://api.github.com/repos/openclaw/openclaw/issues/72072 | RFC: RuntimePlan finalization and embedded runner structural cleanup | ## Summary
This RFC tracks the post-#71722 finalization roadmap for the contract-first Pi/Codex runtime work.
The original RFC #71004 is implemented and closed. The core RuntimePlan boundary is now in `main`: OpenClaw-owned runtime policy is represented in `AgentRuntimePlan`, built before attempts, consumed by Pi/Cod... | open | null | false | 2 | [] | [] | 2026-04-26T08:48:44Z | 2026-05-03T02:15:18Z | null | CONTRIBUTOR | null | 20260503T030030Z | 2026-05-03T03:00:30Z | 100yenadmin | 239,388,517 | U_kgDODkTHZQ | User | false |
openclaw/openclaw | 4,167,158,474 | I_kwDOQb6kR874YcrK | 57,466 | https://github.com/openclaw/openclaw/issues/57466 | https://api.github.com/repos/openclaw/openclaw/issues/57466 | [Bug]: device id resets after browser session restarts | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Hi community, I had setup a cloudflare tunnel and setup a domain and connected local openclaw instance. I then added external device as approved device list through Openclaw devices commands. Earlier I had no problem and thi... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-03-30T04:19:42Z | 2026-04-29T04:40:34Z | 2026-04-26T08:48:44Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | yogesh1981-afi | 158,290,547 | U_kgDOCW9Scw | User | false |
openclaw/openclaw | 4,167,681,552 | I_kwDOQb6kR874acYQ | 57,509 | https://github.com/openclaw/openclaw/issues/57509 | https://api.github.com/repos/openclaw/openclaw/issues/57509 | [Bug]: Telegram helper-agent slash:<user> session keys are incorrectly mapped to group:<group_id>:topic:1 sessions | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
**Title**
Telegram helper-agent session key mapping bug: `slash:<user>` keys are incorrectly projected to `group:<group_id>:topic:1`, causing stale model-switch artifacts
**Body**
## Summary
I found what appears to be a s... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-03-30T05:58:11Z | 2026-04-29T04:29:13Z | 2026-04-26T08:48:48Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | cuijinlin | 33,531,932 | MDQ6VXNlcjMzNTMxOTMy | User | false |
openclaw/openclaw | 4,167,982,911 | I_kwDOQb6kR874bl8_ | 57,536 | https://github.com/openclaw/openclaw/issues/57536 | https://api.github.com/repos/openclaw/openclaw/issues/57536 | [Bug]: feishu:tenant_access_token获取失败，返回400错误 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
升级openclaw到2026.3.28（f9b1079）后，飞书插件获取tenant_access_token失败，返回400错误，鉴权失败。
### Steps to reproduce
正常启动openclaw。
### Expected behavior
飞书插件正常工作
### Actual behavior
飞书插件鉴权失败，飞书消息客户端不能正常对话。
### OpenClaw version
OpenClaw 2... | closed | not_planned | true | 2 | [
"bug",
"regression"
] | [] | 2026-03-30T06:49:27Z | 2026-04-29T04:40:33Z | 2026-04-26T08:48:53Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | david6096 | 26,000,707 | MDQ6VXNlcjI2MDAwNzA3 | User | false |
openclaw/openclaw | 4,168,613,117 | I_kwDOQb6kR874d_z9 | 57,586 | https://github.com/openclaw/openclaw/issues/57586 | https://api.github.com/repos/openclaw/openclaw/issues/57586 | Node system.run fails: Remote app (2026.3.28) does not advertise system.run.prepare | ## Bug Description
`nodes.run` fails with:
```
node command not allowed: the node (platform: macOS 26.3.0) does not support "system.run.prepare"
```
## Environment
- **Gateway**: OpenClaw 2026.3.28 (f9b1079), macOS, Mac Mini M4
- **Remote app**: OpenClaw Remote 2026.3.28 (2026032890), macOS 26.3.0, MacBook Pro (Mac... | closed | not_planned | true | 1 | [] | [] | 2026-03-30T08:21:02Z | 2026-04-29T04:40:32Z | 2026-04-26T08:49:07Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | artdaal | 35,874,009 | MDQ6VXNlcjM1ODc0MDA5 | User | false |
openclaw/openclaw | 4,330,453,523 | I_kwDOQb6kR88AAAABAh16Ew | 72,073 | https://github.com/openclaw/openclaw/issues/72073 | https://api.github.com/repos/openclaw/openclaw/issues/72073 | WebChat / Control UI can duplicate assistant replies in new sessions | ## Summary
In webchat / OpenClaw Control UI, starting a new chat can cause assistant messages to appear twice. The duplication also affected subsequent normal replies in the same session.
Based on logs + session history, this does **not** look like the model generated two different replies. It looks more like the **sa... | closed | completed | true | 2 | [] | [] | 2026-04-26T08:49:11Z | 2026-04-29T04:37:08Z | 2026-04-26T10:48:37Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | xiao398008 | 211,473,177 | U_kgDODJrTGQ | User | false |
openclaw/openclaw | 4,169,608,098 | I_kwDOQb6kR874hyui | 57,653 | https://github.com/openclaw/openclaw/issues/57653 | https://api.github.com/repos/openclaw/openclaw/issues/57653 | [Feature]: add call-level model hooks for accurate tracing and debugging | ### Summary
OpenClaw already provides call-level tool hooks such as `before_tool_call` / `after_tool_call`, but model-side hooks appear to be higher-level. In practice, `llm_input` / `llm_output` seem to represent an overall prompt attempt rather than each real provider model call. For tool-using turns, this makes it ... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-30T10:37:40Z | 2026-04-29T04:40:31Z | 2026-04-26T08:49:15Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | totoyang | 34,332,614 | MDQ6VXNlcjM0MzMyNjE0 | User | false |
openclaw/openclaw | 4,169,609,758 | I_kwDOQb6kR874hzIe | 57,654 | https://github.com/openclaw/openclaw/issues/57654 | https://api.github.com/repos/openclaw/openclaw/issues/57654 | [Bug]: Unexpected event order, got message_start before receiving "message_stop" | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Unexpected event order, got message_start before receiving "message_stop" 大模型为Kimi-code ，接入lark，长时间运行之后一直报这个错
### Steps to reproduce
Unexpected event order, got message_start before receiving "message_stop" 大模型为Kimi-co... | closed | completed | true | 9 | [
"bug"
] | [] | 2026-03-30T10:37:59Z | 2026-04-29T04:40:31Z | 2026-04-26T08:49:19Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | huojinibuhao | 77,958,337 | MDQ6VXNlcjc3OTU4MzM3 | User | false |
openclaw/openclaw | 4,170,239,685 | I_kwDOQb6kR874kM7F | 57,682 | https://github.com/openclaw/openclaw/issues/57682 | https://api.github.com/repos/openclaw/openclaw/issues/57682 | [Bug][Windows] openclaw update spawns visible and hanging CMD window during gateway restart | When running openclaw update on Windows, a visible CMD window with the title `findstr /R /C:":18789 .*LISTENING"` appears and often fails to close automatically, hanging indefinitely. This occurs during the service restart phase of the update process.
## Root Cause Analysis
The issue stems from the temporary batch sc... | closed | completed | true | 4 | [] | [] | 2026-03-30T12:13:07Z | 2026-04-29T04:40:30Z | 2026-04-26T08:49:28Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | zhangxue1985122219 | 54,266,548 | MDQ6VXNlcjU0MjY2NTQ4 | User | false |
openclaw/openclaw | 4,170,519,424 | I_kwDOQb6kR874lROA | 57,699 | https://github.com/openclaw/openclaw/issues/57699 | https://api.github.com/repos/openclaw/openclaw/issues/57699 | Memory leak and excessive memory growth in long-running Gateway process | # Memory leak and excessive memory growth in long-running Gateway process
## Environment
- **OpenClaw version**: 2026.3.28 (f9b1079)
- **Node.js version**: v22.22.0
- **OS**: macOS (Darwin 25.2.0, arm64)
- **Runtime**: Long-running Gateway process (Telegram channel)
## Problem Description
The OpenClaw Gateway proces... | closed | not_planned | true | 1 | [] | [] | 2026-03-30T12:55:01Z | 2026-04-29T04:40:29Z | 2026-04-26T08:49:32Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Fionaaa0124 | 257,260,192 | U_kgDOD1V6oA | User | false |
openclaw/openclaw | 4,171,008,759 | I_kwDOQb6kR874nIr3 | 57,730 | https://github.com/openclaw/openclaw/issues/57730 | https://api.github.com/repos/openclaw/openclaw/issues/57730 | feat(delivery): surface deliveryStatus in --json output for openclaw agent --deliver | ## Problem
`openclaw agent --json --deliver` currently has no machine-readable indication of whether delivery succeeded, failed, or was skipped. The JSON envelope contains `payloads` and `meta` but no delivery status.
This means automation consuming `--json` output cannot detect silent delivery failures — the exact s... | closed | not_planned | true | 1 | [] | [] | 2026-03-30T14:03:29Z | 2026-04-29T04:40:27Z | 2026-04-26T08:49:46Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Kaspre | 36,520,309 | MDQ6VXNlcjM2NTIwMzA5 | User | false |
openclaw/openclaw | 4,330,455,594 | I_kwDOQb6kR88AAAABAh2CKg | 72,074 | https://github.com/openclaw/openclaw/issues/72074 | https://api.github.com/repos/openclaw/openclaw/issues/72074 | subagent session label: add {agentId} auto-prefix template | # Feature Request: subagent session label auto-prefix from agent name
## Problem
When spawning subagents via `sessions_spawn`, the `label` parameter is user-provided text with no convention or auto-prefix mechanism. This makes it difficult to identify which agent a session belongs to when looking at the sessions list... | closed | not_planned | true | 1 | [] | [] | 2026-04-26T08:50:23Z | 2026-04-29T04:34:25Z | 2026-04-26T19:00:12Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | zhengyoufan | 150,060,690 | U_kgDOCPG-kg | User | false |
openclaw/openclaw | 4,330,416,924 | I_kwDOQb6kR88AAAABAhzrHA | 72,058 | https://github.com/openclaw/openclaw/issues/72058 | https://api.github.com/repos/openclaw/openclaw/issues/72058 | [Bug]: npm deps are missing after update to 24.04.2026 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrade from 22.04 - 24.04 OC dont response in telegram and tui isnt starting.
npm install loop with:
```
`Bundled plugin runtime deps are missing. │
│ - @anthropic-ai/sd... | closed | completed | true | 11 | [
"bug",
"regression"
] | [] | 2026-04-26T08:28:38Z | 2026-04-30T04:36:13Z | 2026-04-27T06:51:39Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | amnesia106 | 40,226,796 | MDQ6VXNlcjQwMjI2Nzk2 | User | false |
openclaw/openclaw | 4,171,483,152 | I_kwDOQb6kR874o8gQ | 57,766 | https://github.com/openclaw/openclaw/issues/57766 | https://api.github.com/repos/openclaw/openclaw/issues/57766 | fix(delivery): deliverOutboundPayloads return type cannot distinguish hook cancellations from failures | ## Problem
\`deliverOutboundPayloads()\` returns \`OutboundDeliveryResult[]\`, but an empty array (\`[]\`) is ambiguous — it can mean:
1. **Hook cancellation**: A \`message_sending\` hook cancelled all payloads via \`{ cancel: true }\`. This is intentional policy, not a failure.
2. **Delivery failure**: The send was ... | closed | not_planned | true | 1 | [] | [] | 2026-03-30T14:58:24Z | 2026-04-29T04:40:26Z | 2026-04-26T08:51:12Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Kaspre | 36,520,309 | MDQ6VXNlcjM2NTIwMzA5 | User | false |
openclaw/openclaw | 4,171,751,328 | I_kwDOQb6kR874p9-g | 57,791 | https://github.com/openclaw/openclaw/issues/57791 | https://api.github.com/repos/openclaw/openclaw/issues/57791 | [Feature Request] Gmail hook preset needs system prompt support | ## Problem
The Gmail hook preset creates an agent session to process incoming emails, but it only defines a `messageTemplate` for formatting email content. There is no mechanism to provide system-level instructions to the agent.
When a lightweight model (like Flash Lite) receives an email with no task directive, it h... | closed | not_planned | true | 8 | [] | [] | 2026-03-30T15:33:08Z | 2026-04-29T04:40:25Z | 2026-04-26T08:51:16Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | lobsteryv2 | 270,135,832 | U_kgDOEBnyGA | User | false |
openclaw/openclaw | 4,171,764,388 | I_kwDOQb6kR874qBKk | 57,792 | https://github.com/openclaw/openclaw/issues/57792 | https://api.github.com/repos/openclaw/openclaw/issues/57792 | v2026.3.28: AJV schema validator crashes with Maximum call stack size exceeded, agents produce zero output | ## Summary
After upgrading from v2026.3.24 to v2026.3.28, all agent runs complete in ~8ms with zero output. No model calls are made. The gateway appears to start normally but every agent run silently fails.
## Root cause
The log file reveals `RangeError: Maximum call stack size exceeded` in the AJV schema compiler d... | closed | completed | true | 1 | [] | [] | 2026-03-30T15:35:12Z | 2026-04-29T04:40:24Z | 2026-04-26T08:51:20Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | disolutionsFL | 48,155,514 | MDQ6VXNlcjQ4MTU1NTE0 | User | false |
openclaw/openclaw | 4,171,827,726 | I_kwDOQb6kR874qQoO | 57,796 | https://github.com/openclaw/openclaw/issues/57796 | https://api.github.com/repos/openclaw/openclaw/issues/57796 | Preserve provenance field in plugin hook event messages | ## Summary
Plugin hook events (`agent_end`, `before_compaction`, `before_reset`) strip the `provenance` field from messages before passing them to plugins. The data exists in the session transcript `.jsonl` files but is not available to plugins at runtime.
This causes plugins that need to distinguish message senders ... | closed | not_planned | true | 3 | [] | [] | 2026-03-30T15:44:59Z | 2026-04-29T04:40:23Z | 2026-04-26T08:51:24Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ab0991-oss | 268,141,382 | U_kgDOD_uDRg | User | false |
openclaw/openclaw | 4,172,348,390 | I_kwDOQb6kR874sPvm | 57,844 | https://github.com/openclaw/openclaw/issues/57844 | https://api.github.com/repos/openclaw/openclaw/issues/57844 | Slack Socket Mode: inbound events silently dropped on v2026.3.11+ (works on v2026.3.8) | ## Description
After upgrading from OpenClaw v2026.3.8 to any version v2026.3.11+, Slack Socket Mode stops receiving inbound events entirely. The socket connects, channels resolve, `openclaw channels status --probe` reports "works", but **zero inbound events are logged**. Not even received-and-dropped -- completely si... | closed | completed | true | 2 | [] | [] | 2026-03-30T16:49:51Z | 2026-04-29T04:40:23Z | 2026-04-26T08:51:33Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jhillgvtc | 43,814,591 | MDQ6VXNlcjQzODE0NTkx | User | false |
openclaw/openclaw | 4,174,854,160 | I_kwDOQb6kR8741zgQ | 58,010 | https://github.com/openclaw/openclaw/issues/58010 | https://api.github.com/repos/openclaw/openclaw/issues/58010 | docs/discord: bot-to-bot messaging — guild users list is primary gate, "bot:" prefix in allowFrom not recognized | ## Summary
Two undocumented behaviors make bot-to-bot Discord messaging very difficult to configure correctly, with silent failures at every step.
---
## Issue 1: `guilds.<id>.users` is the primary access gate — bot IDs must be added here
When `guilds.<id>.users` is a non-empty array, it acts as a strict allowlist ... | closed | not_planned | true | 3 | [] | [] | 2026-03-31T00:30:18Z | 2026-04-29T04:40:22Z | 2026-04-26T08:52:04Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | KR-Python | 48,324,619 | MDQ6VXNlcjQ4MzI0NjE5 | User | false |
openclaw/openclaw | 4,174,962,665 | I_kwDOQb6kR8742N_p | 58,027 | https://github.com/openclaw/openclaw/issues/58027 | https://api.github.com/repos/openclaw/openclaw/issues/58027 | normalizeExtraMemoryPaths does not expand tilde (~) in extraPaths | ## Bug
`normalizeExtraMemoryPaths()` in the memory search module does not expand `~/` prefixes in `agents.defaults.memorySearch.extraPaths`. The function uses `path.isAbsolute()` which returns `false` for `~/` paths, causing the path to be resolved relative to each agent's workspace directory.
## Expected Behavior
`... | closed | not_planned | true | 2 | [] | [] | 2026-03-31T01:02:27Z | 2026-04-29T04:40:21Z | 2026-04-26T08:52:13Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | stadman | 10,135,297 | MDQ6VXNlcjEwMTM1Mjk3 | User | false |
openclaw/openclaw | 4,175,238,189 | I_kwDOQb6kR8743RQt | 58,064 | https://github.com/openclaw/openclaw/issues/58064 | https://api.github.com/repos/openclaw/openclaw/issues/58064 | Discord/Telegram channels silently fail to initialize under LaunchAgent after restart | # Bug Report: Discord/Telegram channels silently fail to initialize under LaunchAgent
## Bug type
Bug (channels work in foreground, silently skip under LaunchAgent)
## Summary
When the OpenClaw gateway runs as a macOS LaunchAgent, Discord and Telegram channels silently fail to initialize after the first process lif... | closed | completed | true | 2 | [] | [] | 2026-03-31T02:28:50Z | 2026-04-29T04:40:20Z | 2026-04-26T08:52:17Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jk-liche | 205,320,458 | U_kgDODDzxCg | User | false |
openclaw/openclaw | 4,175,258,106 | I_kwDOQb6kR8743WH6 | 58,071 | https://github.com/openclaw/openclaw/issues/58071 | https://api.github.com/repos/openclaw/openclaw/issues/58071 | Image tool aborts on slow vision models despite successful direct API calls | # Image tool aborts on slow vision models despite successful direct API calls
## Environment
- **OpenClaw**: 2026.3.28 (f9b1079)
- **OS**: Windows_NT 10.0.26200 (x64)
- **Node.js**: v24.13.0
- **Image Model**: zai/glm-4.6v (GLM-4.6V via Z.AI Coding Plan)
- **Model Endpoint**: `https://api.z.ai/api/coding/paas/v4/chat... | closed | not_planned | true | 1 | [] | [] | 2026-03-31T02:35:41Z | 2026-04-29T04:40:20Z | 2026-04-26T08:52:21Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Kaiji-Z | 255,520,104 | U_kgDODzrtaA | User | false |
openclaw/openclaw | 4,175,387,428 | I_kwDOQb6kR87431sk | 58,098 | https://github.com/openclaw/openclaw/issues/58098 | https://api.github.com/repos/openclaw/openclaw/issues/58098 | `HEARTBEAT_OK` in the middle of a reply is not stripped | \`stripTokenAtEdges\` handles the token at the start or end of a message but not in the middle. When a model produces something like:
\`\`\`
Some text before...HEARTBEAT_OK
This is the actual reply.
\`\`\`
the while loop in \`stripTokenAtEdges\` never sets \`didStrip = true\` — \`startsWith\` misses it, the trailing... | closed | not_planned | true | 1 | [] | [] | 2026-03-31T03:13:25Z | 2026-04-29T04:40:19Z | 2026-04-26T08:52:25Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | MaxCHL | 98,648,732 | U_kgDOBeFCnA | User | false |
openclaw/openclaw | 4,175,422,321 | I_kwDOQb6kR8743-Nx | 58,105 | https://github.com/openclaw/openclaw/issues/58105 | https://api.github.com/repos/openclaw/openclaw/issues/58105 | Cron: support multi-channel delivery (fanout) | ## Problem
Cron jobs support a single `delivery` target (one channel + one destination). To deliver the same job output to multiple channels (e.g., Telegram + Discord), you have to duplicate the entire job.
## Proposal
Support an array of delivery targets per job, or add `--also-deliver` / `additionalDelivery` entri... | closed | not_planned | true | 1 | [] | [] | 2026-03-31T03:22:59Z | 2026-04-29T04:40:18Z | 2026-04-26T08:52:29Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | markoalleno | 7,337,839 | MDQ6VXNlcjczMzc4Mzk= | User | false |
openclaw/openclaw | 4,175,963,588 | I_kwDOQb6kR8746CXE | 58,150 | https://github.com/openclaw/openclaw/issues/58150 | https://api.github.com/repos/openclaw/openclaw/issues/58150 | [Bug]: RISC-V64: OpenClaw fails with LLM request failed: network connection error even though direct HTTP requests work | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
OpenClaw on my RISC-V64 machine fails to complete chat turns with LLM request failed: network connection error, even though the same provider is reachable from the same machine via direct HTTP requests.
### St... | closed | not_planned | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-03-31T05:35:38Z | 2026-04-29T04:40:17Z | 2026-04-26T08:52:33Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | youzijin888 | 223,833,710 | U_kgDODVdubg | User | false |
openclaw/openclaw | 4,176,681,593 | I_kwDOQb6kR8748xp5 | 58,201 | https://github.com/openclaw/openclaw/issues/58201 | https://api.github.com/repos/openclaw/openclaw/issues/58201 | Feature Request: Agent-level plugin configuration override | ## Description
Currently, OpenClaw plugin configuration is global (`plugins.entries` in `openclaw.json`). All agents share the same config.
When using stateful plugins like `graph-memory`, different agents (e.g. personal assistant vs writing assistant) need **isolated data storage** (e.g. different `dbPath`), but thi... | closed | not_planned | true | 1 | [] | [] | 2026-03-31T07:26:34Z | 2026-04-29T04:40:16Z | 2026-04-26T08:54:12Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | eriic828 | 86,403,699 | MDQ6VXNlcjg2NDAzNjk5 | User | false |
openclaw/openclaw | 4,176,870,098 | I_kwDOQb6kR8749frS | 58,218 | https://github.com/openclaw/openclaw/issues/58218 | https://api.github.com/repos/openclaw/openclaw/issues/58218 | [Bug]: | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
# [Bug] auto-skill-evolver fails: LEGACY_CONFIG_RULES is undefined in validateConfigObjectWithPluginsBase
## Bug Description
When running `auto-skill-evolver` (optimize_skill.py) and spawning a local agent via `openclaw ag... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-03-31T07:52:02Z | 2026-04-29T04:40:15Z | 2026-04-26T08:54:16Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | dongxujian | 116,413,530 | U_kgDOBvBUWg | User | false |
openclaw/openclaw | 4,330,461,880 | I_kwDOQb6kR88AAAABAh2auA | 72,075 | https://github.com/openclaw/openclaw/issues/72075 | https://api.github.com/repos/openclaw/openclaw/issues/72075 | [Bug]: Gateway crash loop due to unhandled promise rejection in bonjour plugin (CIAO PROBING CANCELLED) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Description:
The OpenClaw gateway enters a crash loop on startup when the bonjour plugin is enabled in certain environments (e.g., VPS or restricted Docker containers).
The crash occurs after the gateway ... | closed | completed | true | 1 | [
"bug",
"duplicate",
"bug:behavior"
] | [] | 2026-04-26T08:54:01Z | 2026-04-29T04:39:37Z | 2026-04-26T09:18:13Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | bithuxxb | 203,383,683 | U_kgDODB9jgw | User | false |
openclaw/openclaw | 4,177,313,895 | I_kwDOQb6kR874_MBn | 58,254 | https://github.com/openclaw/openclaw/issues/58254 | https://api.github.com/repos/openclaw/openclaw/issues/58254 | [Bug]: gateway fails to restart after auto-update on macOS (launchd ThrottleInterval race) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After each nightly auto-update on macOS, the gateway process exits and never comes back up; it must be manually restarted with `openclaw gateway start`. Reproduced on every nightly update cycle across multiple days (confirme... | closed | completed | true | 2 | [] | [] | 2026-03-31T08:53:06Z | 2026-04-29T04:40:15Z | 2026-04-26T08:54:21Z | CONTRIBUTOR | null | 20260429T060025Z | 2026-04-29T06:00:25Z | YiWang24 | 176,262,341 | U_kgDOCoGMxQ | User | false |
openclaw/openclaw | 4,178,565,038 | I_kwDOQb6kR875D9eu | 58,356 | https://github.com/openclaw/openclaw/issues/58356 | https://api.github.com/repos/openclaw/openclaw/issues/58356 | [Bug]: system.run.prepare broken after update to v2026.3.28 — downgrade to v2026.3.24 does not fix | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After OpenClaw auto-updated from v2026.3.24 to v2026.3.28 on 2026-03-29,
the exec tool via the paired macOS node stopped working. Downgrading CLI
and App back to v2026.3.24 does NOT restore functionality.
### Steps to rep... | closed | not_planned | true | 5 | [
"bug"
] | [] | 2026-03-31T11:59:21Z | 2026-04-29T04:40:14Z | 2026-04-26T08:54:25Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | boejoe66-hash | 272,502,181 | U_kgDOED4NpQ | User | false |
openclaw/openclaw | 4,179,730,804 | I_kwDOQb6kR875IaF0 | 58,413 | https://github.com/openclaw/openclaw/issues/58413 | https://api.github.com/repos/openclaw/openclaw/issues/58413 | Add X-Agent-ID header to outgoing provider requests | ## Problem
When routing Anthropic API requests through a usage-tracking proxy, there's no way to identify which agent made the request. All agents on the same gateway share the same `ANTHROPIC_BASE_URL` and API key, so the proxy can't distinguish between them.
## Use Case
I run multiple agents (Alfred, Sam, Layla, M... | closed | not_planned | true | 1 | [] | [] | 2026-03-31T14:33:14Z | 2026-04-29T04:40:13Z | 2026-04-26T08:54:34Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jojopas | 6,856,488 | MDQ6VXNlcjY4NTY0ODg= | User | false |
openclaw/openclaw | 4,179,995,779 | I_kwDOQb6kR875JayD | 58,429 | https://github.com/openclaw/openclaw/issues/58429 | https://api.github.com/repos/openclaw/openclaw/issues/58429 | Feature request: auto-cleanup completed subagent sessions (like cron.sessionRetention) | ## Problem
Subagent sessions spawned with `mode: "run"` remain in the session store (`sessions.json`) indefinitely after completing (`status: done`). They serve no purpose after delivering their result to the parent session — they can't be resumed, their transcript is already persisted as a `.jsonl` file on disk, and ... | closed | not_planned | true | 1 | [] | [] | 2026-03-31T15:03:52Z | 2026-04-29T04:40:12Z | 2026-04-26T08:54:39Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | beto-sudo | 229,640,506 | U_kgDODbAJOg | User | false |
openclaw/openclaw | 4,181,160,361 | I_kwDOQb6kR875N3Gp | 58,505 | https://github.com/openclaw/openclaw/issues/58505 | https://api.github.com/repos/openclaw/openclaw/issues/58505 | [Feature]: Allow `before_prompt_build` hook to abort LLM call and return a custom response | ### Summary
Allow `before_prompt_build` to return `{ abort: true, abortResponse: "..." }` to skip the
LLM call entirely and surface a fixed text response to the user.
### Problem to solve
Today the only pre-LLM hooks available to plugins are:
| Hook | Returnable fields | Can hard-block? |
|------|-------... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-03-31T17:48:00Z | 2026-04-29T04:40:12Z | 2026-04-26T08:54:48Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | thinkycx | 13,138,128 | MDQ6VXNlcjEzMTM4MTI4 | User | false |
openclaw/openclaw | 4,181,800,338 | I_kwDOQb6kR875QTWS | 58,541 | https://github.com/openclaw/openclaw/issues/58541 | https://api.github.com/repos/openclaw/openclaw/issues/58541 | Google Chat (G Suite Add-on) returns 401 Unauthorized — JWT verification failure | # Bug Report: Google Chat (G Suite Add-on) JWT Verification Failure
## Environment
- **OpenClaw version:** 2026.3.28 (f9b1079)
- **macOS host:** Patrik's Mac mini (Darwin/arm64)
- **Gateway:** Running on port 18789, exposed via Tailscale Funnel
- **Public URL:** https://patriks-mac-mini.tail284d5a.ts.net/googlechat
#... | closed | completed | true | 5 | [] | [] | 2026-03-31T19:30:58Z | 2026-04-29T04:40:11Z | 2026-04-26T08:54:52Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Teapot-Agency | 138,458,894 | U_kgDOCEC3Dg | User | false |
openclaw/openclaw | 4,182,636,409 | I_kwDOQb6kR875Tfd5 | 58,581 | https://github.com/openclaw/openclaw/issues/58581 | https://api.github.com/repos/openclaw/openclaw/issues/58581 | Heartbeat schedules by time of day (daypart / variable intervals) | ## Summary
OpenClaw heartbeat currently uses a single fixed `heartbeat.every` interval. That works, but it is a blunt instrument.
It would be useful to support **different heartbeat intervals at different times of day**, so users can run heartbeats more frequently during active hours and less frequently during evenin... | closed | not_planned | true | 2 | [] | [] | 2026-03-31T22:00:25Z | 2026-04-29T04:40:10Z | 2026-04-26T08:54:56Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | claraassistagent | 258,727,452 | U_kgDOD2veHA | User | false |
openclaw/openclaw | 4,182,804,867 | I_kwDOQb6kR875UImD | 58,595 | https://github.com/openclaw/openclaw/issues/58595 | https://api.github.com/repos/openclaw/openclaw/issues/58595 | [Bug]: esbuild binary incompatible with macOS 10.15 Catalina (macOS 12.0+ required) since ~2026.3.8 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
OpenClaw versions after 2026.3.8 fail to start on macOS 10.15.7 (Catalina) due to an esbuild binary that requires macOS 12.0+. The Discord plugin fails to load with Cannot find module '@buape/carbon', and any attempt to inst... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-03-31T22:39:08Z | 2026-04-29T04:40:09Z | 2026-04-26T08:55:00Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | alex-espinoza | 1,746,435 | MDQ6VXNlcjE3NDY0MzU= | User | false |
openclaw/openclaw | 4,184,448,806 | I_kwDOQb6kR875aZ8m | 58,786 | https://github.com/openclaw/openclaw/issues/58786 | https://api.github.com/repos/openclaw/openclaw/issues/58786 | Subagent announce timeout can destabilize gateway | ## Problem
When a sub-agent times out, its announce completion flow can cause gateway instability.
**Observed behavior (2026-04-01):**
1. Sub-agent timed out after ~4.5 minutes (configured `runTimeoutSeconds: 900` in `agents.defaults.subagents`, but actual timeout was much shorter)
2. Announce completion attempted t... | closed | completed | true | 3 | [] | [] | 2026-04-01T06:23:19Z | 2026-04-29T04:40:09Z | 2026-04-26T08:55:17Z | CONTRIBUTOR | null | 20260429T060025Z | 2026-04-29T06:00:25Z | kagura-agent | 268,167,063 | U_kgDOD_vnlw | User | false |
openclaw/openclaw | 4,184,602,713 | I_kwDOQb6kR875a_hZ | 58,806 | https://github.com/openclaw/openclaw/issues/58806 | https://api.github.com/repos/openclaw/openclaw/issues/58806 | feat: pass requesterSenderId and senderIsOwner to ChannelAgentToolFactory | ## Problem
`listChannelAgentTools()` currently only passes `{ cfg }` to the `ChannelAgentToolFactory`:
```typescript
// pi-embedded (line ~119112)
...listChannelAgentTools({ cfg: options?.config }),
```
However, in the **same function**, the immediately following `createOpenClawTools()` call receives the full contex... | closed | not_planned | true | 1 | [] | [] | 2026-04-01T06:54:49Z | 2026-04-29T04:40:08Z | 2026-04-26T08:55:26Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Jerry-Xin | 3,401,616 | MDQ6VXNlcjM0MDE2MTY= | User | false |
openclaw/openclaw | 4,330,465,620 | I_kwDOQb6kR88AAAABAh2pVA | 72,077 | https://github.com/openclaw/openclaw/issues/72077 | https://api.github.com/repos/openclaw/openclaw/issues/72077 | [Bug]: claude-cli harness regression on 2026.4.21+ (related to locked #61093) — workaround: claude-max-api-proxy | Filing as a new issue because #61093 is locked. Posting here so the workaround is searchable for anyone hitting the same regression — bobbercheng noted it was still broken on 2026.4.5; I am hitting it on 2026.4.21 → 2026.4.24.
---
Hit this on OpenClaw 2026.4.21 → 2026.4.24 on Ubuntu 24.04, Node 24.14.1, Claude Max su... | closed | completed | true | 1 | [] | [] | 2026-04-26T08:56:12Z | 2026-04-29T04:38:48Z | 2026-04-26T09:42:09Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | shandutta | 8,773,397 | MDQ6VXNlcjg3NzMzOTc= | User | false |
openclaw/openclaw | 4,185,177,098 | I_kwDOQb6kR875dLwK | 58,875 | https://github.com/openclaw/openclaw/issues/58875 | https://api.github.com/repos/openclaw/openclaw/issues/58875 | Usage footer: token counts missing 'k' suffix, showing raw thousands | ## Description
The usage footer appended to messages (when `/usage full` is enabled) displays token counts like:
```
Usage: 7 in / 545 out · session agent:main:main
```
The `7 in` actually means ~7,000 input tokens, but there's no `k` suffix, making it look like 7 tokens. This is confusing.
## Expected behavior
``... | closed | not_planned | true | 3 | [] | [] | 2026-04-01T08:37:36Z | 2026-04-29T04:40:07Z | 2026-04-26T08:57:26Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | sashakang | 44,499,389 | MDQ6VXNlcjQ0NDk5Mzg5 | User | false |
openclaw/openclaw | 4,186,252,083 | I_kwDOQb6kR875hSMz | 58,983 | https://github.com/openclaw/openclaw/issues/58983 | https://api.github.com/repos/openclaw/openclaw/issues/58983 | [Feature]: Feishu Interactive Approval Cards - One-click Approve/Deny Buttons | ### Summary
Feishu channel approval requests sent as interactive card messages with Approve/Deny buttons instead of plain text commands.
### Problem to solve
Currently when exec commands require approval, users must remember and type a text command like `/approve <id> allow-once` in the chat. This is error-prone and... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-04-01T11:41:01Z | 2026-04-29T04:40:06Z | 2026-04-26T08:57:29Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Lrs-GreenBottle | 253,340,193 | U_kgDODxmqIQ | User | false |
openclaw/openclaw | 4,186,824,481 | I_kwDOQb6kR875jd8h | 59,026 | https://github.com/openclaw/openclaw/issues/59026 | https://api.github.com/repos/openclaw/openclaw/issues/59026 | [Bug]: Discord thread binding not released after ACP task completion — channel messages misrouted to thread | ## Description
After an ACP task (Codex) completes in a Discord thread, subsequent messages sent by the user in the **parent channel** (not the thread) are still routed to the thread session. The channel session only recovers after the second message.
## Environment
- OpenClaw version: 2026.3.31
- Channel: Discord (... | closed | completed | true | 2 | [] | [] | 2026-04-01T13:10:44Z | 2026-04-29T04:40:06Z | 2026-04-26T08:57:33Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | CountofGlamorgan | 59,939,121 | MDQ6VXNlcjU5OTM5MTIx | User | false |
openclaw/openclaw | 4,187,092,039 | I_kwDOQb6kR875kfRH | 59,047 | https://github.com/openclaw/openclaw/issues/59047 | https://api.github.com/repos/openclaw/openclaw/issues/59047 | [Bug]: External plugin tools register in metadata but are not surfaced to agent sessions [in TUI] | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Tools registered via api.registerTool() in external (non-bundled) plugins appear in openclaw plugins inspect output but are never exposed to agent sessions via openclaw agent. Bundled plugin tools (e.g. memory_... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-01T13:48:01Z | 2026-04-29T04:40:05Z | 2026-04-26T08:57:38Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | FroeMic | 15,179,255 | MDQ6VXNlcjE1MTc5MjU1 | User | false |
openclaw/openclaw | 4,187,648,005 | I_kwDOQb6kR875mnAF | 59,090 | https://github.com/openclaw/openclaw/issues/59090 | https://api.github.com/repos/openclaw/openclaw/issues/59090 | Android: Gateway TLS handshake fails on SDK 36 with domain-specific network security config | ## Summary
On Android 16 (SDK 36), the gateway TLS handshake fails immediately during onboarding when connecting via a Tailscale `.ts.net` domain. The app never reaches the pairing stage — the connection is rejected at the TLS layer before any data is exchanged.
## Error
```
Gateway error: Domain specific configurat... | closed | not_planned | true | 1 | [] | [] | 2026-04-01T15:02:12Z | 2026-04-29T04:40:04Z | 2026-04-26T08:57:42Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | bdhwan | 5,001,286 | MDQ6VXNlcjUwMDEyODY= | User | false |
openclaw/openclaw | 4,187,939,003 | I_kwDOQb6kR875nuC7 | 59,124 | https://github.com/openclaw/openclaw/issues/59124 | https://api.github.com/repos/openclaw/openclaw/issues/59124 | security config option not included in openclaw configure wizard | When OpenClaw updates via npm, the tools.exec.security setting gets reset to a default value (not full). Users aren't warned during update, and the configure wizard doesn't offer this option.
## Current behavior
- Update OpenClaw → security resets
- openclaw configure wizard doesn't include security setting
- Result: ... | closed | completed | true | 1 | [] | [] | 2026-04-01T15:40:41Z | 2026-04-29T04:40:03Z | 2026-04-26T08:57:46Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | haschild | 10,702,483 | MDQ6VXNlcjEwNzAyNDgz | User | false |
openclaw/openclaw | 4,189,643,786 | I_kwDOQb6kR875uOQK | 59,238 | https://github.com/openclaw/openclaw/issues/59238 | https://api.github.com/repos/openclaw/openclaw/issues/59238 | [Bug]: Slack per-channel systemPrompt is assembled but agent behavior ignores it | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Per-channel `systemPrompt` configured under `channels.slack.channels.<id>.systemPrompt` is correctly resolved through the Slack adapter, inbound context, and reply pipeline — but channel agent behavior consistently ignores the instructions.... | closed | completed | true | 1 | [] | [] | 2026-04-01T20:22:13Z | 2026-04-29T04:40:02Z | 2026-04-26T08:57:59Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | rimkute | 112,832,114 | U_kgDOBrmucg | User | false |
openclaw/openclaw | 4,190,657,910 | I_kwDOQb6kR875yF12 | 59,304 | https://github.com/openclaw/openclaw/issues/59304 | https://api.github.com/repos/openclaw/openclaw/issues/59304 | Feature proposal: `openclaw update hint` for agent/skill update notifications | ## Problem
When a maintainer invokes an OpenClaw skill in Claude Code, there's no way to know a newer version is available. The user has to remember to run `openclaw update` manually. Tools like [gstack](https://github.com/garrytan/gstack) solve this with a preamble pattern — every skill invocation runs a lightweight ... | closed | not_planned | true | 1 | [] | [] | 2026-04-02T00:29:34Z | 2026-04-29T04:40:01Z | 2026-04-26T08:58:13Z | CONTRIBUTOR | null | 20260429T060025Z | 2026-04-29T06:00:25Z | simonemacario | 2,116,609 | MDQ6VXNlcjIxMTY2MDk= | User | false |
openclaw/openclaw | 4,190,966,615 | I_kwDOQb6kR875zRNX | 59,345 | https://github.com/openclaw/openclaw/issues/59345 | https://api.github.com/repos/openclaw/openclaw/issues/59345 | Control UI: subagent context badge can show historical total tokens as current context usage | # Control UI: subagent context badge can display historical total tokens as current context usage
## Summary
The Control UI subagent/session badge can show impossible values such as:
- `100% context used 468.2k / 48k`
This appears to be a UI/display bug, not a real live-context overflow.
## Why this matters
The bad... | closed | completed | true | 2 | [] | [] | 2026-04-02T02:15:34Z | 2026-04-29T04:40:00Z | 2026-04-26T08:58:17Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | dmitriiforpost-commits | 262,919,577 | U_kgDOD6vVmQ | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.