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 | 4,099,685,893 | I_kwDOQb6kR870XD4F | 50,277 | https://github.com/openclaw/openclaw/issues/50277 | https://api.github.com/repos/openclaw/openclaw/issues/50277 | [Bug]: 聊天页面删除对话时弹框会超出边界 | ### Bug type
Regression (worked before, now fails)
### Summary
会导致弹框超出页面的边界使得无法点击删除按钮。
.chat-delete-confirm {
/* position: absolute; */
bottom: calc(100% + 6px);
left: 0;
background: var(--card, #1a1a1a);
border: 1p... | open | null | false | 3 | [
"bug",
"regression"
] | [] | 2026-03-19T06:41:39Z | 2026-05-04T18:22:21Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | slladmin | 58,767,080 | MDQ6VXNlcjU4NzY3MDgw | User | false |
openclaw/openclaw | 4,099,742,059 | I_kwDOQb6kR870XRlr | 50,287 | https://github.com/openclaw/openclaw/issues/50287 | https://api.github.com/repos/openclaw/openclaw/issues/50287 | Next Sprint B: model-input validation guardrails | ## Decision Snapshot (Confirmed)
- Split issue approach approved
- Policy context: cost-first remains primary
- Execution timing: wait for next sprint
## Objective
Add guardrails so unsupported model input modalities are rejected before deployment.
## In Scope
1. Strengthen schema validation for `models.providers.*.... | open | null | false | 3 | [] | [] | 2026-03-19T06:55:45Z | 2026-05-04T18:22:23Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | Piboonsak | 17,107,792 | MDQ6VXNlcjE3MTA3Nzky | User | false |
openclaw/openclaw | 4,099,770,879 | I_kwDOQb6kR870XYn_ | 50,289 | https://github.com/openclaw/openclaw/issues/50289 | https://api.github.com/repos/openclaw/openclaw/issues/50289 | [Bug]: Performance Issue: Large `ownerAllowFrom` Configuration Causes Severe Slowdown | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
# Performance Issue: Large `ownerAllowFrom` Configuration Causes Severe Slowdown
## Summary
When `commands.ownerAllowFrom` contains a large number of user IDs (9000+ entries), OpenClaw Gateway experiences severe performance degradation. E... | open | null | false | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-03-19T07:02:44Z | 2026-05-04T18:22:25Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | songchenghao | 4,641,267 | MDQ6VXNlcjQ2NDEyNjc= | User | false |
openclaw/openclaw | 4,099,809,107 | I_kwDOQb6kR870Xh9T | 50,291 | https://github.com/openclaw/openclaw/issues/50291 | https://api.github.com/repos/openclaw/openclaw/issues/50291 | Plugin Hooks: Missing trace context for observability (messageId, runId in all hooks, parentSpanId) | # Plugin Hooks: Missing trace context fields for distributed tracing
## Problem
The current Plugin Hooks event/context data is **insufficient for building accurate distributed traces**, particularly in these scenarios:
1. **Concurrent messages in group chats**: Multiple users sending messages simultaneously - cannot... | open | null | false | 6 | [] | [] | 2026-03-19T07:11:04Z | 2026-05-04T18:22:27Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | Hello-World-0X | 25,972,398 | MDQ6VXNlcjI1OTcyMzk4 | User | false |
openclaw/openclaw | 4,100,798,628 | I_kwDOQb6kR870bTik | 50,374 | https://github.com/openclaw/openclaw/issues/50374 | https://api.github.com/repos/openclaw/openclaw/issues/50374 | OpenAI-compat providers: strip extra message keys before sending | ## Problem
When using OpenAI-compatible providers (configured with `api: "openai-completions"`), OpenClaw sends message objects with extra keys beyond the standard `role` and `content` — fields like `name`, `tool_calls`, `tool_call_id`, `cache_control`, etc.
While OpenAI's API silently ignores unknown keys, stricter ... | open | null | false | 4 | [] | [] | 2026-03-19T10:34:20Z | 2026-05-04T20:15:06Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | choutos | 43,113,488 | MDQ6VXNlcjQzMTEzNDg4 | User | false |
openclaw/openclaw | 4,101,198,414 | I_kwDOQb6kR870c1JO | 50,404 | https://github.com/openclaw/openclaw/issues/50404 | https://api.github.com/repos/openclaw/openclaw/issues/50404 | Feature Request: Session Sidebar with Chat History Management | # Feature Request: Session Sidebar with Chat History Management
## Problem Statement
Currently, the Control UI lacks proper session management. When users click the "+" button to start a new chat, previous conversations are not properly saved to the session history list, making it impossible to switch between differe... | closed | not_planned | false | 6 | [] | [] | 2026-03-19T11:56:35Z | 2026-05-04T18:48:11Z | 2026-05-04T18:48:11Z | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | herodv | 112,804,671 | U_kgDOBrlDPw | User | false |
openclaw/openclaw | 4,101,653,399 | I_kwDOQb6kR870ekOX | 50,442 | https://github.com/openclaw/openclaw/issues/50442 | https://api.github.com/repos/openclaw/openclaw/issues/50442 | [Bug] backup create leaves large .tmp files on timeout causing disk space exhaustion | ## Bug Description
Using `openclaw backup create` command, if the process is killed due to timeout, large .tmp temporary files are left behind in the current directory without cleanup.
## Steps to Reproduce
1. Set a short timeout (e.g., 300 seconds)
2. Run `openclaw backup create`
3. Wait for timeout
## Actual Impa... | open | null | false | 5 | [] | [] | 2026-03-19T13:21:52Z | 2026-05-04T18:22:57Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | hillerliao | 5,266,745 | MDQ6VXNlcjUyNjY3NDU= | User | false |
openclaw/openclaw | 4,102,275,920 | I_kwDOQb6kR870g8NQ | 50,481 | https://github.com/openclaw/openclaw/issues/50481 | https://api.github.com/repos/openclaw/openclaw/issues/50481 | Slack: support assistant.threads.setSuggestedPrompts for dynamic prompt suggestions | ## Summary
OpenClaw's Slack integration already calls `assistant.threads.setStatus` for typing indicators, but does not use the related `assistant.threads.setSuggestedPrompts` method. This method allows setting up to 4 contextual prompt suggestions that appear in the Slack UI after each bot message in an assistant thr... | open | null | false | 3 | [] | [] | 2026-03-19T14:56:05Z | 2026-05-04T20:15:11Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | leslielesleyles | 262,164,716 | U_kgDOD6BQ7A | User | false |
openclaw/openclaw | 4,102,346,424 | I_kwDOQb6kR870hNa4 | 50,482 | https://github.com/openclaw/openclaw/issues/50482 | https://api.github.com/repos/openclaw/openclaw/issues/50482 | Feature request: Expose voice message flag in inbound agent context | ## Problem
When a user sends a voice message on Telegram (or other channels that support voice), OpenClaw correctly routes the audio through STT (e.g. Deepgram) and injects the transcript into the agent context. However, the agent has no way to know whether the original message was a voice note or typed text.
This di... | open | null | false | 3 | [] | [] | 2026-03-19T15:08:38Z | 2026-05-04T20:15:13Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | JTOrca | 32,879,384 | MDQ6VXNlcjMyODc5Mzg0 | User | false |
openclaw/openclaw | 4,102,466,897 | I_kwDOQb6kR870hq1R | 50,490 | https://github.com/openclaw/openclaw/issues/50490 | https://api.github.com/repos/openclaw/openclaw/issues/50490 | [Bug]: Feishu 群聊中 activation 模式切换无效，始终响应所有消息 | ### Bug type
Regression (worked before, now fails)
### Summary
在群聊中使用 `/activation mention` 命令切换激活模式后，机器人仍然响应所有消息，未按预期进入"仅当被 @ 时才响应"的模式。
### Steps to reproduce
1. 初始配置 `mentionRequired: false`，机器人在群聊中响应所有消息（符合预期）
2. 在群聊中发送 `/activation mention` 命令
3. 系统返回确认消息（如"激活模式已切换为：mention"）
4. 发送不@机器人的普通消息
5. 观察机器人仍然响应了该消息
... | open | null | false | 4 | [
"bug",
"regression"
] | [] | 2026-03-19T15:28:52Z | 2026-05-04T18:23:03Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | Cafeynman | 54,011,597 | MDQ6VXNlcjU0MDExNTk3 | User | false |
openclaw/openclaw | 4,102,838,556 | I_kwDOQb6kR870jFkc | 50,530 | https://github.com/openclaw/openclaw/issues/50530 | https://api.github.com/repos/openclaw/openclaw/issues/50530 | Telegram: collapse mid-paragraph newlines before sending | ## Problem
LLM output often contains soft line breaks (around 80-100 chars) that are meaningless in prose but get preserved literally by Telegram. This causes:
- Bullet list items split mid-sentence
- Code blocks with unexpected line breaks
- Blockquotes broken across lines
- Generally janky-looking messages
## Prop... | open | null | false | 3 | [] | [] | 2026-03-19T16:29:36Z | 2026-05-04T20:15:17Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | charlotte-parker-kiwazo | 257,559,733 | U_kgDOD1oMtQ | User | false |
openclaw/openclaw | 4,103,084,628 | I_kwDOQb6kR870kBpU | 50,551 | https://github.com/openclaw/openclaw/issues/50551 | https://api.github.com/repos/openclaw/openclaw/issues/50551 | App freezes when text input bar changes vertical size | When the text input bar changes its vertical size (e.g. multi-line expansion), the app freezes briefly. Likely cause: bottom inset recalculation triggers a full re-layout of the message list. Changing the bottom inset/padding should not cause a full relayout — only the bottom padding is changing.
Platform: iOS/iPad, p... | open | null | false | 2 | [] | [] | 2026-03-19T17:15:35Z | 2026-05-04T20:15:19Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | leftspin | 377,231 | MDQ6VXNlcjM3NzIzMQ== | User | false |
openclaw/openclaw | 4,103,203,204 | I_kwDOQb6kR870kemE | 50,561 | https://github.com/openclaw/openclaw/issues/50561 | https://api.github.com/repos/openclaw/openclaw/issues/50561 | Auto-apply safe doctor fixes on gateway start | ## Problem
Gateway shows "Run clawdbot doctor --fix" on startup but does not auto-apply safe fixes. This leads to config drift and repeated warnings.
## Observed Behavior
- Every gateway start shows doctor suggestions
- User must manually run `clawdbot doctor --fix`
- If forgotten, config stays out of sync
## Suggest... | open | null | false | 5 | [] | [] | 2026-03-19T17:39:11Z | 2026-05-04T20:15:21Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | coywolffuturist | 264,284,025 | U_kgDOD8CneQ | User | false |
openclaw/openclaw | 4,103,212,366 | I_kwDOQb6kR870kg1O | 50,563 | https://github.com/openclaw/openclaw/issues/50563 | https://api.github.com/repos/openclaw/openclaw/issues/50563 | Message loss on gateway crash (ack-before-process) | ## Problem
Gateway acks messages to WhatsApp before processing completes. If the gateway crashes mid-processing, the message is lost — WhatsApp thinks it was delivered.
## Impact
- User sends a message
- Gateway acks receipt to WhatsApp
- Gateway crashes before agent processes the message
- Message is lost — no retry ... | open | null | false | 4 | [] | [] | 2026-03-19T17:41:06Z | 2026-05-04T18:23:10Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | coywolffuturist | 264,284,025 | U_kgDOD8CneQ | User | false |
openclaw/openclaw | 4,103,237,898 | I_kwDOQb6kR870knEK | 50,565 | https://github.com/openclaw/openclaw/issues/50565 | https://api.github.com/repos/openclaw/openclaw/issues/50565 | Tool call logs always tagged under feishu subsystem regardless of actual channel | ## Description
All tool call/done/fail log entries are tagged under `gateway/channels/feishu` subsystem logger, regardless of which channel the tool call actually originates from.
For example, when a user sends a message via **Telegram** and the agent executes an `exec` tool, the log shows:
```json
{
"0": "{\"subs... | open | null | false | 4 | [] | [] | 2026-03-19T17:46:12Z | 2026-05-04T18:23:12Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | xucheng | 881,960 | MDQ6VXNlcjg4MTk2MA== | User | false |
openclaw/openclaw | 4,103,975,053 | I_kwDOQb6kR870nbCN | 50,611 | https://github.com/openclaw/openclaw/issues/50611 | https://api.github.com/repos/openclaw/openclaw/issues/50611 | memory flush never triggers when reserveTokensFloor equals contextWindow | ## Bug
When `reserveTokensFloor` is set equal to `contextWindow` (e.g. both 200K), the memory flush threshold becomes negative, so compaction never triggers. Agents fill up the full context window without trimming.
**Expected:** `reserveTokensFloor` should default to a reasonable fraction of contextWindow (e.g. 25%),... | open | null | false | 4 | [] | [] | 2026-03-19T20:07:00Z | 2026-05-04T18:23:17Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | lars-hv | 235,699,712 | U_kgDODgx-AA | User | false |
openclaw/openclaw | 4,104,060,897 | I_kwDOQb6kR870nv_h | 50,619 | https://github.com/openclaw/openclaw/issues/50619 | https://api.github.com/repos/openclaw/openclaw/issues/50619 | Feature: /model switch should check context size against target model limits | ## Problem
When switching models with `/model` in a session with large accumulated context, the new model may immediately hit rate limits or context overflow.
**Real case:** An agent accumulated ~200k+ tokens of context (two rounds of code review) on Claude Opus, then switched to `openai-codex/gpt-5.4` via `/model`. ... | closed | not_planned | true | 4 | [] | [] | 2026-03-19T20:27:10Z | 2026-05-04T04:35:15Z | 2026-05-01T22:06:38Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | HoranCheng | 50,915,351 | MDQ6VXNlcjUwOTE1MzUx | User | false |
openclaw/openclaw | 4,104,142,072 | I_kwDOQb6kR870oDz4 | 50,630 | https://github.com/openclaw/openclaw/issues/50630 | https://api.github.com/repos/openclaw/openclaw/issues/50630 | [Bug]: Tailscale serve + auth.mode=none exposes gateway to full Tailnet without authentication | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 9.3 / 10.0 | 9.3 / 10.0 |
| **Severity** | Critical | Critical |
| **Vector** | CVSS:3.1/AV:A/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:N | CVSS:4.0/AV:A/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:N/SC:H/SI:H/SA:N |
| **Calculator** |... | open | null | false | 5 | [] | [] | 2026-03-19T20:47:00Z | 2026-05-04T18:23:21Z | null | CONTRIBUTOR | null | 20260504T210031Z | 2026-05-04T21:00:31Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 4,104,177,753 | I_kwDOQb6kR870oMhZ | 50,642 | https://github.com/openclaw/openclaw/issues/50642 | https://api.github.com/repos/openclaw/openclaw/issues/50642 | [Bug]: macOS node auto-trusts first TLS certificate and accepts rogue gateway control | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 9.0 / 10.0 | 9.5 / 10.0 |
| **Severity** | Critical | Critical |
| **Vector** | CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:C/C:H/I:H/A:H | CVSS:4.0/AV:N/AC:H/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H |
| **Calculator** |... | open | null | false | 5 | [] | [] | 2026-03-19T20:55:59Z | 2026-05-04T18:23:23Z | null | CONTRIBUTOR | null | 20260504T210031Z | 2026-05-04T21:00:31Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 4,104,601,164 | I_kwDOQb6kR870pz5M | 50,677 | https://github.com/openclaw/openclaw/issues/50677 | https://api.github.com/repos/openclaw/openclaw/issues/50677 | Skills silently truncated - user-created skills invisible with no warning or prioritization | ## Problem
When the number of eligible skills exceeds the prompt token budget, OpenClaw truncates the skills list injected into the system prompt. Currently this happens silently from the users perspective:
- A note ("Skills truncated: included 68 of 90") appears in the system prompt, but the **agent** sees it, not t... | open | null | false | 4 | [] | [] | 2026-03-19T22:41:06Z | 2026-05-04T18:23:31Z | null | CONTRIBUTOR | null | 20260504T210031Z | 2026-05-04T21:00:31Z | barronlroth | 5,567,884 | MDQ6VXNlcjU1Njc4ODQ= | User | false |
openclaw/openclaw | 4,104,642,984 | I_kwDOQb6kR870p-Go | 50,681 | https://github.com/openclaw/openclaw/issues/50681 | https://api.github.com/repos/openclaw/openclaw/issues/50681 | Telegram: support fail-closed unknown-group handling with optional auto-leave | ## Summary
OpenClaw should support a stricter Telegram group security posture for users who want a bot to operate only in explicitly approved groups.
Today, Telegram group handling supports filtering and mention-gating, but there is no clear first-class way to:
- reject unknown groups cleanly
- optionally warn the gr... | open | null | false | 3 | [] | [] | 2026-03-19T22:49:49Z | 2026-05-04T18:23:32Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | monk0bot0 | 269,586,029 | U_kgDOEBGObQ | User | false |
openclaw/openclaw | 4,104,743,655 | I_kwDOQb6kR870qWrn | 50,690 | https://github.com/openclaw/openclaw/issues/50690 | https://api.github.com/repos/openclaw/openclaw/issues/50690 | Prevent fake-progress messages with a runtime pre-send policy layer | ## Summary
OpenClaw currently allows low-value fake-progress messages to reach end users before meaningful execution is visible.
Examples:
- "on it"
- "proceeding"
- "working on it"
- "checking now"
- "starting now"
- "I’m choosing/picking the next slice now"
In practice this creates a family of failures:
1. ack wit... | open | null | false | 4 | [] | [] | 2026-03-19T23:17:45Z | 2026-05-04T18:23:36Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | addelh | 39,250,691 | MDQ6VXNlcjM5MjUwNjkx | User | false |
openclaw/openclaw | 4,104,993,404 | I_kwDOQb6kR870rTp8 | 50,719 | https://github.com/openclaw/openclaw/issues/50719 | https://api.github.com/repos/openclaw/openclaw/issues/50719 | [Bug]: 404 status code (body not found) | ### Bug type
Regression (worked before, now fails)
### Summary
Local model GLM 4.7 Flash stopped working after I upgraded OpenClaw on 19 March 2026.
### Steps to reproduce
- Have a setup with a local GLM 4.7 flash model
- `curl -fsSL https://openclaw.ai/install-cli.sh | bash` model stops working.
- `curl -fsSL htt... | open | null | false | 5 | [
"bug",
"regression"
] | [] | 2026-03-20T00:36:40Z | 2026-05-04T18:23:44Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | sboden | 15,326,242 | MDQ6VXNlcjE1MzI2MjQy | User | false |
openclaw/openclaw | 4,105,143,279 | I_kwDOQb6kR870r4Pv | 50,739 | https://github.com/openclaw/openclaw/issues/50739 | https://api.github.com/repos/openclaw/openclaw/issues/50739 | feature: system event priority/bypass-queue mode for reliable in-session alerts | ## Problem
When a session lane is congested with failing LLM requests (e.g., during a provider rate limit incident), `openclaw system event --mode now` queues behind those requests rather than injecting directly into the session message stream. This makes system events an unreliable alert surface during the exact scen... | open | null | false | 6 | [] | [] | 2026-03-20T01:28:35Z | 2026-05-04T18:23:47Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | data-t3labs | 264,575,540 | U_kgDOD8UaNA | User | false |
openclaw/openclaw | 4,105,319,594 | I_kwDOQb6kR870sjSq | 50,762 | https://github.com/openclaw/openclaw/issues/50762 | https://api.github.com/repos/openclaw/openclaw/issues/50762 | Feature request: split chat and live exec logs in Control UI | ## Background
OpenClaw's current exec experience in the web control UI feels too blocking for longer-running commands.
Right now, when an agent runs `exec`, the chat flow feels "stuck" until output comes back. This works for short commands, but for long-running or interactive commands the experience is poor.
## What... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-20T02:26:06Z | 2026-05-02T04:27:52Z | 2026-04-29T15:32:30Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | laooopooo | 8,705,013 | MDQ6VXNlcjg3MDUwMTM= | User | false |
openclaw/openclaw | 4,105,334,304 | I_kwDOQb6kR870sm4g | 50,768 | https://github.com/openclaw/openclaw/issues/50768 | https://api.github.com/repos/openclaw/openclaw/issues/50768 | feat: sub-agent templates (pre-defined prompt + model + timeout) | ## Problem
Orchestrator agents that spawn sub-agents with recurring configurations (same prompt, model, timeout) have no way to pre-define reusable sub-agent types. Today, the orchestrator must manually read prompt files, assemble task strings, and pass model/timeout params on every `sessions_spawn` call. This is repe... | open | null | false | 3 | [] | [] | 2026-03-20T02:31:56Z | 2026-05-04T18:23:54Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | Servo-AIpex | 259,493,048 | U_kgDOD3eMuA | User | false |
openclaw/openclaw | 4,105,361,592 | I_kwDOQb6kR870sti4 | 50,773 | https://github.com/openclaw/openclaw/issues/50773 | https://api.github.com/repos/openclaw/openclaw/issues/50773 | Heartbeat-cron collision avoidance for local LLM environments | ## Problem
When running with a local LLM (e.g. Ollama), concurrent cron jobs and heartbeats compete for the same inference resource. A running cron job can cause heartbeats to timeout or the LLM to respond slowly.
The heartbeat runner already skips when the **main** command lane is busy (`requests-in-flight`), but do... | closed | completed | true | 3 | [
"stale"
] | [] | 2026-03-20T02:42:18Z | 2026-05-02T04:28:40Z | 2026-04-29T09:12:56Z | CONTRIBUTOR | null | 20260502T060142Z | 2026-05-02T06:01:42Z | scottgl9 | 10,623,288 | MDQ6VXNlcjEwNjIzMjg4 | User | false |
openclaw/openclaw | 4,105,363,396 | I_kwDOQb6kR870st_E | 50,776 | https://github.com/openclaw/openclaw/issues/50776 | https://api.github.com/repos/openclaw/openclaw/issues/50776 | [Bug]: Slack directory reads return empty results despite valid user token and successful Slack API autoke | ## Summary
OpenClaw 2026.3.13 on macOS connects successfully to Slack workspace madisoncrew and reports valid bot/app/user
tokens, but Slack directory reads still return groups: [], self: null, and peers... | closed | completed | true | 4 | [] | [] | 2026-03-20T02:42:57Z | 2026-05-04T04:33:54Z | 2026-05-02T04:17:39Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | pjaillon | 7,649,173 | MDQ6VXNlcjc2NDkxNzM= | User | false |
openclaw/openclaw | 4,105,368,778 | I_kwDOQb6kR870svTK | 50,779 | https://github.com/openclaw/openclaw/issues/50779 | https://api.github.com/repos/openclaw/openclaw/issues/50779 | [Bug]: Control UI does not render tool-returned images inline in Chat | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
OpenClaw Control UI is not visibly rendering images returned from tool output.
Environment
OpenClaw Control UI: 2026.3.8
Gateway/app version: 2026.3.8
macOS
Surface: local Control UI at 127.0.0.1:18789
### Steps to reproduce
Have th... | open | null | false | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-03-20T02:44:46Z | 2026-05-04T20:15:30Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | rquinones84 | 17,855,466 | MDQ6VXNlcjE3ODU1NDY2 | User | false |
openclaw/openclaw | 4,105,466,585 | I_kwDOQb6kR870tHLZ | 50,795 | https://github.com/openclaw/openclaw/issues/50795 | https://api.github.com/repos/openclaw/openclaw/issues/50795 | Bug: Context token count always shows 0 after compaction | # Issue: Context Token Count Always Shows 0 After Compaction
## Summary
The `📚 Context` counter in the TUI always displays `0/1.0m (0%)` even after successful LLM responses. This happens because the `clearStaleAssistantUsageOnSessionMessages()` function incorrectly clears **all** assistant message usage data after a... | open | null | false | 4 | [] | [] | 2026-03-20T03:21:23Z | 2026-05-04T18:23:56Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | wuwahe3 | 268,833,312 | U_kgDOEAYSIA | User | false |
openclaw/openclaw | 4,105,477,577 | I_kwDOQb6kR870tJ3J | 50,798 | https://github.com/openclaw/openclaw/issues/50798 | https://api.github.com/repos/openclaw/openclaw/issues/50798 | [Feature]: Visible agent-to-agent messaging for ACP thread-bound sessions (proxy-only delivery without main session creation) | ## What
A way for a coordinator agent to send **visible** messages into an ACP thread-bound session's Discord thread, **without creating a `agent:main:discord:channel:<thread_id>` session** that causes route pollution.
## Why
### The gap
OpenClaw currently offers two paths to communicate with ACP sessions:
| Path ... | open | null | false | 4 | [] | [] | 2026-03-20T03:25:37Z | 2026-05-04T18:23:57Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | ruan330 | 20,943,445 | MDQ6VXNlcjIwOTQzNDQ1 | User | false |
openclaw/openclaw | 4,105,520,020 | I_kwDOQb6kR870tUOU | 50,809 | https://github.com/openclaw/openclaw/issues/50809 | https://api.github.com/repos/openclaw/openclaw/issues/50809 | Add sms.read command and SMS as a messaging channel | ## Feature Request
### `sms.read` command
Add a new `sms.read` invoke command to the Android app for fetching SMS messages by ID or by thread conversation.
**Parameters:**
- `id` (Long): Fetch a single SMS message by its `_ID`
- `threadId` (Long): Fetch a conversation thread by `THREAD_ID`
- `limit` (Int): Max messag... | open | null | false | 3 | [] | [] | 2026-03-20T03:40:56Z | 2026-05-04T18:24:02Z | null | CONTRIBUTOR | null | 20260504T210031Z | 2026-05-04T21:00:31Z | scottgl9 | 10,623,288 | MDQ6VXNlcjEwNjIzMjg4 | User | false |
openclaw/openclaw | 4,074,871,306 | I_kwDOQb6kR87y4ZoK | 45,803 | https://github.com/openclaw/openclaw/issues/45803 | https://api.github.com/repos/openclaw/openclaw/issues/45803 | Discord: Bot mentions do not trigger notifications (missing allowed_mentions) | ## Summary | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-14T06:33:28Z | 2026-05-02T04:29:18Z | 2026-04-29T05:46:18Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | claw9601-sudo | 262,031,615 | U_kgDOD55I_w | User | false |
openclaw/openclaw | 4,075,651,818 | I_kwDOQb6kR87y7YLq | 46,068 | https://github.com/openclaw/openclaw/issues/46068 | https://api.github.com/repos/openclaw/openclaw/issues/46068 | [Bug]: 401 error | ### Bug type
Regression (worked before, now fails)
### Summary
<img width="631" height="280" alt="Image" src="https://github.com/user-attachments/assets/39553359-8ebf-452e-b590-2bcbafb714aa" />
[agent/embedded] embedded run failover decision: runId=fec25511-fbe3-476f-b251-df9c27004cc2 stage=assistant decision=surfac... | closed | completed | true | 3 | [
"bug",
"stale",
"regression"
] | [] | 2026-03-14T11:55:19Z | 2026-05-02T04:29:17Z | 2026-04-29T05:46:23Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | lovetian1991 | 189,591,834 | U_kgDOC0zxGg | User | false |
openclaw/openclaw | 4,106,018,521 | I_kwDOQb6kR870vN7Z | 50,880 | https://github.com/openclaw/openclaw/issues/50880 | https://api.github.com/repos/openclaw/openclaw/issues/50880 | Steer queue mode silently degrades to followup — messages never injected mid-turn at tool call boundaries | ## Summary
When `messages.queue.mode` is set to `"steer"`, inbound messages are never injected into the active agent turn at tool call boundaries as documented. Instead, they silently degrade to followup behavior (waiting for end-of-turn, then triggering a new turn). There is zero diagnostic logging when this occurs.
... | closed | not_planned | false | 8 | [] | [] | 2026-03-20T06:09:23Z | 2026-05-04T19:55:44Z | 2026-05-04T19:55:44Z | CONTRIBUTOR | null | 20260504T210031Z | 2026-05-04T21:00:31Z | ppamment | 2,122,919 | MDQ6VXNlcjIxMjI5MTk= | User | false |
openclaw/openclaw | 4,106,081,234 | I_kwDOQb6kR870vdPS | 50,887 | https://github.com/openclaw/openclaw/issues/50887 | https://api.github.com/repos/openclaw/openclaw/issues/50887 | [Bug]: Streaming Chinese character corruption - random single characters replaced with U+FFFD | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When agents respond in Chinese via group chat (webchat), random individual Chinese characters are replaced with the U+FFFD replacement character. The pattern suggests a streaming token boundary issue where the last byte of a multi-byte UTF-... | open | null | false | 4 | [] | [] | 2026-03-20T06:26:31Z | 2026-05-04T20:23:04Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | ColaFatty | 211,209,914 | U_kgDODJbOug | User | false |
openclaw/openclaw | 4,106,157,551 | I_kwDOQb6kR870vv3v | 50,900 | https://github.com/openclaw/openclaw/issues/50900 | https://api.github.com/repos/openclaw/openclaw/issues/50900 | Feature request: per-pattern session maintenance retention rules | ## Summary
Currently `session.maintenance.pruneAfter` applies a single uniform age threshold to all session entries. This forces a compromise between aggressive cleanup of ephemeral sessions (cron runs, completed subagents) and conservative retention of interactive sessions (telegram topics, DM sessions).
## Proposed... | open | null | false | 4 | [] | [] | 2026-03-20T06:49:13Z | 2026-05-04T21:59:24Z | null | NONE | null | 20260505T000034Z | 2026-05-05T00:00:34Z | data-t3labs | 264,575,540 | U_kgDOD8UaNA | User | false |
openclaw/openclaw | 4,076,473,003 | I_kwDOQb6kR87y-gqr | 46,514 | https://github.com/openclaw/openclaw/issues/46514 | https://api.github.com/repos/openclaw/openclaw/issues/46514 | Cross-agent channel history: agents on separate bot accounts can't see each other's session history | ## Problem
When multiple OpenClaw agents are deployed on separate Discord bot accounts in the same guild channel, each agent only sees messages routed to its own session. Agent A cannot see Agent B's replies in its history context, and vice versa — even when `allowBots: "mentions"` is configured.
Each agent has an is... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-14T18:52:01Z | 2026-05-02T04:29:17Z | 2026-04-29T05:46:27Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | cam-ross | 239,492,553 | U_kgDODkZdyQ | User | false |
openclaw/openclaw | 4,106,928,515 | I_kwDOQb6kR870ysGD | 50,972 | https://github.com/openclaw/openclaw/issues/50972 | https://api.github.com/repos/openclaw/openclaw/issues/50972 | gh-issues SKILL.md exceeds 500-line best practice (865 lines) | ## Summary
The `gh-issues` skill's SKILL.md is currently **865 lines**, significantly exceeding the recommended 500-line limit from Anthropic's "Lessons from Building Claude Code" best practices.
## Why this matters
- Long SKILL.md files consume significant context window when loaded
- AI attention distribution degr... | open | null | false | 3 | [] | [] | 2026-03-20T09:54:44Z | 2026-05-04T20:23:07Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | ryunosuke-tanaka | 266,285,851 | U_kgDOD98zGw | User | false |
openclaw/openclaw | 4,111,579,816 | I_kwDOQb6kR871Ebqo | 51,455 | https://github.com/openclaw/openclaw/issues/51455 | https://api.github.com/repos/openclaw/openclaw/issues/51455 | bug: pdf extraction fallback renderer omits standardFontDataUrl and emits PDF.js warnings in Node | ## Summary
The PDF extraction-fallback path for non-native PDF analysis can emit repeated PDF.js warnings during page-image rendering in Node:
`UnknownErrorException: Ensure that the standardFontDataUrl API parameter is provided.`
The analysis can still succeed, but the fallback renderer is missing the standard-font... | closed | completed | true | 5 | [] | [] | 2026-03-21T06:11:53Z | 2026-05-03T04:38:05Z | 2026-04-30T07:38:50Z | CONTRIBUTOR | null | 20260503T060030Z | 2026-05-03T06:00:30Z | anyech | 8,743,351 | MDQ6VXNlcjg3NDMzNTE= | User | false |
openclaw/openclaw | 4,079,573,361 | I_kwDOQb6kR87zKVlx | 47,707 | https://github.com/openclaw/openclaw/issues/47707 | https://api.github.com/repos/openclaw/openclaw/issues/47707 | [Bug]: OpenRouter Mistral models fail with strict tool call ID requirements | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When using Mistral models through OpenRouter (e.g., `mistralai/devstral-2512:free`), the tool call IDs are not sanitized to meet Mistral's `strict9` format requirement (alphanumeric only, length 9). This causes failures when the Mistral bac... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-16T00:27:56Z | 2026-05-02T04:29:16Z | 2026-04-29T05:46:44Z | CONTRIBUTOR | null | 20260502T060142Z | 2026-05-02T06:01:42Z | joelnishanth | 140,015,627 | U_kgDOCFh4Cw | User | false |
openclaw/openclaw | 4,081,123,905 | I_kwDOQb6kR87zQQJB | 48,067 | https://github.com/openclaw/openclaw/issues/48067 | https://api.github.com/repos/openclaw/openclaw/issues/48067 | TUI status bar shows agents.defaults.model instead of resolved agent model | ## Bug Description
The TUI status bar displays the model from `agents.defaults.model.primary` instead of the resolved model for the active agent/session.
**Example:** Config has:
- `agents.defaults.model.primary`: `anthropic/claude-sonnet-4-6`
- `agents.list[0]` (main agent): `model.primary`: `anthropic/claude-opus-4... | closed | completed | true | 4 | [
"stale"
] | [] | 2026-03-16T08:57:14Z | 2026-05-02T04:29:15Z | 2026-04-29T05:46:50Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | datalantics | 263,014,425 | U_kgDOD61IGQ | User | false |
openclaw/openclaw | 4,098,800,956 | I_kwDOQb6kR870Tr08 | 50,138 | https://github.com/openclaw/openclaw/issues/50138 | https://api.github.com/repos/openclaw/openclaw/issues/50138 | Mattermost channel health monitor fails to detect and recover from silent WebSocket disconnection | ## Bug Description
The Mattermost WebSocket channel can silently disconnect and remain unresponsive for 12+ hours while the gateway process itself stays alive. The channel health monitor fails to detect and recover from this state.
## Environment
- OpenClaw version: 2026.3.13
- OS: macOS 15 (arm64)
- Channel: Matter... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-19T02:16:27Z | 2026-05-02T04:29:15Z | 2026-04-29T05:46:57Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | ThomasChan | 5,715,335 | MDQ6VXNlcjU3MTUzMzU= | User | false |
openclaw/openclaw | 4,099,041,780 | I_kwDOQb6kR870Umn0 | 50,174 | https://github.com/openclaw/openclaw/issues/50174 | https://api.github.com/repos/openclaw/openclaw/issues/50174 | [Bug]: Windows native: Telegram polling stalls every ~107s (UND_ERR_CONNECT_TIMEOUT) + Discord disconnect restarts while gateway stays healthy | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
## Summary
On native Windows, OpenClaw gateway process stays running (`RPC probe: ok`), but channel runtimes become unstable:
- Telegram long polling stalls every ~107s, then forced restart loop
- Discord periodically disconnects and gets h... | closed | completed | true | 3 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-19T03:26:36Z | 2026-05-02T04:29:14Z | 2026-04-29T05:47:02Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | sunngaow | 230,383,917 | U_kgDODbthLQ | User | false |
openclaw/openclaw | 4,102,138,247 | I_kwDOQb6kR870gamH | 50,472 | https://github.com/openclaw/openclaw/issues/50472 | https://api.github.com/repos/openclaw/openclaw/issues/50472 | [Windows] Gateway exits silently without crash log | ### Bug type
Regression (worked before, now fails)
### Summary
Gateway on Windows exits silently without any crash log or exit reason. No JavaScript error, stack trace, or uncaught exception is logged before the process terminates.
### Steps to reproduce
1. Start OpenClaw Gateway on Windows 11
2. Let it run for a ... | closed | completed | true | 7 | [
"bug",
"stale",
"regression"
] | [] | 2026-03-19T14:34:48Z | 2026-05-02T04:29:13Z | 2026-04-29T05:47:08Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | chinazll | 266,670,425 | U_kgDOD-URWQ | User | false |
openclaw/openclaw | 4,103,812,739 | I_kwDOQb6kR870mzaD | 50,600 | https://github.com/openclaw/openclaw/issues/50600 | https://api.github.com/repos/openclaw/openclaw/issues/50600 | Feature: configurable busy acknowledgment when session lane is occupied | ## Problem
When a message arrives while the session lane is occupied (agent mid-turn), it's silently queued. The sender gets no feedback that their message was received and is waiting for processing. If the current turn is long-running, this feels like the agent is ignoring them.
When the queue eventually drains via ... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-19T19:35:57Z | 2026-05-02T04:29:13Z | 2026-04-29T05:47:13Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | anders-stewart | 262,437,782 | U_kgDOD6R7lg | User | false |
openclaw/openclaw | 4,118,263,918 | I_kwDOQb6kR871d7hu | 52,640 | https://github.com/openclaw/openclaw/issues/52640 | https://api.github.com/repos/openclaw/openclaw/issues/52640 | Feature request: Persistent task-status surface for long-running channel turns | ## Summary
Add a first-class persistent task-status surface for long-running channel turns (Discord first, generic abstraction later).
Current typing indicators, partial streaming, narration, and heartbeat-style updates help, but they do not provide a single authoritative status surface that tells the user whether a ... | open | null | false | 6 | [] | [] | 2026-03-23T05:04:11Z | 2026-05-01T21:59:23Z | null | CONTRIBUTOR | null | 20260502T000037Z | 2026-05-02T00:00:37Z | anyech | 8,743,351 | MDQ6VXNlcjg3NDMzNTE= | User | false |
openclaw/openclaw | 4,105,962,192 | I_kwDOQb6kR870vALQ | 50,876 | https://github.com/openclaw/openclaw/issues/50876 | https://api.github.com/repos/openclaw/openclaw/issues/50876 | Feature: bootstrapFiles config key to extend workspace injection set | ## Problem
The workspace bootstrap injection set is hardcoded (AGENTS.md, SOUL.md, TOOLS.md, USER.md, MEMORY.md, HEARTBEAT.md). There is no config option to add custom files to the injection set.
This creates a bootstrapping problem for files like RULES.md: critical rules that belong in a dedicated file can't be inje... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-20T05:54:12Z | 2026-05-02T04:29:12Z | 2026-04-29T05:47:23Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | data-t3labs | 264,575,540 | U_kgDOD8UaNA | User | false |
openclaw/openclaw | 4,132,486,637 | I_kwDOQb6kR872UL3t | 54,242 | https://github.com/openclaw/openclaw/issues/54242 | https://api.github.com/repos/openclaw/openclaw/issues/54242 | [Bug]: backup archives can contain hardlink targets with '..' that fail broad extraction on macOS tar | ## Summary
`openclaw backup create` can produce archives that verify successfully but fail broad extraction on macOS `bsdtar` because at least one archived hardlink entry has a `..` target path.
I hit this during a real restore drill on macOS while testing catastrophic-loss recovery for `~/.openclaw`.
## What happen... | open | null | false | 3 | [] | [] | 2026-03-25T04:55:04Z | 2026-05-03T06:49:18Z | null | CONTRIBUTOR | null | 20260503T090142Z | 2026-05-03T09:01:42Z | amzzzzzzz | 154,392,693 | U_kgDOCTPYdQ | User | false |
openclaw/openclaw | 4,214,021,242 | I_kwDOQb6kR877LNx6 | 62,120 | https://github.com/openclaw/openclaw/issues/62120 | https://api.github.com/repos/openclaw/openclaw/issues/62120 | [Bug] openclaw-weixin login hangs before QR code appears on OpenClaw 2026.4.5 | ### Summary
On OpenClaw `2026.4.5 (3e72c03)`, Weixin plugin install/login gets stuck before the QR code is shown.
The plugin install flow prints:
- `插件就绪，开始首次连接...`
But the actual host command:
- `openclaw channels login --channel openclaw-weixin --verbose`
hangs before the Weixin plugin login logic starts, so no... | open | null | false | 2 | [] | [] | 2026-04-06T20:54:53Z | 2026-05-02T05:37:32Z | null | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | annanygy | 274,110,558 | U_kgDOEFaYXg | User | false |
openclaw/openclaw | 4,348,217,376 | I_kwDOQb6kR88AAAABAyyIIA | 74,108 | https://github.com/openclaw/openclaw/issues/74108 | https://api.github.com/repos/openclaw/openclaw/issues/74108 | QQBot: /new command fails with 'messages must not be empty' error |
## Bug Description
When sending `/new` command in QQBot private chat, the bot returns:
```
⚠️ Something went wrong while processing your request. Please try again, or use /new to start a fresh session.
```
## Steps to Reproduce
1. Open QQBot private chat
2. Send any message (e.g., "1") - bot responds normally
3. Sen... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T05:45:29Z | 2026-05-02T04:29:11Z | 2026-04-29T05:50:26Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | Zn-Zero | 51,368,854 | MDQ6VXNlcjUxMzY4ODU0 | User | false |
openclaw/openclaw | 4,348,233,792 | I_kwDOQb6kR88AAAABAyzIQA | 74,111 | https://github.com/openclaw/openclaw/issues/74111 | https://api.github.com/repos/openclaw/openclaw/issues/74111 | BUG: media-understanding-core cannot resolve staged sharp dependency for image processing | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Image processing failed with `Failed to optimize image` because `media-understanding-core` could not resolve its required `sharp` dependency. The package declares `sharp` and `stageRuntimeDependencies`, but the... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-29T05:49:12Z | 2026-05-02T04:29:10Z | 2026-04-29T05:54:07Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | kAIborg24 | 261,855,322 | U_kgDOD5uYWg | User | false |
openclaw/openclaw | 4,348,103,885 | I_kwDOQb6kR88AAAABAyrMzQ | 74,093 | https://github.com/openclaw/openclaw/issues/74093 | https://api.github.com/repos/openclaw/openclaw/issues/74093 | Bug: /new reset can be queued behind active run when queue mode is steer | ### Summary
When `messages.queue.mode` is `steer`, a reset command such as `/new` (via Feishu DM) appears to go through the normal active-run queue/steer path instead of interrupting the current run. The result is that the reset is acknowledged by ingress immediately, but the new session/reply can be delayed until the... | closed | completed | true | 4 | [] | [] | 2026-04-29T05:19:27Z | 2026-05-05T04:25:55Z | 2026-05-02T23:54:46Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | ruji9527 | 210,522,954 | U_kgDODIxTSg | User | false |
openclaw/openclaw | 4,336,885,383 | I_kwDOQb6kR88AAAABAn-ehw | 72,876 | https://github.com/openclaw/openclaw/issues/72876 | https://api.github.com/repos/openclaw/openclaw/issues/72876 | [Feature]: Support multiple Microsoft Teams bot accounts in one gateway | ### Summary
Allow the Microsoft Teams channel to define multiple bot/app accounts, similar to Discord accounts, so one OpenClaw gateway can route different Teams apps to different agents.
### Problem to solve
The current Microsoft Teams channel appears to support only one top-level `channels.msteams` config. This m... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-27T15:26:01Z | 2026-05-02T04:29:09Z | 2026-04-29T05:54:53Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | SimSef | 56,121,054 | MDQ6VXNlcjU2MTIxMDU0 | User | false |
openclaw/openclaw | 4,348,268,685 | I_kwDOQb6kR88AAAABAy1QjQ | 74,119 | https://github.com/openclaw/openclaw/issues/74119 | https://api.github.com/repos/openclaw/openclaw/issues/74119 | [Bug]: Telegram DM startup/runtime context and async follow-up scaffolding can leak into user-visible messages | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Fresh Telegram DM sessions and some async follow-up turns can surface internal runtime/startup context as user-visible assistant messages instead of keeping it internal.
### Steps to reproduce
1. Use OpenClaw... | closed | completed | true | 1 | [] | [] | 2026-04-29T05:57:22Z | 2026-05-02T04:29:08Z | 2026-04-29T06:03:34Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | gkoch02 | 76,127,910 | MDQ6VXNlcjc2MTI3OTEw | User | false |
openclaw/openclaw | 4,348,261,251 | I_kwDOQb6kR88AAAABAy0zgw | 74,115 | https://github.com/openclaw/openclaw/issues/74115 | https://api.github.com/repos/openclaw/openclaw/issues/74115 | [Bug]: IT SHOULD NOT BE REQUIRED TO set up all settings before gateway and WEBUI starts. This is unwise design practice. | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
the webui and any other system related to that should not have to rely on needing every single config set up just to be able to run, also, it should not be required to set up models before the webui works, it should not crash w... | open | null | false | 2 | [
"bug",
"bug:crash"
] | [] | 2026-04-29T05:55:43Z | 2026-05-03T01:23:09Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | cdvankammen | 31,197,020 | MDQ6VXNlcjMxMTk3MDIw | User | false |
openclaw/openclaw | 4,348,264,112 | I_kwDOQb6kR88AAAABAy0-sA | 74,117 | https://github.com/openclaw/openclaw/issues/74117 | https://api.github.com/repos/openclaw/openclaw/issues/74117 | ensureRuntimePluginsLoaded cache-misses every inbound dispatch (~5-6s wasted per message) | ### Bug type
Behavior bug (incorrect output/state without crash) — performance.
### Beta release blocker
No
### Summary
`ensureRuntimePluginsLoaded` (`src/agents/runtime-plugins.ts:6`) cache-misses on every inbound message dispatch, triggering a full `loadOpenClawPlugins` rebuild and re-running every plugin's `regist... | closed | completed | true | 2 | [] | [] | 2026-04-29T05:56:17Z | 2026-05-05T04:27:10Z | 2026-05-02T13:42:47Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | poolside-ventures | 215,486,291 | U_kgDODNgPUw | User | false |
openclaw/openclaw | 4,348,286,009 | I_kwDOQb6kR88AAAABAy2UOQ | 74,122 | https://github.com/openclaw/openclaw/issues/74122 | https://api.github.com/repos/openclaw/openclaw/issues/74122 | Watcher should debounce rapid multi-file edits into a single reload | # Watcher: rapid multi-file edits trigger multiple full plugin reloads instead of one coalesced reload
## Summary
Editing several watched files within a short window (e.g. `openclaw.json` plus one or two other watched files) triggers a separate full plugin reload for **each** save. There is currently no:
1. **Deboun... | open | null | false | 1 | [] | [] | 2026-04-29T06:01:10Z | 2026-05-03T09:49:34Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | kenneth2light-bot | 261,371,820 | U_kgDOD5Q3rA | User | false |
openclaw/openclaw | 4,348,299,047 | I_kwDOQb6kR88AAAABAy3HJw | 74,125 | https://github.com/openclaw/openclaw/issues/74125 | https://api.github.com/repos/openclaw/openclaw/issues/74125 | Gateway-restart wake-ups for Claude CLI Discord sessions arrive cold — neither --resume nor openClawHistoryPrompt rebuild fires | ## Summary
After a gateway restart, queued wake-up / system-event turns delivered to a Claude-CLI-bound Discord session arrive **without prior conversational context**. Normal user-message turns on the same session post-restart DO retain context, so this is specific to the wake-up / system-event path.
The result is t... | closed | completed | true | 1 | [] | [] | 2026-04-29T06:03:57Z | 2026-05-02T04:28:27Z | 2026-04-29T11:14:50Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | aounakram | 57,077,179 | MDQ6VXNlcjU3MDc3MTc5 | User | false |
openclaw/openclaw | 4,348,294,857 | I_kwDOQb6kR88AAAABAy22yQ | 74,124 | https://github.com/openclaw/openclaw/issues/74124 | https://api.github.com/repos/openclaw/openclaw/issues/74124 | [Feature]: Support global skill installation target | ### Summary
Add a way for `openclaw skills install` to install skills into the global skills directory, e.g. `~/.openclaw/skills` / `~/.agents/skills`, instead of only installing into an agent workspace such as `~/.openclaw/workspace-<agent>/skills`.
### Problem to solve
Running:
```sh
openclaw skills install <skil... | open | null | false | 1 | [] | [] | 2026-04-29T06:03:01Z | 2026-05-03T08:26:00Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | abnershang | 75,654,486 | MDQ6VXNlcjc1NjU0NDg2 | User | false |
openclaw/openclaw | 4,348,288,193 | I_kwDOQb6kR88AAAABAy2cwQ | 74,123 | https://github.com/openclaw/openclaw/issues/74123 | https://api.github.com/repos/openclaw/openclaw/issues/74123 | BUG: media:// image URI resolves to invalid workspace media path | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Passing an inbound image as a `media://...` URI to the image tool fails because the URI is resolved to an invalid local path containing `media:/inbound/...` under the workspace. The same image succeeds when pas... | open | null | false | 1 | [
"bug"
] | [] | 2026-04-29T06:01:36Z | 2026-05-03T10:19:53Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | kAIborg24 | 261,855,322 | U_kgDOD5uYWg | User | false |
openclaw/openclaw | 4,348,322,680 | I_kwDOQb6kR88AAAABAy4jeA | 74,128 | https://github.com/openclaw/openclaw/issues/74128 | https://api.github.com/repos/openclaw/openclaw/issues/74128 | [Bug]: Gateway boot still blocks ~75s on pricing fetch despite fix 7d9dc8cf24 (macOS, 2026.4.26 be8c246, Node 25.9.0) | ### Bug type
Regression after partial fix
### Summary
Gateway boot blocks ~75 seconds waiting for `OpenRouter pricing fetch` and `LiteLLM pricing fetch` to time out at 60s, despite running on a build that contains the fix from #72971 (commit `7d9dc8cf24`, "reuse plugin manifests for model pricing refresh"). Standalone... | closed | completed | true | 2 | [] | [] | 2026-04-29T06:09:20Z | 2026-05-05T04:27:46Z | 2026-05-02T10:08:19Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | ctbritt | 4,516,034 | MDQ6VXNlcjQ1MTYwMzQ= | User | false |
openclaw/openclaw | 4,348,384,116 | I_kwDOQb6kR88AAAABAy8TdA | 74,141 | https://github.com/openclaw/openclaw/issues/74141 | https://api.github.com/repos/openclaw/openclaw/issues/74141 | session_status throws "Unknown sessionKey: current" for channel-plugin agents (~4s wasted per turn) | ### Bug type
Behavior bug (incorrect output/state without crash) — performance + UX (~4s wasted per affected turn, agent visibly retries).
### Beta release blocker
No
### Summary
The `session_status` tool's docstring (`tool-description-presets.ts:63`) advises agents to pass `sessionKey="current"` universally, but the... | closed | completed | true | 1 | [] | [] | 2026-04-29T06:24:54Z | 2026-05-05T04:27:37Z | 2026-05-02T10:59:01Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | poolside-ventures | 215,486,291 | U_kgDODNgPUw | User | false |
openclaw/openclaw | 4,348,363,810 | I_kwDOQb6kR88AAAABAy7EIg | 74,138 | https://github.com/openclaw/openclaw/issues/74138 | https://api.github.com/repos/openclaw/openclaw/issues/74138 | [Bug]: feishu createScopedPairingAccess | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
Yes
### Summary
After upgrade to latest version, my feishu channel broken.My message can be processed.even if revert the change of the channel config
### Steps to reproduce
1.upgrade to latest version
2.config multi accounts
### Expecte... | open | null | false | 1 | [
"bug"
] | [] | 2026-04-29T06:19:39Z | 2026-05-03T05:00:04Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | dkisser | 24,600,271 | MDQ6VXNlcjI0NjAwMjcx | User | false |
openclaw/openclaw | 4,348,414,971 | I_kwDOQb6kR88AAAABAy-L-w | 74,149 | https://github.com/openclaw/openclaw/issues/74149 | https://api.github.com/repos/openclaw/openclaw/issues/74149 | Bug: Control UI sends unsupported 'view' parameter to models.list, causing INVALID_REQUEST in v2026.4.26 | ## Describe the bug
The Control UI sends `{view: "configured"}` as params to `models.list`, but the Gateway's `ModelsListParamsSchema` is defined as `Type.Object({}, { additionalProperties: false })` which rejects any parameters. This causes repeated `INVALID_REQUEST` errors and the model selector dropdown in the Cont... | closed | completed | true | 2 | [] | [
"vincentkoc"
] | 2026-04-29T06:32:57Z | 2026-05-02T04:28:15Z | 2026-04-29T13:09:14Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | wangk123 | 20,060,636 | MDQ6VXNlcjIwMDYwNjM2 | User | false |
openclaw/openclaw | 4,348,363,760 | I_kwDOQb6kR88AAAABAy7D8A | 74,137 | https://github.com/openclaw/openclaw/issues/74137 | https://api.github.com/repos/openclaw/openclaw/issues/74137 | [Bug] Telegram channel occasionally injects blank user messages, causing API errors | ## Bug Description
Telegram channel occasionally injects empty/blank user messages into the session, which causes the LLM provider API to reject the request with an error like:
```
One of "input" or "previous_response_id" or "prompt" or "conversation_id" must be provided.
```
## Environment
- **OS**: macOS (arm64),... | closed | completed | true | 5 | [] | [] | 2026-04-29T06:19:39Z | 2026-05-03T04:36:12Z | 2026-04-30T15:48:48Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Gracker | 2,181,004 | MDQ6VXNlcjIxODEwMDQ= | User | false |
openclaw/openclaw | 4,348,421,595 | I_kwDOQb6kR88AAAABAy-l2w | 74,154 | https://github.com/openclaw/openclaw/issues/74154 | https://api.github.com/repos/openclaw/openclaw/issues/74154 | [Bug]: Session stuck + High memory on v2026.4.26 — works fine on 4.15 | ## Bug Description
After upgrading from 4.15 to 4.26, gateway becomes unresponsive after a short period. Session locks up in \"processing\" state, Telegram polling stalls, and memory usage jumps from ~40MB to 1.2GB.
## Environment
- **Version**: 2026.4.26 (macOS, npm global install)
- **Previous version**: 2026.4.15... | closed | not_planned | true | 2 | [] | [] | 2026-04-29T06:34:45Z | 2026-05-02T04:29:07Z | 2026-04-29T06:38:32Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | kyle20026 | 274,152,656 | U_kgDOEFc80A | User | false |
openclaw/openclaw | 4,348,419,856 | I_kwDOQb6kR88AAAABAy-fEA | 74,153 | https://github.com/openclaw/openclaw/issues/74153 | https://api.github.com/repos/openclaw/openclaw/issues/74153 | [Bug]: Session stuck + High memory on v2026.4.26 — works fine on 4.15 | ## Bug Description
After upgrading from 4.15 to 4.26, gateway becomes unresponsive after a short period. Session locks up in \"processing\" state, Telegram polling stalls, and memory usage jumps from ~40MB to 1.2GB.
## Environment
- **Version**: 2026.4.26 (macOS, npm global install)
- **Previous version**: 2026.4.15... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T06:34:17Z | 2026-05-02T04:29:06Z | 2026-04-29T06:38:52Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | kyle20026 | 274,152,656 | U_kgDOEFc80A | User | false |
openclaw/openclaw | 4,348,612,158 | I_kwDOQb6kR88AAAABAzKOPg | 74,178 | https://github.com/openclaw/openclaw/issues/74178 | https://api.github.com/repos/openclaw/openclaw/issues/74178 | [Bug]: Gateway startup pricing fetch fails despite network being available — OpenRouter TypeError + LiteLLM 60s timeout on Windows 11 | ## Summary
Gateway 启动时，model pricing fetch 总是失败，导致启动变慢。OpenRouter 报 `TypeError: fetch failed`，LiteLLM 报 `TimeoutError: 60s timeout`。但**本地网络和 openclaw agent 都能正常访问这些端点**——说明不是网络不可达的问题，而是 gateway 进程内的 fetch 存在某种特殊限制。
## Environment
- OS: Windows 11 Pro (China Only), 10.0.26200
- Node: (Node 24.x)
- openclaw version: 2... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T07:16:09Z | 2026-05-02T04:29:00Z | 2026-04-29T07:21:18Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | stayif | 7,867,389 | MDQ6VXNlcjc4NjczODk= | User | false |
openclaw/openclaw | 4,348,625,679 | I_kwDOQb6kR88AAAABAzLDDw | 74,183 | https://github.com/openclaw/openclaw/issues/74183 | https://api.github.com/repos/openclaw/openclaw/issues/74183 | [Feature]: MCP bundle materializer needs exponential backoff for rate-limited server init | ## Feature Request
OpenClaw's MCP bundle materializer (`pi-bundle-mcp-tools.ts`) does not implement exponential backoff when MCP server initialization fails with rate-limit errors. This causes rapid-fire retry loops that burn through rate-limit quotas.
## Problem
When a streamable HTTP MCP server (e.g. Cloudflare MC... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T07:18:52Z | 2026-05-03T04:35:03Z | 2026-04-30T20:37:52Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | kai-the-man[bot] | 259,478,402 | BOT_kgDOD3dTgg | Bot | false |
openclaw/openclaw | 4,348,625,606 | I_kwDOQb6kR88AAAABAzLCxg | 74,182 | https://github.com/openclaw/openclaw/issues/74182 | https://api.github.com/repos/openclaw/openclaw/issues/74182 | Embedded LLM run objects retain memory after timeout/abort — causes multi-GB RSS growth in single-process Gateway | After running the OpenClaw Gateway for ~2 hours with embedded LLM runs, the process RSS grew from a normal baseline to **2.85GB**, causing event loop saturation (100% CPU, 5s+ RPC response times).
The root cause appears to be **incomplete cleanup of embedded LLM run objects after timeout**. When an embedded run times ... | closed | completed | true | 4 | [] | [] | 2026-04-29T07:18:51Z | 2026-05-03T04:33:59Z | 2026-05-01T04:24:14Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | cjboy007 | 64,246,343 | MDQ6VXNlcjY0MjQ2MzQz | User | false |
openclaw/openclaw | 4,348,387,144 | I_kwDOQb6kR88AAAABAy8fSA | 74,143 | https://github.com/openclaw/openclaw/issues/74143 | https://api.github.com/repos/openclaw/openclaw/issues/74143 | feishu: streaming card duplicates content when block payloads overlap with partial snapshots | # Feishu streaming card duplicates content when block payloads overlap with partial snapshots
## Summary
When a model emits content that arrives via both `onPartialReply` (snapshot mode) **and** `deliver({kind: "block"})` (delta mode), the streaming card UI shows the same text repeated multiple times. The bug surface... | open | null | false | 2 | [] | [] | 2026-04-29T06:25:36Z | 2026-05-03T08:26:02Z | null | CONTRIBUTOR | null | 20260503T090142Z | 2026-05-03T09:01:42Z | ottodeng | 20,509,364 | MDQ6VXNlcjIwNTA5MzY0 | User | false |
openclaw/openclaw | 4,348,343,209 | I_kwDOQb6kR88AAAABAy5zqQ | 74,135 | https://github.com/openclaw/openclaw/issues/74135 | https://api.github.com/repos/openclaw/openclaw/issues/74135 | Gateway intermittently stalls: WebSocket preauth handshakes time out late during model catalog/provider discovery | # Gateway intermittently stalls: WebSocket preauth handshakes time out late during model catalog/provider discovery
## Summary
The OpenClaw Gateway intermittently becomes unresponsive for tens of seconds. During these windows, WebSocket preauth handshakes time out, but the timeout callbacks fire much later than the c... | closed | completed | true | 5 | [] | [] | 2026-04-29T06:14:16Z | 2026-05-02T04:26:36Z | 2026-04-30T02:35:24Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | DerFlash | 185,013 | MDQ6VXNlcjE4NTAxMw== | User | false |
openclaw/openclaw | 4,348,624,973 | I_kwDOQb6kR88AAAABAzLATQ | 74,181 | https://github.com/openclaw/openclaw/issues/74181 | https://api.github.com/repos/openclaw/openclaw/issues/74181 | [Bug]: Session compaction deadlock blocks child subagent SessionManager.open() | ## Bug Description
When a session reaches the compaction threshold (~200k tokens), the parent agent's compaction process acquires a **WRITE lock** on the session file. If a child subagent spawn is attempted simultaneously, `SessionManager.open()` tries to acquire a **READ lock** on the same file. The child blocks inde... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T07:18:43Z | 2026-05-02T04:28:59Z | 2026-04-29T07:37:04Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | kai-the-man[bot] | 259,478,402 | BOT_kgDOD3dTgg | Bot | false |
openclaw/openclaw | 4,348,751,144 | I_kwDOQb6kR88AAAABAzStKA | 74,192 | https://github.com/openclaw/openclaw/issues/74192 | https://api.github.com/repos/openclaw/openclaw/issues/74192 | [Bug]: Delay in provider/transport path | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After updating to 2026.4.24 there was a huge delay introduced from the moment the message is sent till the moment the actual reply from the agent shows up. I have tried to reinstall OpenClaw before wiping all the files (clea... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-04-29T07:44:45Z | 2026-05-03T04:35:21Z | 2026-04-30T19:19:09Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | karlis0 | 39,435,777 | MDQ6VXNlcjM5NDM1Nzc3 | User | false |
openclaw/openclaw | 4,348,803,477 | I_kwDOQb6kR88AAAABAzV5lQ | 74,197 | https://github.com/openclaw/openclaw/issues/74197 | https://api.github.com/repos/openclaw/openclaw/issues/74197 | [Bug]: qmd collection add hangs under gateway init, triggers SIGKILL restart loop | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
On gateway start, `qmd collection add ... --name memory-root-main --mask MEMORY.md` can hang past `TimeoutStopSec=30`, causing systemd to SIGKILL the gateway and enter an auto-restart loop. The same `qmd collection add` invocat... | closed | completed | true | 1 | [] | [] | 2026-04-29T07:55:17Z | 2026-05-02T04:28:59Z | 2026-04-29T07:59:47Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | maxpuppet | 131,998,469 | U_kgDOB94jBQ | User | false |
openclaw/openclaw | 4,348,806,060 | I_kwDOQb6kR88AAAABAzWDrA | 74,198 | https://github.com/openclaw/openclaw/issues/74198 | https://api.github.com/repos/openclaw/openclaw/issues/74198 | Memory flush write paths should be configurable (currently hardcoded to single daily log) | # OpenClaw Runtime: Memory Flush Write Paths Should Be Configurable
## Problem
The memory flush mechanism (triggered when context window fills) currently restricts `write` tool calls to a single path: `memory/YYYY-MM-DD.md`. This is hardcoded in the binary at `openclaw-tools-CUmYpN1l.js:775`.
The system prompt (`AGEN... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T07:55:48Z | 2026-05-02T04:28:58Z | 2026-04-29T07:59:58Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | joelrendall | 8,376,419 | MDQ6VXNlcjgzNzY0MTk= | User | false |
openclaw/openclaw | 4,348,806,638 | I_kwDOQb6kR88AAAABAzWF7g | 74,199 | https://github.com/openclaw/openclaw/issues/74199 | https://api.github.com/repos/openclaw/openclaw/issues/74199 | [Bug]: runtime-deps manifest still missing 9 dist bare-imports in 2026.4.26 (follow-up to #71420) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The runtime-deps mirror manifest at `~/.openclaw/plugin-runtime-deps/openclaw-<version>/.openclaw-runtime-deps.json` is missing 9 packages that the dist code does `import x from 'pkg'` for. They never get insta... | closed | completed | true | 5 | [] | [] | 2026-04-29T07:55:56Z | 2026-05-02T04:26:29Z | 2026-04-30T02:52:22Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | maxpuppet | 131,998,469 | U_kgDOB94jBQ | User | false |
openclaw/openclaw | 4,348,784,023 | I_kwDOQb6kR88AAAABAzUtlw | 74,195 | https://github.com/openclaw/openclaw/issues/74195 | https://api.github.com/repos/openclaw/openclaw/issues/74195 | CLI text commands hang with no output while --json variants work (agents list/status) | ## Summary
Several OpenClaw CLI commands hang indefinitely in their default text output path with no stdout/stderr, while their `--json` variants complete successfully. The gateway is healthy and responsive, so this appears to be a CLI text/bootstrap/rendering path issue rather than a gateway availability issue.
This... | closed | completed | true | 2 | [] | [] | 2026-04-29T07:51:09Z | 2026-05-03T04:38:12Z | 2026-04-30T06:36:25Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | kirmak81-spec | 235,636,757 | U_kgDODguIFQ | User | false |
openclaw/openclaw | 4,348,834,237 | I_kwDOQb6kR88AAAABAzXxvQ | 74,203 | https://github.com/openclaw/openclaw/issues/74203 | https://api.github.com/repos/openclaw/openclaw/issues/74203 | [Bug]: Telegram safe-send retry misses grammY Network request failed envelope | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
OpenClaw 2026.4.26 still treats grammY `HttpError: Network request for 'sendMessage' failed!` as a terminal Telegram delivery failure in the final reply path, even though this is a transient pre-response networ... | closed | completed | true | 2 | [] | [] | 2026-04-29T08:00:33Z | 2026-05-02T04:28:20Z | 2026-04-29T11:43:25Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | nanli2000cn | 259,904,494 | U_kgDOD33T7g | User | false |
openclaw/openclaw | 4,348,856,537 | I_kwDOQb6kR88AAAABAzZI2Q | 74,204 | https://github.com/openclaw/openclaw/issues/74204 | https://api.github.com/repos/openclaw/openclaw/issues/74204 | memory.qmd.update.embedTimeoutMs default (120 s) is too low for local GGUF; error message doesn't surface the fix | ## Environment
- **OpenClaw version:** 2026.4.25
- **QMD version:** 2.1.0 (bab86d5)
- **Platform:** Ubuntu 24.04, GCP e2-standard-4 (4 vCPU, 16 GB RAM)
- **Workspace size:** 37+ Markdown files in memory root + `memory/` tree
- **GGUF model:** default (`embeddinggemma-300m-qat-Q8_0.gguf`, ~0.6 GB, auto-downloaded)
- **... | open | null | false | 1 | [] | [] | 2026-04-29T08:04:35Z | 2026-05-03T08:26:04Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Skeptomenos | 22,348,986 | MDQ6VXNlcjIyMzQ4OTg2 | User | false |
openclaw/openclaw | 4,348,881,001 | I_kwDOQb6kR88AAAABAzaoaQ | 74,206 | https://github.com/openclaw/openclaw/issues/74206 | https://api.github.com/repos/openclaw/openclaw/issues/74206 | MiniMax API error 2013 'chat content is empty' on new session (v2026.4.26, regressed from v2026.4.23) | ## Bug Description
**Affected versions**: v2026.4.25-beta.10 through v2026.4.26 (v2026.4.23 is NOT affected)
When a user sends a message in Feishu that triggers a new session (via /new or session reset), OpenClaw sends an empty user message content to MiniMax API, which rejects it with:
```
400 invalid params, chat ... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T08:09:28Z | 2026-05-02T04:28:57Z | 2026-04-29T08:13:31Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | Huolele | 131,687,351 | U_kgDOB9ljtw | User | false |
openclaw/openclaw | 4,348,859,285 | I_kwDOQb6kR88AAAABAzZTlQ | 74,205 | https://github.com/openclaw/openclaw/issues/74205 | https://api.github.com/repos/openclaw/openclaw/issues/74205 | [Bug]: kimi/k2p6 returns incomplete terminal response and gateway status probe times out | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
OpenClaw 2026.4.26 repeatedly fails turns using `kimi/k2p6` with `ended with an incomplete terminal response`, and `openclaw gateway status` times out while the gateway service is running.
### Steps to reproduce
Steps to... | open | null | false | 2 | [
"bug",
"regression"
] | [] | 2026-04-29T08:05:03Z | 2026-05-03T10:41:43Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | jondeaf | 194,342,313 | U_kgDOC5VtqQ | User | false |
openclaw/openclaw | 4,348,931,238 | I_kwDOQb6kR88AAAABAzdspg | 74,208 | https://github.com/openclaw/openclaw/issues/74208 | https://api.github.com/repos/openclaw/openclaw/issues/74208 | [Bug]: The 2026.4.26 version has difficulties in using the local model. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The 2026.4.26 version has difficulties in using the local model.(run by llama.cpp) The first conversation takes 7 minutes to access the local large model, and it often gets disconnected. The 2026.4.15 version can invoke the ... | closed | not_planned | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-29T08:18:08Z | 2026-05-02T04:28:54Z | 2026-04-29T08:22:55Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | kinlioncd | 83,751,473 | MDQ6VXNlcjgzNzUxNDcz | User | false |
openclaw/openclaw | 4,348,958,904 | I_kwDOQb6kR88AAAABAzfYuA | 74,209 | https://github.com/openclaw/openclaw/issues/74209 | https://api.github.com/repos/openclaw/openclaw/issues/74209 | [Bug]: Default bundled plugins, especially bonjour, can block gateway startup after 2026.4.26 upgrade | ## Bug type
Regression (worked before, now fails)
## Beta release blocker
No
## Summary
After upgrading a Linux OpenClaw installation that previously worked on 2026.4.2 to the 2026.4.26 plugin set, the gateway became effectively unusable: many bundled plugins were enabled/loaded by default, and the `bonjour` advertis... | closed | completed | false | 6 | [] | [] | 2026-04-29T08:23:09Z | 2026-05-03T18:07:57Z | 2026-05-03T18:07:57Z | NONE | null | 20260503T210034Z | 2026-05-03T21:00:34Z | ndj888 | 11,382,964 | MDQ6VXNlcjExMzgyOTY0 | User | false |
openclaw/openclaw | 4,348,449,991 | I_kwDOQb6kR88AAAABAzAUxw | 74,159 | https://github.com/openclaw/openclaw/issues/74159 | https://api.github.com/repos/openclaw/openclaw/issues/74159 | github-copilot: Gemini models route to unsupported /responses endpoint, and openai-completions transport bypasses Copilot IDE auth headers | ## Summary
`github-copilot/gemini-3-flash-preview` (and likely all Gemini/`/chat/completions`-only Copilot models) fail 100% of the time and silently fall back to the configured non-Copilot fallback (e.g. `anthropic/claude-sonnet-4-6`), generating substantial unintended Anthropic API spend.
Two bugs in series in `dis... | open | null | false | 1 | [] | [] | 2026-04-29T06:41:18Z | 2026-05-03T11:25:40Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | xonaman | 45,240,351 | MDQ6VXNlcjQ1MjQwMzUx | User | false |
openclaw/openclaw | 4,348,494,455 | I_kwDOQb6kR88AAAABAzDCdw | 74,164 | https://github.com/openclaw/openclaw/issues/74164 | https://api.github.com/repos/openclaw/openclaw/issues/74164 | [Bug]: OpenClaw send message with big file via wecom show gateway timeout after 10000ms | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
OpenClaw send message with big file via wecom show error: gateway timeout after 10000ms
### Steps to reproduce
Run the following command in the Linux Terminal (the attach file is about 10M) :
openclaw message send --ch... | open | null | false | 1 | [
"bug",
"regression"
] | [] | 2026-04-29T06:51:00Z | 2026-05-03T13:22:07Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | caoyinchuan | 177,899,473 | U_kgDOCpqH0Q | User | false |
openclaw/openclaw | 4,348,503,157 | I_kwDOQb6kR88AAAABAzDkdQ | 74,165 | https://github.com/openclaw/openclaw/issues/74165 | https://api.github.com/repos/openclaw/openclaw/issues/74165 | [Feature]: structured abort logging for embedded run aborts (OPENCLAW_LOG_ABORT_SOURCES env var) | ### Summary
Add an opt-in structured diagnostic for embedded run aborts, gated by `OPENCLAW_LOG_ABORT_SOURCES=1`. When enabled, every `abortRun()` entry in `src/agents/pi-embedded-runner/run/attempt.ts` emits one `[abort-source]` line classifying the trigger and capturing five caller stack frames.
### Problem to solv... | open | null | false | 1 | [] | [] | 2026-04-29T06:52:48Z | 2026-05-03T05:00:06Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | chphch | 19,252,487 | MDQ6VXNlcjE5MjUyNDg3 | User | false |
openclaw/openclaw | 4,348,503,326 | I_kwDOQb6kR88AAAABAzDlHg | 74,166 | https://github.com/openclaw/openclaw/issues/74166 | https://api.github.com/repos/openclaw/openclaw/issues/74166 | [Feature]: configurable gateway chat-delta broadcast throttle (OPENCLAW_CHAT_DELTA_THROTTLE_MS env var) | ### Summary
Make the gateway chat-delta broadcast throttle configurable via `OPENCLAW_CHAT_DELTA_THROTTLE_MS`. Default stays at the existing **150 ms**, so behavior is unchanged unless operators opt in.
### Problem to solve
`emitChatDelta` in `src/gateway/server-chat.ts` rate-limits assistant text broadcasts to one ... | open | null | false | 1 | [] | [] | 2026-04-29T06:52:50Z | 2026-05-03T20:25:13Z | null | NONE | null | 20260503T210034Z | 2026-05-03T21:00:34Z | chphch | 19,252,487 | MDQ6VXNlcjE5MjUyNDg3 | User | false |
openclaw/openclaw | 4,348,981,712 | I_kwDOQb6kR88AAAABAzgx0A | 74,212 | https://github.com/openclaw/openclaw/issues/74212 | https://api.github.com/repos/openclaw/openclaw/issues/74212 | [Bug]: models auth login --provider openai-codex swallows the device-pairing code that upstream codex login prints to the terminal | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
openclaw models auth login --provider openai-codex with Device Pairing Auth selected does not surface the device-pairing code in an SSH session, whereas running upstream codex login directly in the same SSH ses... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-29T08:27:22Z | 2026-05-05T04:27:50Z | 2026-05-02T10:00:27Z | CONTRIBUTOR | null | 20260505T060034Z | 2026-05-05T06:00:34Z | da22le123 | 106,140,419 | U_kgDOBlOTAw | User | false |
openclaw/openclaw | 4,349,016,799 | I_kwDOQb6kR88AAAABAzi63w | 74,217 | https://github.com/openclaw/openclaw/issues/74217 | https://api.github.com/repos/openclaw/openclaw/issues/74217 | Heartbeat model leaks into user session status/runtime model | ## Summary
A heartbeat / async completion run that uses `agents.defaults.heartbeat.model` appears to persist its runtime model into the normal chat session (`sessionEntry.model` / `modelProvider`). After that, `/status` / session status reports the user-facing WhatsApp direct session as running the heartbeat model, ev... | open | null | false | 1 | [] | [] | 2026-04-29T08:33:35Z | 2026-05-03T08:26:08Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Sitham94 | 224,994,873 | U_kgDODWkmOQ | User | false |
openclaw/openclaw | 4,349,025,155 | I_kwDOQb6kR88AAAABAzjbgw | 74,218 | https://github.com/openclaw/openclaw/issues/74218 | https://api.github.com/repos/openclaw/openclaw/issues/74218 | [Bug]: WhatsApp dmPolicy allowlist bypassed during gateway reconnection | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
WhatsApp dmPolicy allowlist is bypassed during gateway reconnection, allowing queued messages from non-allowed contacts to reach the agent.
### Steps to reproduce
1. Configure channels.whatsapp.dmPolicy to "allowlist" with... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-29T08:35:07Z | 2026-05-02T04:28:53Z | 2026-04-29T08:40:17Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | charliehermes31416-create | 280,301,567 | U_kgDOELUP_w | User | false |
openclaw/openclaw | 4,349,052,222 | I_kwDOQb6kR88AAAABAzlFPg | 74,225 | https://github.com/openclaw/openclaw/issues/74225 | https://api.github.com/repos/openclaw/openclaw/issues/74225 | [Bug]: sessions_spawn(runtime:"acp", agentId:"claude") fails with AcpRuntimeError: Internal error — direct acpx claude exec works | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
`sessions_spawn(runtime:"acp", agentId:"claude")` fails with `AcpRuntimeError: Internal error` (~1.3s, 0 tokens processed). Direct `acpx claude exec` CLI on the same machine works fine, confirming the `claude-agen... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T08:40:03Z | 2026-05-02T04:28:52Z | 2026-04-29T08:43:59Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | slideshow-dingo | 258,996,445 | U_kgDOD2_43Q | User | false |
openclaw/openclaw | 4,349,063,626 | I_kwDOQb6kR88AAAABAzlxyg | 74,227 | https://github.com/openclaw/openclaw/issues/74227 | https://api.github.com/repos/openclaw/openclaw/issues/74227 | [Bug] Discord `/new` slash command times out — preflight compaction holds WRITE lock past InteractionEventListener 30s deadline | ## Bug Report
**Title:** `/new` Discord slash command times out due to preflight compaction WRITE lock holding session for 60–360s
---
### Description
When a user invokes `/new` (or `/reset`) via Discord's slash command UI, the InteractionEventListener times out at 30 seconds because `runPreflightCompactionIfNeeded... | closed | completed | true | 1 | [] | [] | 2026-04-29T08:42:01Z | 2026-05-02T04:28:52Z | 2026-04-29T08:48:34Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | slideshow-dingo | 258,996,445 | U_kgDOD2_43Q | User | false |
openclaw/openclaw | 4,349,086,997 | I_kwDOQb6kR88AAAABAznNFQ | 74,229 | https://github.com/openclaw/openclaw/issues/74229 | https://api.github.com/repos/openclaw/openclaw/issues/74229 | Context-engine assembled prompts are bypassed by overflow precheck using pre-assembly history | ## Summary
OpenClaw's `ContextEngine.assemble()` contract says the engine returns ordered messages ready for the model under a token budget, but the embedded runner still hard-gates prompt submission on a larger pre-assembly message view (`unwindowedMessages`). With `lossless-claw`/LCM as the context engine, this can ... | closed | not_planned | true | 4 | [] | [] | 2026-04-29T08:45:35Z | 2026-05-02T04:28:51Z | 2026-04-29T08:50:15Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | electricsheepops | 276,317,342 | U_kgDOEHhEng | User | false |
openclaw/openclaw | 4,349,111,069 | I_kwDOQb6kR88AAAABAzorHQ | 74,233 | https://github.com/openclaw/openclaw/issues/74233 | https://api.github.com/repos/openclaw/openclaw/issues/74233 | Context-engine assembled prompts are bypassed by overflow precheck using pre-assembly history | ## Summary
OpenClaw's `ContextEngine.assemble()` contract says the engine returns ordered messages ready for the model under a token budget, but the embedded runner still hard-gates prompt submission on a larger pre-assembly message view (`unwindowedMessages`). With `lossless-claw`/LCM as the context engine, this can ... | closed | completed | true | 2 | [] | [] | 2026-04-29T08:49:37Z | 2026-05-04T04:36:09Z | 2026-05-01T14:43:43Z | CONTRIBUTOR | null | 20260504T060031Z | 2026-05-04T06:00:31Z | 100yenadmin | 239,388,517 | U_kgDODkTHZQ | User | false |
openclaw/openclaw | 4,052,146,259 | I_kwDOQb6kR87xhthT | 42,219 | https://github.com/openclaw/openclaw/issues/42219 | https://api.github.com/repos/openclaw/openclaw/issues/42219 | [Bug]: When `fetchUser("@me")` fails during Discord provider startup (e.g. due to transient network issues after a proxy reconnect), the bot continues running with `botUserId = undefined`. This causes multiple security and routing failures. | ### Bug type
Regression (worked before, now fails)
### Summary
### Impact
1. **Mention gating bypassed**: The guard `if (botId && mentionGate.shouldSkip)` in `message-handler.preflight.ts:686` short-circuits when `botId` is undefined, letting **all** guild messages bypass `requireMention: true`.
2. **Self-message f... | closed | completed | true | 2 | [
"bug",
"security",
"regression"
] | [] | 2026-03-10T14:20:42Z | 2026-05-02T04:28:50Z | 2026-04-29T08:56:17Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | huangserva | 201,197,786 | U_kgDOC_4I2g | User | false |
openclaw/openclaw | 4,349,135,570 | I_kwDOQb6kR88AAAABAzqK0g | 74,234 | https://github.com/openclaw/openclaw/issues/74234 | https://api.github.com/repos/openclaw/openclaw/issues/74234 | [Bug]: Possibly incorrect gpg key fingerprint validation in Dockerfile | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
```
expected_fingerprint="$(printf '%s' "$OPENCLAW_DOCKER_GPG_FINGERPRINT" | tr '[:lower:]' '[:upper:]' | tr -d '[:space:]')" && \
actual_fingerprint="$(gpg --batch --show-keys --with-colons /tmp/docker.gpg.asc |... | closed | completed | true | 1 | [
"bug"
] | [
"sallyom"
] | 2026-04-29T08:53:36Z | 2026-05-03T04:34:08Z | 2026-05-01T02:47:45Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | peledins-zimperium | 146,088,545 | U_kgDOCLUiYQ | User | false |
openclaw/openclaw | 4,349,163,733 | I_kwDOQb6kR88AAAABAzr41Q | 74,237 | https://github.com/openclaw/openclaw/issues/74237 | https://api.github.com/repos/openclaw/openclaw/issues/74237 | Feishu channel can hot-loop gateway on stale or corrupt channel session state | ## Summary
`openclaw gateway` can enter a sustained high-CPU state when the Feishu channel is enabled and the local Feishu channel/session state contains stale or corrupt data.
In this case the Feishu app credentials were valid and the bundled Feishu plugin itself was present. The gateway became stable only after reb... | open | null | false | 2 | [] | [] | 2026-04-29T08:58:17Z | 2026-05-03T08:26:10Z | null | CONTRIBUTOR | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Lightningxxl | 221,346,793 | U_kgDODTF76Q | User | false |
openclaw/openclaw | 4,349,038,710 | I_kwDOQb6kR88AAAABAzkQdg | 74,222 | https://github.com/openclaw/openclaw/issues/74222 | https://api.github.com/repos/openclaw/openclaw/issues/74222 | [Bug] transcript repair inserts 'missing tool result' after LCM compaction | ## Bug Report
**Summary**: After LCM (Lossless Context Management) compaction, the session transcript repair logic inserts synthetic error messages: `[openclaw] missing tool result in session history; inserted synthetic error result for transcript repair.`
**Root Cause**:
- File: `session-transcript-repair-D9T_omS-.j... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T08:37:38Z | 2026-05-02T04:28:50Z | 2026-04-29T08:58:56Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | skwalb12 | 144,493,392 | U_kgDOCJzLUA | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.