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 | 3,989,155,874 | I_kwDOQb6kR87txbAi | 26,466 | https://github.com/openclaw/openclaw/issues/26466 | https://api.github.com/repos/openclaw/openclaw/issues/26466 | Thinking content leaks to channel output when Think: off | ## Description
When using Claude Opus 4.5 with `<thinking_mode>interleaved</thinking_mode>` in the system prompt, the thinking/reasoning content is being sent to the channel (Feishu in my case) even when `Think: off` is set in the session.
## Expected Behavior
When `Think: off` is configured, the gateway should filt... | closed | not_planned | true | 7 | [
"stale"
] | [] | 2026-02-25T12:21:39Z | 2026-04-19T04:24:46Z | 2026-04-16T04:23:51Z | NONE | null | 20260419T052035Z | 2026-04-19T05:20:35Z | Vicky-v7 | 189,712,541 | U_kgDOC07InQ | User | false |
openclaw/openclaw | 3,969,268,192 | I_kwDOQb6kR87sljng | 21,985 | https://github.com/openclaw/openclaw/issues/21985 | https://api.github.com/repos/openclaw/openclaw/issues/21985 | [Bug]: Cross-provider failover creates phantom toolResult with empty toolCallId, permanently corrupting session | ### Summary
When a fallback model (Kimi K2.5 via NVIDIA NIM) returns `stopReason: "toolUse"` but the response content contains no actual `toolCall` block (only thinking + text), OpenClaw's transcript repair generates a phantom `toolResult` with empty `toolCallId`: "" and toolName: "". After safeguard compaction, this ... | closed | not_planned | true | 7 | [
"bug",
"stale"
] | [] | 2026-02-20T16:06:31Z | 2026-04-16T04:23:51Z | 2026-04-14T04:20:16Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | ethduke | 2,248,185 | MDQ6VXNlcjIyNDgxODU= | User | false |
openclaw/openclaw | 3,984,511,807 | I_kwDOQb6kR87tftM_ | 25,592 | https://github.com/openclaw/openclaw/issues/25592 | https://api.github.com/repos/openclaw/openclaw/issues/25592 | Text between tool calls leaks to messaging channels | ## Problem
When the agent produces text between tool calls (e.g., error handling, processing acknowledgments, or narration), that text gets routed to the active messaging channel (Slack, iMessage, etc.) as a visible message. This is a significant UX problem — internal processing output, failed exec error traces, and s... | open | null | false | 24 | [] | [] | 2026-02-24T15:59:23Z | 2026-05-05T02:56:00Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | doomclaw | 259,108,889 | U_kgDOD3GwGQ | User | false |
openclaw/openclaw | 3,976,363,893 | I_kwDOQb6kR87tAn91 | 24,173 | https://github.com/openclaw/openclaw/issues/24173 | https://api.github.com/repos/openclaw/openclaw/issues/24173 | [Bug]: compaction memoryFlush not triggered on gateway restart/shutdown — session context lost | ### Summary
When the gateway is restarted (e.g., after an update), active sessions with compaction.memoryFlush.enabled: true do not trigger the configured flush before shutdown. Session history is permanently lost.
### Steps to reproduce
1. Configure compaction.memoryFlush.enabled: true with a systemPrompt instructi... | closed | not_planned | true | 9 | [
"bug",
"stale"
] | [] | 2026-02-23T05:12:01Z | 2026-04-16T04:23:50Z | 2026-04-14T04:20:29Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | chadmbrown | 43,625,896 | MDQ6VXNlcjQzNjI1ODk2 | User | false |
openclaw/openclaw | 3,970,354,972 | I_kwDOQb6kR87sps8c | 22,183 | https://github.com/openclaw/openclaw/issues/22183 | https://api.github.com/repos/openclaw/openclaw/issues/22183 | Email channel MVP: SMTP outbound + channel plugin support | ## Summary
Add an Email channel plugin so OpenClaw can send messages to email recipients via SMTP as a first-class channel.
## Scope (MVP)
- New `email` channel plugin under `extensions/email`
- Outbound send via SMTP (host/port/secure/user/pass/from config)
- Target normalization (`email:user@example.com`, `mailto:us... | closed | not_planned | true | 7 | [
"stale"
] | [] | 2026-02-20T21:00:21Z | 2026-04-16T04:23:50Z | 2026-04-14T04:20:18Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | yhindy | 11,757,328 | MDQ6VXNlcjExNzU3MzI4 | User | false |
openclaw/openclaw | 3,992,378,284 | I_kwDOQb6kR87t9tus | 26,952 | https://github.com/openclaw/openclaw/issues/26952 | https://api.github.com/repos/openclaw/openclaw/issues/26952 | Support provider-specific request parameters (e.g. Inception Mercury `diffusing` flag) | ## Summary
When using custom model providers via `openai-completions` API adapter, there is no way to pass provider-specific parameters in the request body. This affects providers like **Inception (Mercury/Mercury-2)** which require non-standard parameters for optimal operation.
## Problem
Inception's Mercury-2 diff... | closed | not_planned | true | 6 | [
"stale"
] | [] | 2026-02-25T23:30:31Z | 2026-04-16T04:23:49Z | 2026-04-14T04:20:37Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | ericsr | 2,302,911 | MDQ6VXNlcjIzMDI5MTE= | User | false |
openclaw/openclaw | 4,002,662,633 | I_kwDOQb6kR87uk8jp | 29,107 | https://github.com/openclaw/openclaw/issues/29107 | https://api.github.com/repos/openclaw/openclaw/issues/29107 | Agents periodically stop responding — EventQueue listener timeouts during heavy session I/O | ## Summary
Agents occasionally stop responding to messages entirely. During these episodes, the gateway error log shows repeated `EventQueue listener timed out after 30000ms` for both `MESSAGE_CREATE` and `INTERACTION_CREATE` events. The typing indicator also hits its 2-minute TTL and stops. Recovery is eventually aut... | closed | not_planned | true | 5 | [
"stale"
] | [] | 2026-02-27T20:04:55Z | 2026-04-16T04:23:48Z | 2026-04-14T04:20:59Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | petersykim | 35,867,020 | MDQ6VXNlcjM1ODY3MDIw | User | false |
openclaw/openclaw | 3,982,898,829 | I_kwDOQb6kR87tZjaN | 25,303 | https://github.com/openclaw/openclaw/issues/25303 | https://api.github.com/repos/openclaw/openclaw/issues/25303 | [Bug]: Skill-triggered exec calls not shown in verbose output | ### Summary
When a skill is triggered automatically (via `was_mentioned: true` in message metadata), the subsequent `exec` tool calls made by the agent are not logged to the verbose output in the console.
However, when the agent manually reads the skill's SKILL.md and calls `exec` directly (without auto-trigger), the... | closed | not_planned | true | 7 | [
"bug",
"stale"
] | [] | 2026-02-24T10:43:01Z | 2026-04-19T04:24:47Z | 2026-04-16T04:23:48Z | NONE | null | 20260419T052035Z | 2026-04-19T05:20:35Z | notaLonelyDay | 33,029,221 | MDQ6VXNlcjMzMDI5MjIx | User | false |
openclaw/openclaw | 4,009,844,727 | I_kwDOQb6kR87vAV_3 | 31,447 | https://github.com/openclaw/openclaw/issues/31447 | https://api.github.com/repos/openclaw/openclaw/issues/31447 | Discord: only last text block delivered when response contains tool calls | ## Bug Description
When using the Claude Code CLI backend (`claude-cli`) with Discord, only the **last text block** of a multi-block response is delivered to the user. Intermediate text blocks (before/between tool calls) are silently dropped.
For example, if the agent produces:
1. Text: "Let me investigate the logs..... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T07:56:10Z | 2026-04-16T04:23:47Z | 2026-04-14T04:21:49Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Ragingwasabi | 111,104,485 | U_kgDOBp9R5Q | User | false |
openclaw/openclaw | 4,009,839,193 | I_kwDOQb6kR87vAUpZ | 31,442 | https://github.com/openclaw/openclaw/issues/31442 | https://api.github.com/repos/openclaw/openclaw/issues/31442 | [Bug]:The new installation is still OpenClaw 2026.2.26 (bc50708) | ### Bug type
Regression (worked before, now fails)
### Summary
One-liner or npm，The new installation is still OpenClaw 2026.2.26 (bc50708)
### Steps to reproduce
One-liner or npm，The new installation is still OpenClaw 2026.2.26 (bc50708)
### Expected behavior
One-liner or npm，The new installation is still Op... | closed | not_planned | true | 3 | [
"bug",
"stale",
"regression"
] | [] | 2026-03-02T07:54:36Z | 2026-04-16T04:23:47Z | 2026-04-14T04:21:47Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | iwufox | 165,463,159 | U_kgDOCdzEdw | User | false |
openclaw/openclaw | 4,009,903,741 | I_kwDOQb6kR87vAkZ9 | 31,464 | https://github.com/openclaw/openclaw/issues/31464 | https://api.github.com/repos/openclaw/openclaw/issues/31464 | cron list should filter by current group context | ## Problem
When running `openclaw cron list` in a group chat, it shows ALL cron jobs across all groups. This causes:
1. **Confusion** - Users see jobs from other groups they may not know about
2. **Privacy leak** - Job names and schedules from other groups are visible
3. **Noise** - Hard to see relevant jobs for curr... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T08:10:55Z | 2026-04-16T04:23:46Z | 2026-04-14T04:21:50Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Endlessedwork | 166,708,662 | U_kgDOCe_Ftg | User | false |
openclaw/openclaw | 4,009,992,278 | I_kwDOQb6kR87vA6BW | 31,476 | https://github.com/openclaw/openclaw/issues/31476 | https://api.github.com/repos/openclaw/openclaw/issues/31476 | [Bug]: 3.1 regression — message tool "Unknown channel" + health-monitor false-positive stuck detection | ## Environment
- **OpenClaw**: 2026.3.1 (upgraded from 2026.2.26)
- **OS**: Linux x86_64
- **Node**: v22.22.0
- **Channel**: Discord
- **Setup**: Two gateway instances on the same host — main bot + a secondary bot using `--profile`
---
## Bug 1: `message` tool returns "Unknown channel" for valid, resolved channels
... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T08:34:02Z | 2026-04-16T04:23:45Z | 2026-04-14T04:21:52Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Sid-Qin | 201,593,046 | U_kgDODAQQ1g | User | false |
openclaw/openclaw | 4,010,026,999 | I_kwDOQb6kR87vBCf3 | 31,481 | https://github.com/openclaw/openclaw/issues/31481 | https://api.github.com/repos/openclaw/openclaw/issues/31481 | [Bug]: token missing | ### Bug type
Regression (worked before, now fails)
### Summary
Issue 标题
Gateway rejects WebSocket connections with "token missing" despite correct configuration and environment variables
问题描述
OpenClaw 网关启动后，浏览器（Chrome/Safari/QQ 浏览器）尝试连接 WebSocket 时始终返回 unauthorized: gateway token missing，提示需要在 Control UI 设置中粘贴 t... | closed | not_planned | true | 2 | [
"bug",
"stale",
"regression"
] | [] | 2026-03-02T08:42:41Z | 2026-04-16T04:23:44Z | 2026-04-14T04:21:56Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | laixintian | 264,949,997 | U_kgDOD8rQ7Q | User | false |
openclaw/openclaw | 4,010,024,438 | I_kwDOQb6kR87vBB32 | 31,480 | https://github.com/openclaw/openclaw/issues/31480 | https://api.github.com/repos/openclaw/openclaw/issues/31480 | Feature Request: Self-healing gateway with restart countdown, config backup & crash recovery | ## Problem
When the OpenClaw gateway crashes or enters a restart loop due to a bad config change, there is currently no built-in recovery mechanism. The user has to:
- Notice the gateway is down
- Manually diagnose the cause
- Restore a previous config by hand
- Restart manually
This is especially painful on headless... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T08:42:00Z | 2026-04-16T04:23:44Z | 2026-04-14T04:21:54Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Mb29661 | 51,697,539 | MDQ6VXNlcjUxNjk3NTM5 | User | false |
openclaw/openclaw | 4,010,057,754 | I_kwDOQb6kR87vBKAa | 31,487 | https://github.com/openclaw/openclaw/issues/31487 | https://api.github.com/repos/openclaw/openclaw/issues/31487 | Feat: Add timezone configuration for logs | ## Description
Currently, OpenClaw Gateway logs display timestamps in UTC time. For users in other timezones (e.g., China Standard Time, UTC+8), this is confusing as the displayed time doesn't match their local time.
## Expected Behavior
Logs should display timestamps in the local timezone of the system (e.g., China... | closed | not_planned | true | 4 | [
"stale"
] | [] | 2026-03-02T08:50:41Z | 2026-04-16T04:23:43Z | 2026-04-14T04:21:57Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | caesarloo | 11,091,983 | MDQ6VXNlcjExMDkxOTgz | User | false |
openclaw/openclaw | 3,973,999,346 | I_kwDOQb6kR87s3mry | 23,414 | https://github.com/openclaw/openclaw/issues/23414 | https://api.github.com/repos/openclaw/openclaw/issues/23414 | [Bug]: mode="session" requires thread=true — blocks orchestrator pattern on non-Discord channels | ## Summary
`sessions_spawn` with `mode: "session"` is hardcoded to require `thread: true`. Since thread binding currently only works on Discord (via `subagent_spawning` hooks), **persistent sub-agent sessions are completely unavailable on Telegram, Slack, WhatsApp, and all other channels**.
This blocks the orchestrat... | closed | not_planned | true | 9 | [
"stale"
] | [] | 2026-02-22T09:14:31Z | 2026-04-19T04:24:48Z | 2026-04-16T04:23:43Z | NONE | null | 20260419T052035Z | 2026-04-19T05:20:35Z | eminos | 1,682,784 | MDQ6VXNlcjE2ODI3ODQ= | User | false |
openclaw/openclaw | 4,010,079,179 | I_kwDOQb6kR87vBPPL | 31,494 | https://github.com/openclaw/openclaw/issues/31494 | https://api.github.com/repos/openclaw/openclaw/issues/31494 | Telegram topic gets split into two main-thread sessions (thread:<topicId> vs thread:<chatId>:<topicId>) causing context divergence | ### Summary
There is a session routing inconsistency for Telegram topics.
For the same Telegram topic, OpenClaw can create/use **two different session keys** depending on message source:
- `agent:main:main:thread:<topicId>`
- `agent:main:main:thread:<chatId>:<topicId>`
As a result, the same TG topic ends up with two... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T08:56:16Z | 2026-04-16T04:23:42Z | 2026-04-14T04:21:59Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | goxofy | 1,695,416 | MDQ6VXNlcjE2OTU0MTY= | User | false |
openclaw/openclaw | 4,010,102,436 | I_kwDOQb6kR87vBU6k | 31,501 | https://github.com/openclaw/openclaw/issues/31501 | https://api.github.com/repos/openclaw/openclaw/issues/31501 | [Feature Request] Feishu: 支持 useUAT 参数以用户身份创建文档 | ## 描述
当前 feishu_doc 工具创建文档时只支持机器人身份（tenant_access_token），无法以用户身份创建。
## 需求
添加 useUAT 参数支持，允许以用户身份（user_access_token）创建文档。
## 背景
Claude Code 的飞书 MCP 通过 --oauth 参数启动 OAuth 流程获取用户访问令牌，然后使用 useUAT: true 来以用户身份调用 API。
飞书 API 凭证类型：
- tenant_access_token (t-): 机器人身份，无需用户授权
- user_access_token (u-): 用户身份，需要用户授权
## 参考
Cl... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T09:02:14Z | 2026-04-16T04:23:41Z | 2026-04-14T04:22:02Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | louseer | 4,948,307 | MDQ6VXNlcjQ5NDgzMDc= | User | false |
openclaw/openclaw | 4,010,094,616 | I_kwDOQb6kR87vBTAY | 31,499 | https://github.com/openclaw/openclaw/issues/31499 | https://api.github.com/repos/openclaw/openclaw/issues/31499 | Feature request: sessions_await — block until multiple sub-agents complete | ## Problem
When an orchestrator sub-agent (depth-1) spawns multiple worker sub-agents (depth-2), there is no reliable way to wait for all of them to complete before proceeding.
The current approach requires the orchestrator LLM to poll `sessions_list` in a loop, checking the status of each child session. This is unre... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T09:00:14Z | 2026-04-16T04:23:41Z | 2026-04-14T04:22:01Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Bwarhness | 3,010,513 | MDQ6VXNlcjMwMTA1MTM= | User | false |
openclaw/openclaw | 4,010,155,225 | I_kwDOQb6kR87vBhzZ | 31,518 | https://github.com/openclaw/openclaw/issues/31518 | https://api.github.com/repos/openclaw/openclaw/issues/31518 | feat: Slack adapter support for ACP thread-bound sessions (spawnAcpSessions) | ## Feature Request
**Is your feature request related to a problem?**
Thread-bound ACP sessions currently only support Discord via `channels.discord.threadBindings.spawnAcpSessions`. Slack has native thread support but the Slack adapter doesn't expose `spawnAcpSessions` or thread binding capabilities for ACP sessions.
... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T09:15:29Z | 2026-04-16T04:23:40Z | 2026-04-14T04:22:05Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Fato07 | 40,800,448 | MDQ6VXNlcjQwODAwNDQ4 | User | false |
openclaw/openclaw | 4,010,189,695 | I_kwDOQb6kR87vBqN_ | 31,533 | https://github.com/openclaw/openclaw/issues/31533 | https://api.github.com/repos/openclaw/openclaw/issues/31533 | Bot-to-bot streaming messages arrive truncated in channels | ## Problem
When two OpenClaw bots communicate via a Telegram channel with `streaming: "partial"` enabled, the receiving bot processes the initial `channel_post` immediately — before the sending bot finishes streaming edits. This means the receiving bot sees truncated/partial text.
The workaround is setting `streaming... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T09:23:35Z | 2026-04-16T04:23:39Z | 2026-04-14T04:22:09Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | medvedev1088 | 930,603 | MDQ6VXNlcjkzMDYwMw== | User | false |
openclaw/openclaw | 4,010,155,836 | I_kwDOQb6kR87vBh88 | 31,519 | https://github.com/openclaw/openclaw/issues/31519 | https://api.github.com/repos/openclaw/openclaw/issues/31519 | gpt-5.3-codex-spark not routable: inherits chatgpt.com/backend-api endpoint instead of Cerebras endpoint | ## Summary
`gpt-5.3-codex-spark` is configured in `agents.defaults.models` and appears correctly in `/models`, but attempting to use it fails with:
```
run error: {"detail":"The 'gpt-5.3-codex-spark' model is not supported when using Codex with a ChatGPT account."}
```
## Root Cause
In `model-catalog-BiejcmsY.js`, ... | closed | not_planned | true | 4 | [
"stale"
] | [] | 2026-03-02T09:15:38Z | 2026-04-16T04:23:39Z | 2026-04-14T04:22:07Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | data-t3labs | 264,575,540 | U_kgDOD8UaNA | User | false |
openclaw/openclaw | 4,010,256,820 | I_kwDOQb6kR87vB6m0 | 31,542 | https://github.com/openclaw/openclaw/issues/31542 | https://api.github.com/repos/openclaw/openclaw/issues/31542 | macOS CI job runs full TS test suite for iOS-only changes, causing timeout | ## Problem
PR #30885 only touches a single iOS file (`apps/ios/Sources/Voice/TalkModeManager.swift`), but the `macOS` CI job fails after 3h15m (timeout).
The `changed-scope` job in CI groups `apps/ios/*` under `run_macos=true`, which is reasonable since Xcode builds require macOS. However, the `macos` job runs `pnpm ... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T09:38:27Z | 2026-04-16T04:23:38Z | 2026-04-14T04:22:10Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | leepokai | 109,857,817 | U_kgDOBoxMGQ | User | false |
openclaw/openclaw | 3,969,709,247 | I_kwDOQb6kR87snPS_ | 22,085 | https://github.com/openclaw/openclaw/issues/22085 | https://api.github.com/repos/openclaw/openclaw/issues/22085 | openclaw agent: --session-id doesn't create isolated sessions | ## Problem
`openclaw agent --agent <id> --session-id <uuid>` sets the session UUID but does NOT change the session key. The session key is always `agent:<id>:main` regardless of `--session-id` or `--to` flags.
This means all `openclaw agent` calls for the same agent share the same session, even when different `--sess... | closed | not_planned | true | 8 | [
"stale"
] | [] | 2026-02-20T17:59:24Z | 2026-04-19T04:24:50Z | 2026-04-16T04:23:38Z | NONE | null | 20260419T052035Z | 2026-04-19T05:20:35Z | compass-soul | 260,311,058 | U_kgDOD4QIEg | User | false |
openclaw/openclaw | 4,010,273,678 | I_kwDOQb6kR87vB-uO | 31,544 | https://github.com/openclaw/openclaw/issues/31544 | https://api.github.com/repos/openclaw/openclaw/issues/31544 | [Bug]: Warning Message When Selecting a Model with API KEY Not Configured is Not Helpful | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Warning message when selecting a new model but API key not configured is not helpful.
### Steps to reproduce
1. In console, run `openclaw tui`.
2. Run `/model`.
3. Choose `google/gemini-3-flash-preview`.
4. Run any query.
### Expected be... | closed | not_planned | true | 3 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-02T09:41:58Z | 2026-04-16T04:23:37Z | 2026-04-14T04:22:12Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | arjunaskykok | 32,124,593 | MDQ6VXNlcjMyMTI0NTkz | User | false |
openclaw/openclaw | 4,010,300,874 | I_kwDOQb6kR87vCFXK | 31,552 | https://github.com/openclaw/openclaw/issues/31552 | https://api.github.com/repos/openclaw/openclaw/issues/31552 | Webchat message dropped and replaced by heartbeat execution on LLM Provider Error | # Title: Webchat message dropped and replaced by heartbeat execution on LLM Provider Error
## Bug Summary
When sending a message to the agent via the Webchat UI, if the underlying LLM provider (Anthropic, OpenAI, DeepSeek, etc.) returns a fatal error (like an `HTTP 401 authentication_error`), the user's message com... | closed | not_planned | true | 4 | [
"stale"
] | [] | 2026-03-02T09:48:06Z | 2026-04-16T04:23:36Z | 2026-04-14T04:22:14Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | ravpathak1 | 4,957,525 | MDQ6VXNlcjQ5NTc1MjU= | User | false |
openclaw/openclaw | 3,964,039,188 | I_kwDOQb6kR87sRnAU | 21,097 | https://github.com/openclaw/openclaw/issues/21097 | https://api.github.com/repos/openclaw/openclaw/issues/21097 | Support thinkingDefault in per-agent config (agents.list entries) | ## Feature Request
### Problem
`thinkingDefault` is supported in `agents.defaults` but not in individual agent entries under `agents.list`. This means all agents inherit the same default thinking level, which is wasteful when you have agents with different complexity needs (e.g., Opus for interactive use vs Sonnet for... | closed | not_planned | true | 10 | [
"stale",
"dedupe:parent"
] | [] | 2026-02-19T17:08:04Z | 2026-04-19T04:24:51Z | 2026-04-16T04:23:36Z | NONE | null | 20260419T052035Z | 2026-04-19T05:20:35Z | tdavis009 | 223,409,012 | U_kgDODVDzdA | User | false |
openclaw/openclaw | 4,010,434,345 | I_kwDOQb6kR87vCl8p | 31,575 | https://github.com/openclaw/openclaw/issues/31575 | https://api.github.com/repos/openclaw/openclaw/issues/31575 | fix(tui): activity status stuck on 'streaming' after cross-channel run completes | ## Bug
When chatting with an agent via Telegram (or any other channel) while the TUI is open on the same session, the TUI status bar gets stuck showing "streaming" indefinitely after the conversation ends.
## Steps to Reproduce
1. Open TUI, viewing an agent session (e.g. `agent:agent_name:main`)
2. Chat with the sam... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T10:17:29Z | 2026-04-16T04:23:35Z | 2026-04-14T04:22:18Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | cheyuni | 1,212,892 | MDQ6VXNlcjEyMTI4OTI= | User | false |
openclaw/openclaw | 4,010,403,405 | I_kwDOQb6kR87vCeZN | 31,570 | https://github.com/openclaw/openclaw/issues/31570 | https://api.github.com/repos/openclaw/openclaw/issues/31570 | [Bug]: No longer able to run commands on his own | ### Bug type
Regression (worked before, now fails)
### Summary
After upgrading to 2026.3.1 - it's pretty much useless:
> I’m sorry, but I don’t have the ability to run commands on your machine.
### Steps to reproduce
1. Configure Telegram
2. Send message
3. Get back response - it can't run anything itself.
### ... | closed | not_planned | true | 4 | [
"bug",
"stale",
"regression"
] | [] | 2026-03-02T10:10:23Z | 2026-04-16T04:23:35Z | 2026-04-14T04:22:16Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | slyapustin | 370,774 | MDQ6VXNlcjM3MDc3NA== | User | false |
openclaw/openclaw | 4,010,446,436 | I_kwDOQb6kR87vCo5k | 31,578 | https://github.com/openclaw/openclaw/issues/31578 | https://api.github.com/repos/openclaw/openclaw/issues/31578 | Feature: post-restart notification to channels | ## Problem
When the gateway restarts (e.g. after `openclaw update` or `openclaw gateway restart`), all active agent sessions are terminated. The agent cannot proactively notify the user that it's back online — it only wakes up when the user sends a new message.
This creates a frustrating UX where the user has to keep... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T10:20:03Z | 2026-04-16T04:23:34Z | 2026-04-14T04:22:20Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | nickjlamb | 129,560,626 | U_kgDOB7jwMg | User | false |
openclaw/openclaw | 3,960,769,177 | I_kwDOQb6kR87sFIqZ | 20,568 | https://github.com/openclaw/openclaw/issues/20568 | https://api.github.com/repos/openclaw/openclaw/issues/20568 | Telegram streamMode: "partial" broken in 2026.2.17 (regression from 2026.2.14) | ## Description
`streamMode: "partial"` for Telegram DM stops working after upgrading from 2026.2.14 to 2026.2.17. Instead of smooth character-by-character streaming (with trailing ellipsis indicator), messages are delivered in large chunks — effectively falling back to block-style delivery.
## Steps to Reproduce
1. ... | closed | not_planned | true | 9 | [
"stale"
] | [] | 2026-02-19T03:11:47Z | 2026-04-19T04:24:52Z | 2026-04-16T04:23:34Z | NONE | null | 20260419T052035Z | 2026-04-19T05:20:35Z | tacnaci | 40,619,572 | MDQ6VXNlcjQwNjE5NTcy | User | false |
openclaw/openclaw | 4,010,612,290 | I_kwDOQb6kR87vDRZC | 31,598 | https://github.com/openclaw/openclaw/issues/31598 | https://api.github.com/repos/openclaw/openclaw/issues/31598 | [Bug]: local VLLM model returns empty response (no generated text) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When I use OpenClaw to call the large model deployed by vLLM, it returns an empty response. However, other third-party applications work fine.
vllm'version is 0.11.0
the model is DeepSeek-V3.2-Exp-w8a8
Here are the OpenClaw configuration... | closed | not_planned | true | 2 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-02T10:55:57Z | 2026-04-16T04:23:33Z | 2026-04-14T04:22:23Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | PerroDoge | 44,633,568 | MDQ6VXNlcjQ0NjMzNTY4 | User | false |
openclaw/openclaw | 4,010,593,519 | I_kwDOQb6kR87vDMzv | 31,595 | https://github.com/openclaw/openclaw/issues/31595 | https://api.github.com/repos/openclaw/openclaw/issues/31595 | Telegram supergroup config (groups, requireMention) doesn't persist after restart | ## Summary
Telegram supergroup `groups` configuration keys (`groups`, `allowFrom`, `groupAllowFrom`, `requireMention`) do not persist after gateway restart. Config values get rewritten to `null`, making it impossible to configure bots in supergroups without @mention requirement.
## Environment
- **OpenClaw version:*... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T10:51:50Z | 2026-04-16T04:23:33Z | 2026-04-14T04:22:21Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | aramshakh | 140,006,457 | U_kgDOCFhUOQ | User | false |
openclaw/openclaw | 4,010,686,797 | I_kwDOQb6kR87vDjlN | 31,614 | https://github.com/openclaw/openclaw/issues/31614 | https://api.github.com/repos/openclaw/openclaw/issues/31614 | [Android] Canvas WebView fails to load due to missing authentication headers | # [Android] Canvas WebView fails to load due to missing authentication headers
## Description
The Android app's Canvas feature shows "Canvas unavailable" because the WebView cannot access the Gateway's canvas host endpoint which requires authentication.
## Root Cause
In `CanvasController.kt`, the `reload()` method ... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T11:13:19Z | 2026-04-16T04:23:31Z | 2026-04-14T04:22:27Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Louisone1 | 166,708,625 | U_kgDOCe_FkQ | User | false |
openclaw/openclaw | 4,010,637,826 | I_kwDOQb6kR87vDXoC | 31,603 | https://github.com/openclaw/openclaw/issues/31603 | https://api.github.com/repos/openclaw/openclaw/issues/31603 | [Feature]: Provider-agnostic batch tool to reduce tool call context overhead | ## Summary
A plugin tool (`batch`) that lets the LLM specify multiple read/fetch operations in a single tool call, executes them in parallel, and returns one combined structured result — reducing the number of context-consuming tool result messages.
## Problem to solve
When the model needs to perform N similar opera... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T11:01:38Z | 2026-04-16T04:23:31Z | 2026-04-14T04:22:25Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | LikunYDev | 53,739,955 | MDQ6VXNlcjUzNzM5OTU1 | User | false |
openclaw/openclaw | 4,014,718,160 | I_kwDOQb6kR87vS7zQ | 32,579 | https://github.com/openclaw/openclaw/issues/32579 | https://api.github.com/repos/openclaw/openclaw/issues/32579 | [Bug]: Gateway broadcasts chat events to all WebSocket clients regardless of session | ## Bug Description
The gateway broadcasts all `chat` events (deltas, finals, errors) to **every** connected WebSocket client via `broadcast("chat", payload)` without any session-scoped filtering. While the payload includes a `sessionKey` field, the server makes no attempt to route events only to clients associated wit... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-03T05:00:50Z | 2026-04-16T04:23:30Z | 2026-04-14T04:22:58Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | YUJIE2002 | 123,847,463 | U_kgDOB2HDJw | User | false |
openclaw/openclaw | 4,014,768,862 | I_kwDOQb6kR87vTILe | 32,598 | https://github.com/openclaw/openclaw/issues/32598 | https://api.github.com/repos/openclaw/openclaw/issues/32598 | Feature: Gateway-level liveness check — instant agent status without waiting for LLM turn | ## Summary
When an agent is busy (long tool calls, subagent orchestration, streaming a response), there's no way to get a quick "are you alive?" answer. Status queries enter the same message queue and wait behind the current turn, sometimes for minutes.
Users need an instant liveness signal — "yes, I'm here, currentl... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-03T05:22:31Z | 2026-04-16T04:23:28Z | 2026-04-14T04:23:04Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | AytuncYildizli | 47,717,026 | MDQ6VXNlcjQ3NzE3MDI2 | User | false |
openclaw/openclaw | 4,014,776,593 | I_kwDOQb6kR87vTKER | 32,601 | https://github.com/openclaw/openclaw/issues/32601 | https://api.github.com/repos/openclaw/openclaw/issues/32601 | Feature Request: Per-channel sessions for Discord (and other multi-channel platforms) | ## Feature Request
**Is your feature request related to a problem?**
Currently, all Discord channels within a guild route to a single agent session. When a user has topic-specific channels (e.g. #finance, #research, #ops), the conversations share the same context window. This causes:
- Context pollution between unrela... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-03T05:25:36Z | 2026-04-16T04:23:27Z | 2026-04-14T04:23:08Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | otcombo | 8,352,042 | MDQ6VXNlcjgzNTIwNDI= | User | false |
openclaw/openclaw | 4,014,775,476 | I_kwDOQb6kR87vTJy0 | 32,599 | https://github.com/openclaw/openclaw/issues/32599 | https://api.github.com/repos/openclaw/openclaw/issues/32599 | Slack channel plugin: add signing secret verification (HMAC-SHA256) | ## Problem
The Slack channel plugin currently accepts all inbound webhook payloads without verifying the `x-slack-signature` header. This means any HTTP request to the gateway's Slack endpoint is accepted as legitimate, regardless of source.
Slack provides a signing secret mechanism specifically to prevent this — eve... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-03T05:25:12Z | 2026-04-16T04:23:27Z | 2026-04-14T04:23:06Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | ben1440 | 258,074,641 | U_kgDOD2HoEQ | User | false |
openclaw/openclaw | 3,948,637,951 | I_kwDOQb6kR87rW27_ | 18,237 | https://github.com/openclaw/openclaw/issues/18237 | https://api.github.com/repos/openclaw/openclaw/issues/18237 | Feature Request: Async exec callback — inject result back to session when process exits | ## Problem
When an agent launches a long-running process (e.g., Claude Code via `exec`), there is no way to receive the result without either:
1. **Blocking the agent turn** — The agent sits idle waiting for the process (wastes resources, and sub-agent sessions get killed by system lifetime limits before the process ... | closed | not_planned | true | 11 | [
"stale"
] | [] | 2026-02-16T16:48:18Z | 2026-04-19T04:24:53Z | 2026-04-16T04:23:27Z | NONE | null | 20260419T052035Z | 2026-04-19T05:20:35Z | albert-mochi | 258,370,801 | U_kgDOD2Zs8Q | User | false |
openclaw/openclaw | 3,923,670,870 | I_kwDOQb6kR87p3ndW | 13,688 | https://github.com/openclaw/openclaw/issues/13688 | https://api.github.com/repos/openclaw/openclaw/issues/13688 | Discord: WebSocket 1005/1006 disconnects with failing resume logic and unbounded backoff | ## Summary
Discord WebSocket connections repeatedly drop (codes 1005/1006), resume attempts fail, and exponential backoff grows without bound, leaving the bot offline for 30+ minutes. DMs are silently lost during downtime.
## Environment
- **OpenClaw version:** 2026.1.24-3
- **OS:** macOS (Darwin 25.3.0, arm64)
- **Ch... | closed | not_planned | true | 22 | [
"bug",
"stale"
] | [] | 2026-02-10T21:34:49Z | 2026-04-19T04:24:53Z | 2026-04-16T04:23:23Z | NONE | null | 20260419T052035Z | 2026-04-19T05:20:35Z | carmandale | 2,348,819 | MDQ6VXNlcjIzNDg4MTk= | User | false |
openclaw/openclaw | 4,011,023,064 | I_kwDOQb6kR87vE1rY | 31,679 | https://github.com/openclaw/openclaw/issues/31679 | https://api.github.com/repos/openclaw/openclaw/issues/31679 | Discord: multi-block model responses sent as separate messages when streaming=off | ## Summary
When a model returns a single response split across multiple content blocks (e.g., `choices[0].message.content` arriving in fragments), each block is dispatched as a **separate Discord message** even when `channels.discord.streaming` is set to `"off"`. This results in garbled mid-word splits visible to end ... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T12:30:14Z | 2026-04-16T04:23:20Z | 2026-04-13T04:29:45Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | rafafortesopenclaw | 259,107,268 | U_kgDOD3GpxA | User | false |
openclaw/openclaw | 3,888,790,324 | I_kwDOQb6kR87nyjs0 | 7,663 | https://github.com/openclaw/openclaw/issues/7663 | https://api.github.com/repos/openclaw/openclaw/issues/7663 | Slack DM replies from embedded/main agent don’t get delivered back to Slack | **Environment**
- OpenClaw version: `2026.1.30 (76b5208)`
- OS: Windows 10
- Gateway: Scheduled Task, loopback only
- Slack config (redacted):
```jsonc
"channels": {
"slack": {
"mode": "socket",
"webhookPath": "/slack/events",
"enabled": true,
"botToken": "xoxb-…",
"appToken": "xapp-1-…",
"u... | closed | not_planned | true | 16 | [
"bug",
"stale"
] | [] | 2026-02-03T03:09:03Z | 2026-04-19T04:24:54Z | 2026-04-16T04:23:12Z | NONE | null | 20260419T052035Z | 2026-04-19T05:20:35Z | glovesave66 | 127,753,282 | U_kgDOB51cQg | User | false |
openclaw/openclaw | 3,960,739,919 | I_kwDOQb6kR87sFBhP | 20,562 | https://github.com/openclaw/openclaw/issues/20562 | https://api.github.com/repos/openclaw/openclaw/issues/20562 | Feature Request: Add Serper.dev as a web_search provider | ## Summary
Add [Serper.dev](https://serper.dev) as a native `web_search` provider alongside Brave and Perplexity.
## Motivation
- **Brave Search killed its free tier** (Feb 2026) — now $5/mo for ~1K queries, which many users will hit quickly
- **Serper returns real-time Google results** — critical for time-sensitive... | closed | not_planned | true | 10 | [
"stale"
] | [] | 2026-02-19T03:01:03Z | 2026-04-16T04:23:10Z | 2026-04-13T04:26:32Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | travislius | 64,033,454 | MDQ6VXNlcjY0MDMzNDU0 | User | false |
openclaw/openclaw | 4,092,387,980 | I_kwDOQb6kR87z7OKM | 49,395 | https://github.com/openclaw/openclaw/issues/49395 | https://api.github.com/repos/openclaw/openclaw/issues/49395 | BlueBubbles webhook POST returns 404 despite plugin reporting 'webhook listening' | ## BlueBubbles webhook POST returns 404 despite plugin reporting "webhook listening"
**Version:** OpenClaw 2026.3.13 (61d171a)
**OS:** macOS Tahoe 26.4.0 (arm64)
**BlueBubbles Server:** 1.9.9
### Problem
POST requests to the BlueBubbles webhook path return 404, while GET returns 200 (SPA catch-all HTML).
### Config
... | closed | completed | true | 7 | [] | [] | 2026-03-18T02:59:26Z | 2026-04-18T04:13:57Z | 2026-04-16T04:05:22Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | imaznation | 1,430,148 | MDQ6VXNlcjE0MzAxNDg= | User | false |
openclaw/openclaw | 4,086,155,223 | I_kwDOQb6kR87zjcfX | 48,734 | https://github.com/openclaw/openclaw/issues/48734 | https://api.github.com/repos/openclaw/openclaw/issues/48734 | Plugin HTTP route dispatch uses stale registry - webhook POST returns 404 | ## Bug Description
BlueBubbles (and likely other channel plugins) webhook HTTP routes return 404 despite being correctly registered during gateway startup. The route IS registered and appears in logs (`BlueBubbles webhook listening on /bluebubbles-webhook`), but POST requests fall through to the 404 fallback.
## Root... | closed | completed | true | 5 | [] | [] | 2026-03-17T04:36:28Z | 2026-04-18T04:13:58Z | 2026-04-16T04:05:20Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | reidmcgill | 207,939,519 | U_kgDODGTnvw | User | false |
openclaw/openclaw | 4,079,283,337 | I_kwDOQb6kR87zJOyJ | 47,650 | https://github.com/openclaw/openclaw/issues/47650 | https://api.github.com/repos/openclaw/openclaw/issues/47650 | [2026.3.13 Regression] Multiple critical bugs: CLI scope validation broken, BlueBubbles webhook route not registered | ## Summary
OpenClaw 2026.3.13 has multiple critical regressions:
1. CLI authentication completely broken
2. BlueBubbles webhook route fails to register silently
3. Memory leak in doctor command
## Affected Version
2026.3.13
## Issue 1: CLI Connection Failure
CLI commands fail to connect to the gateway with error:
``... | closed | completed | true | 3 | [] | [] | 2026-03-15T21:58:58Z | 2026-04-18T04:13:59Z | 2026-04-16T04:05:18Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | davekurt | 7,126,839 | MDQ6VXNlcjcxMjY4Mzk= | User | false |
openclaw/openclaw | 4,117,940,040 | I_kwDOQb6kR871csdI | 52,605 | https://github.com/openclaw/openclaw/issues/52605 | https://api.github.com/repos/openclaw/openclaw/issues/52605 | [Bug]: iMessage inbound messages not received on v2026.3.x — BlueBubbles webhook route returns 404 after upgrade | ### Bug type
Regression (worked before, now fails)
### Summary
After upgrading from v2026.2.23 to any v2026.3.x release (tested v2026.3.13), iMessage messages are no longer received by the agent. The BlueBubbles webhook endpoint (`/bluebubbles-webhook`) returns `404 Not Found` for POST requests, silently dropping al... | closed | completed | true | 2 | [] | [] | 2026-03-23T03:26:32Z | 2026-04-18T04:14:00Z | 2026-04-16T04:05:16Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | chataclaw | 259,815,951 | U_kgDOD3x6Dw | User | false |
openclaw/openclaw | 4,076,247,117 | I_kwDOQb6kR87y9phN | 46,404 | https://github.com/openclaw/openclaw/issues/46404 | https://api.github.com/repos/openclaw/openclaw/issues/46404 | BlueBubbles webhook 404, plugin registry drift, and session metadata parity | ## Summary
Three related bugs introduced in or around `2026.3.12` that break BlueBubbles webhook routing, cause plugin registry state loss, and prevent TUI session tracking for BlueBubbles conversations.
## Issues
### 1. BlueBubbles webhook returns 404
**POST** to `/bluebubbles-webhook` returns "Not Found" even thou... | closed | completed | true | 3 | [] | [] | 2026-03-14T17:00:26Z | 2026-04-18T04:14:00Z | 2026-04-16T04:05:14Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | joshskyhi | 245,895,373 | U_kgDODqgQzQ | User | false |
openclaw/openclaw | 4,085,701,349 | I_kwDOQb6kR87zhtrl | 48,624 | https://github.com/openclaw/openclaw/issues/48624 | https://api.github.com/repos/openclaw/openclaw/issues/48624 | BlueBubbles webhook route returns 404 on gateway port (v2026.3.13) | ## Bug Report
**Version:** OpenClaw 2026.3.13 (61d171a)
**OS:** macOS Tahoe 26.2 (Mac Mini M4)
**Channel:** BlueBubbles
**Node:** v25.5.0
### Description
The BlueBubbles webhook handler logs `BlueBubbles webhook listening on /bluebubbles-webhook` during gateway startup, but HTTP POST requests to `http://127.0.... | closed | completed | true | 4 | [] | [] | 2026-03-17T01:37:21Z | 2026-04-18T04:14:01Z | 2026-04-16T04:05:13Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | KingIVthe1st | 212,242,084 | U_kgDODKaOpA | User | false |
openclaw/openclaw | 4,105,858,169 | I_kwDOQb6kR870umx5 | 50,856 | https://github.com/openclaw/openclaw/issues/50856 | https://api.github.com/repos/openclaw/openclaw/issues/50856 | session-memory and boot-md hooks break BlueBubbles webhook HTTP route (404) | ## Environment
- OpenClaw: 2026.3.13
- macOS: 15.6 (Sequoia) / Apple M4
- Node: 25.7.0
- Channel: BlueBubbles (bundled plugin)
## Description
When `hooks.internal.enabled: true` is set, the BlueBubbles webhook HTTP endpoint (`/bluebubbles-webhook`) returns **404 Not Found** instead of accepting POST requests. The gat... | closed | completed | true | 2 | [] | [] | 2026-03-20T05:24:18Z | 2026-04-18T04:14:02Z | 2026-04-16T04:05:09Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | mids-neo | 117,532,760 | U_kgDOBwFoWA | User | false |
openclaw/openclaw | 4,096,550,120 | I_kwDOQb6kR870LGTo | 49,932 | https://github.com/openclaw/openclaw/issues/49932 | https://api.github.com/repos/openclaw/openclaw/issues/49932 | BlueBubbles webhook route can disappear at runtime and start returning 404 | ## Summary
BlueBubbles webhook delivery can start returning `404 Not Found` after the gateway has been running for a while, even though OpenClaw still reports that the webhook is listening.
## Version / Environment
- **OpenClaw version:** `2026.3.13`
- **Channel:** `stable`
- **OS:** `macOS 15.7.3 (arm64)`
- **Node:... | closed | completed | true | 4 | [] | [] | 2026-03-18T17:09:28Z | 2026-04-18T04:14:03Z | 2026-04-16T04:05:07Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | yiqingsir | 111,381,629 | U_kgDOBqOMfQ | User | false |
openclaw/openclaw | 4,082,342,218 | I_kwDOQb6kR87zU5lK | 48,212 | https://github.com/openclaw/openclaw/issues/48212 | https://api.github.com/repos/openclaw/openclaw/issues/48212 | bug(bluebubbles): HTTP POST webhook routes return 404 in 2026.3.13 | ## Summary
All HTTP POST routes to the OpenClaw gateway return **404** in v2026.3.13. Only GET and WebSocket connections work. This breaks BlueBubbles webhook delivery entirely — inbound messages from BlueBubbles never reach the agent.
## Environment
- **OpenClaw version:** 2026.3.13 (61d171a)
- **macOS:** Darwin 25... | closed | completed | true | 4 | [] | [] | 2026-03-16T12:58:53Z | 2026-04-18T04:14:04Z | 2026-04-16T04:05:04Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | calvintandjung | 48,805,358 | MDQ6VXNlcjQ4ODA1MzU4 | User | false |
openclaw/openclaw | 4,084,618,459 | I_kwDOQb6kR87zdlTb | 48,473 | https://github.com/openclaw/openclaw/issues/48473 | https://api.github.com/repos/openclaw/openclaw/issues/48473 | [Bug]: Plugin HTTP routes return 404 — registry mismatch in setActivePluginRegistry across multiple bundle chunks | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
setActivePluginRegistry() is duplicated across 6 bundle chunks; when loadOpenClawPlugins() is re-invoked after startup (e.g. session bootstrap), it replaces the live plugin registry
with a fresh empty one, silently dropping all plugin-reg... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-03-16T20:25:04Z | 2026-04-18T04:14:04Z | 2026-04-16T04:05:02Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | Nurozen | 36,425,183 | MDQ6VXNlcjM2NDI1MTgz | User | false |
openclaw/openclaw | 4,115,868,996 | I_kwDOQb6kR871Uy1E | 52,235 | https://github.com/openclaw/openclaw/issues/52235 | https://api.github.com/repos/openclaw/openclaw/issues/52235 | Context counter shows 0/200k with LiteLLM proxy — compaction never triggers | ## Environment
- OpenClaw: 2026.3.13 (61d171a)
- Model: `litellm/claude-opus-4-6` via LiteLLM proxy (localhost:4000)
- OS: Debian 13 (arm64, Raspberry Pi 5)
## Problem
`/status` always shows `Context: 0/200k (0%)` and `Compactions: 0`, even after long conversations with extensive tool use.
## Expected behavior
Cont... | closed | completed | true | 5 | [] | [] | 2026-03-22T12:22:13Z | 2026-04-28T04:36:15Z | 2026-04-26T03:57:12Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | shadow-enthusiast | 252,925,178 | U_kgDODxNU-g | User | false |
openclaw/openclaw | 4,017,003,603 | I_kwDOQb6kR87vbpxT | 33,106 | https://github.com/openclaw/openclaw/issues/33106 | https://api.github.com/repos/openclaw/openclaw/issues/33106 | [Feature]: Runtime trust verification for agent-to-agent interactions via TrustChain | ## Problem
When two AI agents talk, neither can verify who the other is. No ID. No audit trail. No way to verify.
VirusTotal scanning (announced Feb 7) catches malicious code in skills — static analysis. But it can't catch:
- A legitimate skill that starts behaving maliciously after gaining trust
- An agent impersona... | closed | not_planned | true | 6 | [
"stale"
] | [] | 2026-03-03T14:31:13Z | 2026-04-27T05:03:26Z | 2026-04-24T19:11:21Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | viftode4 | 37,593,074 | MDQ6VXNlcjM3NTkzMDc0 | User | false |
openclaw/openclaw | 4,057,998,212 | I_kwDOQb6kR87x4COE | 43,177 | https://github.com/openclaw/openclaw/issues/43177 | https://api.github.com/repos/openclaw/openclaw/issues/43177 | Cron announce may report delivered=true without actual Telegram delivery | ## Summary
Isolated cron jobs configured with explicit Telegram announce delivery can be recorded as `delivered: true` / `deliveryStatus: delivered` in cron run history even when no Telegram message is actually sent to the user.
This makes scheduled follow-ups look successful internally while the user receives nothin... | closed | completed | true | 6 | [] | [] | 2026-03-11T12:39:31Z | 2026-04-27T04:51:22Z | 2026-04-24T22:18:22Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | darklion84 | 95,440,896 | U_kgDOBbBQAA | User | false |
openclaw/openclaw | 4,109,013,072 | I_kwDOQb6kR8706pBQ | 51,154 | https://github.com/openclaw/openclaw/issues/51154 | https://api.github.com/repos/openclaw/openclaw/issues/51154 | Session hangs when context reaches ~198k tokens — no user reply, infinite tool loop; file-based reset doesn't clear in-memory state | ## Bug Report
### Summary
When a session approaches the context limit (~198k/200k tokens), the agent enters an infinite tool-call loop and never sends a final reply to the user. The chat appears dead — the user sees no response. Additionally, resetting the session by deleting/renaming the `.jsonl` file does not clear ... | closed | completed | true | 3 | [] | [] | 2026-03-20T17:05:02Z | 2026-04-30T04:52:40Z | 2026-04-25T05:05:42Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | IIIyban | 137,302,231 | U_kgDOCC8Q1w | User | false |
openclaw/openclaw | 4,050,307,597 | I_kwDOQb6kR87xasoN | 41,972 | https://github.com/openclaw/openclaw/issues/41972 | https://api.github.com/repos/openclaw/openclaw/issues/41972 | Per-agent skills filter not applied in live gateway sessions | ## Summary
The `agents.list[].skills` filter configured in `openclaw.json` is not applied consistently across all code paths, and cached skill snapshots are never refreshed when the filter changes.
## Steps to Reproduce
1. Configure a per-agent skills filter:
```json
{
"agents": {
"list": [{
... | closed | completed | true | 3 | [] | [] | 2026-03-10T08:45:03Z | 2026-04-27T04:53:50Z | 2026-04-24T22:09:44Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | syedair | 20,495,779 | MDQ6VXNlcjIwNDk1Nzc5 | User | false |
openclaw/openclaw | 4,156,451,562 | I_kwDOQb6kR873vmrq | 56,075 | https://github.com/openclaw/openclaw/issues/56075 | https://api.github.com/repos/openclaw/openclaw/issues/56075 | Shared skills should propagate reliably to existing chats, not just new sessions | ## Summary
OpenClaw should support a first-class workflow where **shared/global skills added after chats already exist** become available to those chats **reliably and predictably**, without requiring manual prompting, per-workspace duplication, or recreating sessions.
Today, shared skill loading appears to exist at ... | closed | completed | true | 2 | [] | [] | 2026-03-27T23:26:36Z | 2026-04-28T05:22:30Z | 2026-04-25T11:24:55Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | juhaclaw | 269,972,226 | U_kgDOEBdzAg | User | false |
openclaw/openclaw | 4,088,899,215 | I_kwDOQb6kR87zt6aP | 49,059 | https://github.com/openclaw/openclaw/issues/49059 | https://api.github.com/repos/openclaw/openclaw/issues/49059 | Bug: Skills snapshot not refreshed for existing sessions after gateway restart | ## Bug: Skills snapshot not refreshed for existing sessions after gateway restart
### Summary
When new skills are installed while the gateway is running, existing sessions continue using a stale skill snapshot even after gateway restart. Only new sessions (created after skill installation) see the updated skill list.... | closed | completed | true | 3 | [] | [] | 2026-03-17T14:21:53Z | 2026-04-27T04:34:22Z | 2026-04-25T04:03:21Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | rivercrab26 | 259,764,886 | U_kgDOD3uylg | User | false |
openclaw/openclaw | 4,121,592,439 | I_kwDOQb6kR871qoJ3 | 52,952 | https://github.com/openclaw/openclaw/issues/52952 | https://api.github.com/repos/openclaw/openclaw/issues/52952 | Windows Telegram execs fail in allowlist/command-rebuild path | ### Summary
On Windows, Telegram-triggered execs still fail in OpenClaw's allowlist/command-rebuild path before approval can help. The gateway restarts cleanly, but a simple PowerShell command is rebuilt into an invalid PowerShell invocation and/or denied with `exec denied: allowlist miss`.
### Environment
- OpenClaw... | closed | completed | true | 2 | [] | [] | 2026-03-23T15:08:01Z | 2026-04-30T04:46:45Z | 2026-04-25T06:10:57Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | jim777156 | 49,817,933 | MDQ6VXNlcjQ5ODE3OTMz | User | false |
openclaw/openclaw | 4,205,248,358 | I_kwDOQb6kR876pv9m | 60,935 | https://github.com/openclaw/openclaw/issues/60935 | https://api.github.com/repos/openclaw/openclaw/issues/60935 | Feature Request: Complete Chinese (zh-CN) i18n translation for Control UI | The OpenClaw Control UI has partial Chinese (zh-CN) translation but several sections are missing.
**Missing sections:**
- Usage page (usage section) - completely missing
- Dreams tab (tabs.dreams, subtitles.dreams)
- Cron jobs: schedule, lastRun, reset fields
- Common: theme field
- Chat: toolCallsToggle field
**Im... | closed | completed | true | 2 | [] | [] | 2026-04-04T16:04:35Z | 2026-04-28T05:10:07Z | 2026-04-25T13:48:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | xiep-dot | 266,485,161 | U_kgDOD-I9qQ | User | false |
openclaw/openclaw | 4,117,354,672 | I_kwDOQb6kR871adiw | 52,510 | https://github.com/openclaw/openclaw/issues/52510 | https://api.github.com/repos/openclaw/openclaw/issues/52510 | doctor: false positive cleanup hint for active gateway LaunchAgent plist | ## Bug Description
`openclaw doctor` suggests removing the active gateway LaunchAgent plist as a "cleanup hint," even when that plist is the currently running gateway service.
## Steps to Reproduce
1. Install OpenClaw on macOS with LaunchAgent-managed gateway (standard setup)
2. Run `openclaw doctor`
3. Observe outp... | open | null | false | 3 | [] | [] | 2026-03-22T23:16:11Z | 2026-05-04T19:31:34Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | daniel-knox | 34,885,600 | MDQ6VXNlcjM0ODg1NjAw | User | false |
openclaw/openclaw | 4,221,640,526 | I_kwDOQb6kR877oR9O | 62,844 | https://github.com/openclaw/openclaw/issues/62844 | https://api.github.com/repos/openclaw/openclaw/issues/62844 | [Bug]: Telegram loader/config-read bug | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Summary
OpenClaw intermittently fails while reading config in CLI commands when Telegram is enabled, with this error:
```text
Failed to read config at /home/node/.openclaw/openclaw.json
TypeError: Cannot read properties... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-04-08T02:14:59Z | 2026-04-28T05:05:10Z | 2026-04-25T14:46:34Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | GGaulin | 2,190,491 | MDQ6VXNlcjIxOTA0OTE= | User | false |
openclaw/openclaw | 4,271,403,893 | I_kwDOQb6kR87-mHN1 | 67,367 | https://github.com/openclaw/openclaw/issues/67367 | https://api.github.com/repos/openclaw/openclaw/issues/67367 | zeroclaw v0.6.9: Four blocking issues for Python-based agentic skills (native runtime, PYTHONPATH, prompt injection, sandbox defaults) | # zeroclaw v0.6.9: Four Blocking Issues for Real-World Python-Based Agentic Skills
**Reporter**: Jason Perlow
**Affiliation**: Contributing in a personal capacity, independent of employer
**zeroclaw version**: 0.6.9
**Platform**: aarch64 (Raspberry Pi 4, 2GB), Raspberry Pi OS Bookworm
**Skill under development... | closed | not_planned | true | 1 | [] | [] | 2026-04-15T20:08:21Z | 2026-04-30T04:39:56Z | 2026-04-26T05:12:43Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | perlowja | 162,375,972 | U_kgDOCa2pJA | User | false |
openclaw/openclaw | 4,070,707,116 | I_kwDOQb6kR87yog-s | 45,064 | https://github.com/openclaw/openclaw/issues/45064 | https://api.github.com/repos/openclaw/openclaw/issues/45064 | [Bug]: 2026.3.12 Memory Leak - OOM on basic commands (gateway status, doctor) | ### Bug type
Crash (process/app exits or hangs)
### Summary
After upgrading to 2026.3.12, OpenClaw crashes with JavaScript heap out of memory on basic CLI commands, making the CLI completely unusable.
### Steps to reproduce
1. Install OpenClaw 2026.3.12:
npm install -g openclaw@2026.3.12
2. Run any simple comm... | closed | completed | true | 31 | [
"bug",
"bug:crash"
] | [
"vincentkoc"
] | 2026-03-13T12:06:42Z | 2026-04-18T04:14:12Z | 2026-04-15T19:15:53Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | binbin1213 | 32,067,712 | MDQ6VXNlcjMyMDY3NzEy | User | false |
openclaw/openclaw | 4,271,123,929 | I_kwDOQb6kR87-lC3Z | 67,348 | https://github.com/openclaw/openclaw/issues/67348 | https://api.github.com/repos/openclaw/openclaw/issues/67348 | feat(agent): add --model flag to openclaw agent for per-invocation model selection | ## Summary
`openclaw agent` currently has no `--model` flag. The only way to switch the model for an agent invocation is `openclaw models set <id>`, which writes to `openclaw.json` and triggers a gateway restart via the file watcher. This makes programmatic multi-model workflows (test harnesses, CI pipelines, per-task... | closed | completed | true | 2 | [] | [] | 2026-04-15T19:14:11Z | 2026-04-21T04:21:47Z | 2026-04-18T17:21:05Z | NONE | null | 20260421T060038Z | 2026-04-21T06:00:38Z | perlowja | 162,375,972 | U_kgDOCa2pJA | User | false |
openclaw/openclaw | 4,188,520,175 | I_kwDOQb6kR875p77v | 59,162 | https://github.com/openclaw/openclaw/issues/59162 | https://api.github.com/repos/openclaw/openclaw/issues/59162 | [Bug]: Telegram/mobile approvals can resolve as "unknown or expired approval id" after pending approval creation | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Telegram/mobile approval flows sometimes fail with unknown or expired approval id even after approval creation was observed in logs.
### Steps to reproduce
1. Trigger a command that requires approval from a T... | closed | completed | true | 5 | [
"bug",
"bug:behavior"
] | [] | 2026-04-01T17:09:27Z | 2026-05-01T04:43:52Z | 2026-04-28T10:38:58Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | wikithoughts | 7,737,137 | MDQ6VXNlcjc3MzcxMzc= | User | false |
openclaw/openclaw | 4,168,466,905 | I_kwDOQb6kR874dcHZ | 57,574 | https://github.com/openclaw/openclaw/issues/57574 | https://api.github.com/repos/openclaw/openclaw/issues/57574 | Add MiniMax as image-generation provider (support image-01 model) | ## Feature Request: MiniMax Image Generation Provider
### What do you want?
Add MiniMax as a registered image-generation provider in OpenClaw, similar to how Google Gemini is already supported. This would allow users to use MiniMax's image-01 model via the `image_generate` tool.
### What's the current situation?
1. O... | closed | completed | true | 2 | [] | [] | 2026-03-30T08:01:17Z | 2026-04-26T04:29:31Z | 2026-04-24T00:48:21Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | 0xcjl | 38,884,292 | MDQ6VXNlcjM4ODg0Mjky | User | false |
openclaw/openclaw | 4,103,043,152 | I_kwDOQb6kR870j3hQ | 50,547 | https://github.com/openclaw/openclaw/issues/50547 | https://api.github.com/repos/openclaw/openclaw/issues/50547 | [Bug]: Discord bot does not reply to messages after onboarding" | ### Bug type
Regression (worked before, now fails)
### Summary
# Discord Integration Bug Report
## Issue
OpenClaw onboarding wizard completes successfully, but Discord bot does not respond to messages.
## Environment
- **OpenClaw Version:** 2026.3.13
- **Host:** macOS (Mac mini)
- **Node:** v22.22.1
- **Model:** a... | closed | not_planned | true | 5 | [
"bug",
"regression"
] | [] | 2026-03-19T17:07:59Z | 2026-04-30T04:54:27Z | 2026-04-25T04:49:33Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | ernestclawmac-lgtm | 269,536,155 | U_kgDOEBDLmw | User | false |
openclaw/openclaw | 4,169,747,099 | I_kwDOQb6kR874iUqb | 57,661 | https://github.com/openclaw/openclaw/issues/57661 | https://api.github.com/repos/openclaw/openclaw/issues/57661 | [Bug]: agent-to-agent session_send timeout | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
项目是开发团队模拟，使用了session_send 功能，一对多的方式沟通，大部分时间会出现session_send timeout。仅很小的几率会成功
<img width="497" height="383" alt="Image" src="https://github.com/user-attachments/assets/f994c737-2c39-48b9-af80-1fe34152a532" />
... | open | null | false | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-03-30T11:01:21Z | 2026-05-02T03:02:46Z | null | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | opensm | 11,435,512 | MDQ6VXNlcjExNDM1NTEy | User | false |
openclaw/openclaw | 4,223,190,605 | I_kwDOQb6kR877uMZN | 63,035 | https://github.com/openclaw/openclaw/issues/63035 | https://api.github.com/repos/openclaw/openclaw/issues/63035 | [Bug]: `JSON.stringify` silently drops UUID keys from array-typed `pending.json`, breaking all device pairing | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
`pending.json` and `paired.json` in the devices state directory can end up containing `[]` (empty JSON arrays) instead of `{}` (empty objects). When `loadState` reads these files, the `pending ?? {}` fallback p... | closed | completed | true | 6 | [
"bug",
"bug:behavior"
] | [] | 2026-04-08T08:32:26Z | 2026-05-01T04:43:54Z | 2026-04-28T10:38:55Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | sar618 | 214,745,104 | U_kgDODMzAEA | User | false |
openclaw/openclaw | 4,234,864,988 | I_kwDOQb6kR878aulc | 63,921 | https://github.com/openclaw/openclaw/issues/63921 | https://api.github.com/repos/openclaw/openclaw/issues/63921 | [Bug]: Dreaming/REM surfaces raw session-corpus metadata (untrusted, user, assistant) instead of meaningful reflections | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
# Dreaming bug report: raw session-corpus noise pollutes REM/Dreaming output
## Summary
On OpenClaw 2026.4.9, Dreaming appears healthy at a service level but the REM/Dreaming pipeline is ingesting noisy raw session-corpus l... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-09T20:56:53Z | 2026-04-30T04:42:39Z | 2026-04-26T04:43:01Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | chall-gcpl | 163,865,947 | U_kgDOCcRlWw | User | false |
openclaw/openclaw | 4,236,529,959 | I_kwDOQb6kR878hFEn | 64,074 | https://github.com/openclaw/openclaw/issues/64074 | https://api.github.com/repos/openclaw/openclaw/issues/64074 | Support Aliyun MaaS as Custom Model Provider | ### Summary
Allow custom OpenAI-compatible model providers via config.patch
### Problem to solve
Relax models.providers schema to allow arbitrary provider keys following OpenAI-compatible structure: {api, apiKey, baseUrl, models:[{id,name,...}]}
### Proposed solution
Relax models.providers schema to allow arbitrar... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-04-10T04:19:20Z | 2026-04-28T05:02:00Z | 2026-04-25T15:56:37Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | m013020082-dev | 274,596,041 | U_kgDOEF4AyQ | User | false |
openclaw/openclaw | 4,163,581,456 | I_kwDOQb6kR874KzYQ | 56,920 | https://github.com/openclaw/openclaw/issues/56920 | https://api.github.com/repos/openclaw/openclaw/issues/56920 | [Feature]: Support custom inference servers (OpenAI-compatible) | ### Summary
Support custom inference servers (OpenAI-compatible) with Basic authentication and enable secure private network access for self-hosted and local endpoints.
My Motivation is users need flexibility to:
- Integrate with self-hosted or third-party inference APIs
- Use non-Bearer authentication methods (e.g.,... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-03-29T09:42:34Z | 2026-04-28T05:19:54Z | 2026-04-25T11:52:55Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | leemgs | 82,404 | MDQ6VXNlcjgyNDA0 | User | false |
openclaw/openclaw | 4,052,181,859 | I_kwDOQb6kR87xh2Nj | 42,225 | https://github.com/openclaw/openclaw/issues/42225 | https://api.github.com/repos/openclaw/openclaw/issues/42225 | openai-codex/gpt-5.4 still uses 272000 context for catalog + runtime context engineering despite PR #37876 | ## Summary
After updating to the current OpenClaw release, `openclaw models list` still shows `openai-codex/gpt-5.4` as ~266k context, and the runtime context-engineering path is also using `272000` for pruning, compaction, and token budgeting.
PR #37876 added a forward-compat patch for GPT-5.4 with `contextWindow = 1... | closed | completed | true | 9 | [] | [] | 2026-03-10T14:26:09Z | 2026-04-27T04:53:21Z | 2026-04-24T22:11:24Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | luizlf | 25,336,587 | MDQ6VXNlcjI1MzM2NTg3 | User | false |
openclaw/openclaw | 4,163,297,696 | I_kwDOQb6kR874JuGg | 56,874 | https://github.com/openclaw/openclaw/issues/56874 | https://api.github.com/repos/openclaw/openclaw/issues/56874 | Discord voice: @discordjs/voice MODULE_NOT_FOUND after v2026.3.28 update + voice.tts config breaking change | ### Bug Description
After updating to **v2026.3.28**, Discord voice joining fails with `Cannot find module '@discordjs/voice'`. Additionally, the `channels.discord.voice.tts` config schema changed in a breaking way without migration or changelog notice.
---
### Issue 1: `@discordjs/voice` not resolvable from runtime... | closed | completed | true | 1 | [] | [] | 2026-03-29T08:10:34Z | 2026-04-18T04:14:18Z | 2026-04-15T14:26:30Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | darealgege | 30,112,470 | MDQ6VXNlcjMwMTEyNDcw | User | false |
openclaw/openclaw | 4,179,191,919 | I_kwDOQb6kR875GWhv | 58,385 | https://github.com/openclaw/openclaw/issues/58385 | https://api.github.com/repos/openclaw/openclaw/issues/58385 | Bug: web_fetch tool validation error - 'url: must have required property' | **Describe the bug**
The `web_fetch` tool fails with a validation error complaining about the missing `url` property when used with specific models (e.g., `kimi/k2p5`). The tool works fine with other models (e.g., `google/gemini-3.1-pro-preview`).
**To Reproduce**
Steps to reproduce the behavior:
1. Set the model to `... | closed | completed | true | 3 | [] | [] | 2026-03-31T13:24:31Z | 2026-04-18T04:14:19Z | 2026-04-15T14:26:17Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | darealgege | 30,112,470 | MDQ6VXNlcjMwMTEyNDcw | User | false |
openclaw/openclaw | 4,198,766,164 | I_kwDOQb6kR876RBZU | 60,113 | https://github.com/openclaw/openclaw/issues/60113 | https://api.github.com/repos/openclaw/openclaw/issues/60113 | TypeError: Cannot read properties of undefined (reading "replace") during embedded agent run (v2026.4.1) | ## Bug Description
Embedded agent runs intermittently crash with:
```
TypeError: Cannot read properties of undefined (reading 'replace')
```
Logged as `embedded run agent end: isError=true` with no full stack trace (only `error.message` is captured).
This occurs in **two distinct trigger patterns**:
### Pattern 1:... | open | null | false | 2 | [] | [] | 2026-04-03T06:44:55Z | 2026-05-02T09:22:03Z | null | CONTRIBUTOR | null | 20260502T120029Z | 2026-05-02T12:00:29Z | wujiaming88 | 7,887,545 | MDQ6VXNlcjc4ODc1NDU= | User | false |
openclaw/openclaw | 4,208,222,842 | I_kwDOQb6kR8761GJ6 | 61,455 | https://github.com/openclaw/openclaw/issues/61455 | https://api.github.com/repos/openclaw/openclaw/issues/61455 | [Bug]: Tool calling timeouts. | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
I've been getting consistent timeout errors when OpenClaw is running tool calls, even with a good gateway connection. I've had OpenClaw try to diagnose itself. These were the findings:
Still the same error. Even with allowIns... | closed | completed | true | 4 | [
"bug",
"bug:crash"
] | [] | 2026-04-05T17:54:40Z | 2026-04-28T04:33:35Z | 2026-04-26T04:27:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | dtroupe2 | 70,831,799 | MDQ6VXNlcjcwODMxNzk5 | User | false |
openclaw/openclaw | 4,202,105,251 | I_kwDOQb6kR876dwmj | 60,518 | https://github.com/openclaw/openclaw/issues/60518 | https://api.github.com/repos/openclaw/openclaw/issues/60518 | Groq media understanding provider dropped from registry after first message — voice transcription works once then fails | ## Description
Voice message transcription (Groq Whisper) works exactly **once** after gateway restart, then silently fails for all subsequent voice messages with:
```
Error: Media provider not available: groq
```
The error is caught and silently swallowed by `runAttachmentEntries` in `runner.ts`, logged only via `l... | closed | completed | true | 2 | [] | [] | 2026-04-03T20:33:05Z | 2026-04-28T05:10:54Z | 2026-04-25T13:40:01Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | KurtVogel | 22,785,743 | MDQ6VXNlcjIyNzg1NzQz | User | false |
openclaw/openclaw | 4,269,036,454 | I_kwDOQb6kR87-dFOm | 67,180 | https://github.com/openclaw/openclaw/issues/67180 | https://api.github.com/repos/openclaw/openclaw/issues/67180 | bug(web-ui): streaming markdown flickers — full DOM replacement on every token via unsafeHTML | ## Summary
- **Problem:** During streaming, every new token causes the full message container to flicker. Text visibly disappears and re-appears, especially noticeable with paragraphs and code fences.
- **Why it matters:** Any skill that produces structured markdown output (code blocks, JSON fences, multi-paragraph re... | closed | completed | true | 1 | [] | [] | 2026-04-15T12:50:49Z | 2026-04-21T04:21:50Z | 2026-04-18T17:21:07Z | NONE | null | 20260421T060038Z | 2026-04-21T06:00:38Z | perlowja | 162,375,972 | U_kgDOCa2pJA | User | false |
openclaw/openclaw | 4,217,125,426 | I_kwDOQb6kR877XDoy | 62,438 | https://github.com/openclaw/openclaw/issues/62438 | https://api.github.com/repos/openclaw/openclaw/issues/62438 | [Bug]: security audit "config.secrets.hooks_token_in_config" fires for ${VAR} env refs because the check runs on the resolved config | ## Summary
`openclaw security audit --deep` reports `config.secrets.hooks_token_in_config` (INFO) and `config.secrets.gateway_password_in_config` (WARN) for fields that are *correctly* using the documented `${ENV_VAR}` substitution syntax.
The audit's `looksLikeEnvRef()` helper is intended to suppress those findings ... | open | null | false | 3 | [] | [] | 2026-04-07T10:44:31Z | 2026-05-02T06:31:24Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | ahuray | 54,942,761 | MDQ6VXNlcjU0OTQyNzYx | User | false |
openclaw/openclaw | 4,186,072,817 | I_kwDOQb6kR875gmbx | 58,963 | https://github.com/openclaw/openclaw/issues/58963 | https://api.github.com/repos/openclaw/openclaw/issues/58963 | [Feature Proposal] Migrate oh-my-openclaw multi-agent orchestration to official OpenClaw ecosystem | # Feature Proposal: oh-my-openclaw Migration & Community Maintenance
## 🎯 Summary
Propose migrating the **oh-my-openclaw** project (OmO-style multi-agent orchestration) into the official OpenClaw ecosystem, with community-driven maintenance and continued development to align with OpenClaw v2026.3.x+ plugin SDK.
## ... | closed | completed | true | 1 | [] | [] | 2026-04-01T11:09:24Z | 2026-04-18T04:14:23Z | 2026-04-15T11:53:16Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | lileilei-camera | 7,010,843 | MDQ6VXNlcjcwMTA4NDM= | User | false |
openclaw/openclaw | 4,175,800,105 | I_kwDOQb6kR8745acp | 58,140 | https://github.com/openclaw/openclaw/issues/58140 | https://api.github.com/repos/openclaw/openclaw/issues/58140 | [Bug]: Running official install command triggers Windows Defender - malicious batch file and registry keys added | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When running the official installation command from https://openclaw.ai, malicious changes are made to your computer that Windows Defender complains about loudly.
### Steps to reproduce
1. Open link to https:... | closed | completed | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-03-31T05:00:18Z | 2026-04-28T05:16:44Z | 2026-04-25T12:31:12Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | CodeHatchling | 7,232,874 | MDQ6VXNlcjcyMzI4NzQ= | User | false |
openclaw/openclaw | 4,191,142,978 | I_kwDOQb6kR875z8RC | 59,370 | https://github.com/openclaw/openclaw/issues/59370 | https://api.github.com/repos/openclaw/openclaw/issues/59370 | Bug: `logging.file` config ignored — bundled require("../config/config.js") fails at runtime | ## Description
The `logging.file` config option in `openclaw.json` is completely ignored. Logs are always written to the default `/tmp/openclaw/openclaw-YYYY-MM-DD.log` regardless of what `logging.file` is set to.
All other config options (models, channels, cron, etc.) load correctly — only logging is affected.
## R... | closed | completed | true | 5 | [] | [
"vincentkoc"
] | 2026-04-02T03:09:53Z | 2026-04-30T04:39:06Z | 2026-04-26T05:23:22Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | KeaneYan | 18,630,645 | MDQ6VXNlcjE4NjMwNjQ1 | User | false |
openclaw/openclaw | 4,221,312,754 | I_kwDOQb6kR877nB7y | 62,777 | https://github.com/openclaw/openclaw/issues/62777 | https://api.github.com/repos/openclaw/openclaw/issues/62777 | cron: failureAlert never fires + channel: prefix stripped in delivery path | ### Bug type
Bug / Silent Failure
### OpenClaw version
OpenClaw 2026.4.5 (3e72c03)
### OS and install method
Linux 6.8.0-106-generic (x64), Node.js v24.14.1, npm global, systemd user service.
### Summary
Cron job `failureAlert` is configured on multiple jobs but never fires. Every error run shows `deliveryStatus... | closed | completed | true | 2 | [] | [] | 2026-04-08T00:19:58Z | 2026-04-30T04:36:31Z | 2026-04-27T06:14:26Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | slideshow-dingo | 258,996,445 | U_kgDOD2_43Q | User | false |
openclaw/openclaw | 4,167,631,819 | I_kwDOQb6kR874aQPL | 57,506 | https://github.com/openclaw/openclaw/issues/57506 | https://api.github.com/repos/openclaw/openclaw/issues/57506 | tts tool ignores messages.tts.openai.baseUrl — hardcoded to Edge TTS regardless of config | ## Summary
The `tts` tool available to agents always uses Edge TTS internally, regardless of what is configured under `messages.tts`. This makes it impossible to use a custom OpenAI-compatible TTS endpoint (like a self-hosted Polly proxy, Piper, Coqui, etc.) from within agent tool calls.
## Environment
- **OpenClaw ... | closed | completed | true | 3 | [] | [] | 2026-03-30T05:48:52Z | 2026-04-27T04:38:06Z | 2026-04-25T03:02:51Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | cvictorv | 263,056,780 | U_kgDOD63tjA | User | false |
openclaw/openclaw | 4,105,509,840 | I_kwDOQb6kR870tRvQ | 50,806 | https://github.com/openclaw/openclaw/issues/50806 | https://api.github.com/repos/openclaw/openclaw/issues/50806 | allowBots: "mentions" — bot-authored messages with real @mention not triggering sessions | ## Description
Running a multi-agent setup (7 OpenClaw instances on the same Mac, each with its own Discord bot) and trying to enable bot-to-bot communication via `allowBots: "mentions"`.
## Setup
- OpenClaw version: 2026.3.13
- 7 agents, each with their own Discord bot token and gateway
- All in the same Discord gu... | closed | completed | true | 4 | [] | [] | 2026-03-20T03:37:03Z | 2026-04-30T04:53:42Z | 2026-04-25T04:56:14Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | franduranv | 41,132,175 | MDQ6VXNlcjQxMTMyMTc1 | User | false |
openclaw/openclaw | 4,189,583,069 | I_kwDOQb6kR875t_bd | 59,234 | https://github.com/openclaw/openclaw/issues/59234 | https://api.github.com/repos/openclaw/openclaw/issues/59234 | [Bug]: QMD boot probe triggers llama.cpp build on ARM Linux — searchMode: "search" should skip vector probe | ## Summary
When `memory.backend: "qmd"` is enabled with `searchMode: "search"` (BM25-only), OpenClaw still calls `qmd status` at boot to probe vector capability. On ARM Linux without Vulkan (e.g. Oracle Cloud aarch64), this causes QMD to attempt a full CMake + llama.cpp compile, fail, hit OpenClaw's 5s containment tim... | closed | completed | true | 8 | [] | [] | 2026-04-01T20:09:13Z | 2026-04-30T04:34:05Z | 2026-04-27T13:10:07Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | PrinceOfEgypt | 2,649,235 | MDQ6VXNlcjI2NDkyMzU= | User | false |
openclaw/openclaw | 3,929,772,503 | I_kwDOQb6kR87qO5HX | 14,438 | https://github.com/openclaw/openclaw/issues/14438 | https://api.github.com/repos/openclaw/openclaw/issues/14438 | [Feature]: Plugin hot-reload without container restart (jiti cache invalidation) | ## Description
When developing OpenClaw plugins (TypeScript), every code change requires:
1. Restart the OpenClaw process/container
2. Clear the jiti cache (`rm /tmp/jiti/plugin-name.*.cjs`)
3. Wait for the plugin to recompile and the server to reinitialize
This makes the plugin development cycle slow, especially in... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-12T04:14:01Z | 2026-05-05T02:00:42Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | badespider | 80,556,496 | MDQ6VXNlcjgwNTU2NDk2 | User | false |
openclaw/openclaw | 4,206,020,078 | I_kwDOQb6kR876ssXu | 61,077 | https://github.com/openclaw/openclaw/issues/61077 | https://api.github.com/repos/openclaw/openclaw/issues/61077 | Security: Exec approval prompts leak to chat channels | ## Bug Description
When OpenClaw's exec obfuscation detection triggers an approval prompt, the prompt text — including the full command with embedded secrets — can be routed to a connected chat channel (Microsoft Teams in our case) instead of being isolated to the internal terminal session.
## What Happened
1. VICTO... | closed | completed | true | 0 | [] | [] | 2026-04-04T23:38:01Z | 2026-04-18T04:14:24Z | 2026-04-15T10:02:12Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | blockmos | 247,715,528 | U_kgDODsPWyA | User | false |
openclaw/openclaw | 4,186,875,593 | I_kwDOQb6kR875jqbJ | 59,034 | https://github.com/openclaw/openclaw/issues/59034 | https://api.github.com/repos/openclaw/openclaw/issues/59034 | [macOS] screen.record command not available on node | ## 问题描述
macOS 节点缺少屏幕录制/截图相关命令。
## 环境
- Gateway: OpenClaw 2026.3.31
- macOS 节点: 曾蔚的MacBook Pro, macOS 26.3.0
## 复现步骤
1. 执行 `openclaw nodes screen record --node "曾蔚的MacBook Pro" --duration 1s`
2. 返回错误: `node command not allowed: "screen.record" is not in the allowlist for platform "macOS 26.3.0"`
## 实际行为
- 节点 ca... | closed | completed | true | 1 | [] | [] | 2026-04-01T13:17:35Z | 2026-04-28T05:14:41Z | 2026-04-25T12:55:40Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | nightq | 3,429,433 | MDQ6VXNlcjM0Mjk0MzM= | User | false |
openclaw/openclaw | 4,068,607,642 | I_kwDOQb6kR87yggaa | 44,610 | https://github.com/openclaw/openclaw/issues/44610 | https://api.github.com/repos/openclaw/openclaw/issues/44610 | Session stops processing new messages after prompt-error: aborted | ## Description
After a `prompt-error: aborted` event is recorded in a session JSONL, the session stops processing any new inbound messages. Messages sent to the group chat are silently dropped — they never appear in the session JSONL and no reply is generated.
## Steps to Reproduce
1. A group chat session is activel... | closed | completed | true | 3 | [] | [] | 2026-03-13T03:32:48Z | 2026-04-27T04:47:54Z | 2026-04-24T23:34:31Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | mcfn | 1,106,518 | MDQ6VXNlcjExMDY1MTg= | User | false |
openclaw/openclaw | 4,209,988,429 | I_kwDOQb6kR87671NN | 61,717 | https://github.com/openclaw/openclaw/issues/61717 | https://api.github.com/repos/openclaw/openclaw/issues/61717 | [Feature]: Add GitHub Copilot as memory search embedding provider | ### Summary
Add GitHub Copilot as a memory search embedding provider so users with Copilot subscriptions can use embeddings without a separate OpenAI/Gemini API key.
### Problem to solve
Users who have GitHub Copilot subscriptions (Individual, Business, or Enterprise) currently must configure a separate OpenAI, Gemi... | closed | completed | true | 0 | [] | [] | 2026-04-06T06:23:31Z | 2026-04-18T04:14:25Z | 2026-04-15T09:39:30Z | CONTRIBUTOR | null | 20260418T042035Z | 2026-04-18T04:20:35Z | feiskyer | 676,637 | MDQ6VXNlcjY3NjYzNw== | User | false |
openclaw/openclaw | 4,040,940,768 | I_kwDOQb6kR87w29zg | 39,672 | https://github.com/openclaw/openclaw/issues/39672 | https://api.github.com/repos/openclaw/openclaw/issues/39672 | [Feature]: Filesystem Access Control (allowedPaths / denyPaths) & PathGuard Canonicalization | ### Summary
Add filesystem-level access control (`allowedPaths`/`denyPaths`) to OpenClaw config safely by implementing a `PathGuard` canonicalization utility to prevent symlink and traversal bypasses.
### Problem to solve
OpenClaw currently lacks path-level file restrictions (agents with filesystem access can read `... | closed | not_planned | true | 4 | [
"enhancement"
] | [] | 2026-03-08T09:32:54Z | 2026-04-28T04:43:05Z | 2026-04-26T02:03:10Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Ar3ss12 | 150,044,748 | U_kgDOCPGATA | User | false |
openclaw/openclaw | 4,113,222,768 | I_kwDOQb6kR871Ksxw | 51,794 | https://github.com/openclaw/openclaw/issues/51794 | https://api.github.com/repos/openclaw/openclaw/issues/51794 | secretref-env: not working for Discord token in channels.discord.token | ## Description
The `secretref-env:` format is not working correctly for the Discord token in `channels.discord.token` configuration.
## Steps to Reproduce
1. Set Discord token in systemd service environment:
```
Environment=OPENCLAW_DISCORD_TOKEN=MTQ2OTgyNjA0MjU0NDkxODYyMA.GvPdPl...
```
2. Configure Discor... | closed | completed | true | 3 | [] | [] | 2026-03-21T18:01:51Z | 2026-04-27T04:38:54Z | 2026-04-25T02:49:50Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | halointellicore | 251,796,857 | U_kgDODwIdeQ | User | false |
openclaw/openclaw | 4,181,724,567 | I_kwDOQb6kR875QA2X | 58,535 | https://github.com/openclaw/openclaw/issues/58535 | https://api.github.com/repos/openclaw/openclaw/issues/58535 | [Bug]: Discord announce removing fields from my input | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
I have a cron job running that specifically says 'Output this file verbatim' - the file is generated by a `system cron` 5 minutes before that, simply because it was timing out when being run by `openclaw cron` ... | closed | completed | true | 5 | [
"bug",
"bug:behavior"
] | [] | 2026-03-31T19:18:09Z | 2026-04-18T04:14:26Z | 2026-04-15T08:40:29Z | NONE | null | 20260418T042035Z | 2026-04-18T04:20:35Z | JayCroghan | 1,171,148 | MDQ6VXNlcjExNzExNDg= | User | false |
openclaw/openclaw | 4,062,131,305 | I_kwDOQb6kR87yHzRp | 43,680 | https://github.com/openclaw/openclaw/issues/43680 | https://api.github.com/repos/openclaw/openclaw/issues/43680 | Agent model configuration ignored—defaults override explicit agent.list[*].model settings | ## Bug Report
### Version
OpenClaw 2026.3.8 (latest)
### Description
Agent-level model configurations specified in `agents.list[*].model` are being ignored. Instead, the system defaults to `agents.defaults.model.primary` (anthropic/claude-haiku-4-5) regardless of explicit agent config.
This occurs even when:
- Agent... | closed | completed | true | 3 | [] | [] | 2026-03-12T04:49:43Z | 2026-04-27T04:50:25Z | 2026-04-24T22:21:39Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Seabass-up | 184,999,458 | U_kgDOCwbeIg | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.