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 | 4,008,993,707 | I_kwDOQb6kR87u9GOr | 31,206 | https://github.com/openclaw/openclaw/issues/31206 | https://api.github.com/repos/openclaw/openclaw/issues/31206 | Feature: skills.descriptionMode for token-efficient skill injection | ## Problem
OpenClaw injects all skill descriptions into the system prompt every session. With 70+ skills, descriptions consume ~1,100-3,400 tokens depending on verbosity — most of which are wasted because only 1-2 skills are used per session.
This mirrors the problem Cursor solved with **lazy MCP tool loading** (46.9... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T02:57:56Z | 2026-04-15T04:21:21Z | 2026-04-12T04:25:33Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | aspenas | 25,066,580 | MDQ6VXNlcjI1MDY2NTgw | User | false |
openclaw/openclaw | 4,008,897,935 | I_kwDOQb6kR87u8u2P | 31,175 | https://github.com/openclaw/openclaw/issues/31175 | https://api.github.com/repos/openclaw/openclaw/issues/31175 | Windows Node: exec-approvals socket/Named Pipe not created automatically | ## Summary
Windows nodes fail to create the `exec-approvals` socket/Named Pipe automatically on startup, causing all remote command executions to fail with timeout errors.
## Environment
- **OS**: Windows 11 Pro Build 26200.7840
- **OpenClaw Version**: 2026.2.15 (confirmed still present in 2026.2.26)
- **Hardware**:... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T02:16:46Z | 2026-04-15T04:21:21Z | 2026-04-12T04:25:30Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | jgarciac | 7,825,683 | MDQ6VXNlcjc4MjU2ODM= | User | false |
openclaw/openclaw | 4,008,994,654 | I_kwDOQb6kR87u9Gde | 31,208 | https://github.com/openclaw/openclaw/issues/31208 | https://api.github.com/repos/openclaw/openclaw/issues/31208 | Feature: tools.byChannel for channel-specific tool profiles | ## Problem
All tools are loaded into every session regardless of channel or session type. Heartbeat cron sessions get the full `browser` schema (~2,453 tokens) despite never using it. iMessage sessions get `canvas` despite it being a webchat-only feature.
Tool schemas are a significant chunk of context — the docs ex... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T02:58:18Z | 2026-04-15T04:21:20Z | 2026-04-12T04:25:35Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | aspenas | 25,066,580 | MDQ6VXNlcjI1MDY2NTgw | User | false |
openclaw/openclaw | 4,009,020,182 | I_kwDOQb6kR87u9MsW | 31,214 | https://github.com/openclaw/openclaw/issues/31214 | https://api.github.com/repos/openclaw/openclaw/issues/31214 | Security: Possible phishing/typosquatting domain (og-openclaw.com) | Potential phishing / typosquatting domain detected: `https://og-openclaw.com/`
## Summary
A user reported this domain as potentially related to OpenClaw. It appears suspicious and likely impersonation/typosquatting.
## Why it looks suspicious
- Uses lookalike naming (`og-openclaw.com`) instead of official OpenClaw do... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T03:08:26Z | 2026-04-15T04:21:19Z | 2026-04-12T04:25:38Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | AZRIELA | 21,047,521 | MDQ6VXNlcjIxMDQ3NTIx | User | false |
openclaw/openclaw | 4,009,019,271 | I_kwDOQb6kR87u9MeH | 31,212 | https://github.com/openclaw/openclaw/issues/31212 | https://api.github.com/repos/openclaw/openclaw/issues/31212 | message:received hook does not fire for Discord guild channel messages | ## Summary
`message:received` internal hooks do not fire when messages are received in Discord guild channels. They only fire for DMs and WhatsApp messages. Guild messages are routed directly to guild sessions, bypassing the hook event system.
## Steps to Reproduce
1. Create an internal hook listening for `message:r... | closed | not_planned | true | 4 | [
"stale"
] | [] | 2026-03-02T03:08:01Z | 2026-04-15T04:21:19Z | 2026-04-12T04:25:37Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | wave-workflow | 264,320,090 | U_kgDOD8E0Wg | User | false |
openclaw/openclaw | 4,009,041,101 | I_kwDOQb6kR87u9RzN | 31,222 | https://github.com/openclaw/openclaw/issues/31222 | https://api.github.com/repos/openclaw/openclaw/issues/31222 | WhatsApp voice note media download produces truncated/corrupt files | ## Bug Report
**Version:** 2026.2.26
**Channel:** WhatsApp (Baileys)
**OS:** macOS (Apple Silicon M-series)
### Description
WhatsApp voice notes are being saved to disk but the audio payload is truncated/corrupt. Files that should be 15-30KB for a 15-second voice note are only ~1KB. When decoded with ffmpeg, the PCM ... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T03:17:11Z | 2026-04-15T04:21:18Z | 2026-04-12T04:25:40Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | emilywinslow-source | 260,186,948 | U_kgDOD4IjRA | User | false |
openclaw/openclaw | 4,009,082,938 | I_kwDOQb6kR87u9cA6 | 31,236 | https://github.com/openclaw/openclaw/issues/31236 | https://api.github.com/repos/openclaw/openclaw/issues/31236 | Feature: Configurable display name for typing indicator (e.g. gateway.displayName) | ## Feature Request
### Problem
When using the Google Chat channel, the typing indicator message displays **"OpenClaw"** regardless of the bot's configured app name. This is confusing for end-users who see their bot (e.g. "Slay") show an "OpenClaw is typing" indicator.
### Proposed Solution
Add a `gateway.displayNam... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T03:36:55Z | 2026-04-15T04:21:17Z | 2026-04-12T04:25:42Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | jrimmer | 629 | MDQ6VXNlcjYyOQ== | User | false |
openclaw/openclaw | 4,016,839,786 | I_kwDOQb6kR87vbBxq | 33,087 | https://github.com/openclaw/openclaw/issues/33087 | https://api.github.com/repos/openclaw/openclaw/issues/33087 | 请求:在 Webchat 界面显示 AI 响应时长 | ## 功能请求
### 描述
在 Webchat 聊天界面中,希望能够显示 AI 模型的响应时长(response time/duration)。
### 当前情况
- 界面只显示时间戳
- 没有显示每次对话的响应耗时
### 期望
- 在消息旁边显示响应时长,比如 "2.3s" 或 "2.3秒"
- 或者在会话详情中显示总响应时间
### 场景
- 方便用户了解 AI 响应速度
- 用于性能监控和优化
--- | closed | not_planned | true | 4 | [
"stale"
] | [] | 2026-03-03T14:00:36Z | 2026-04-22T04:30:19Z | 2026-04-20T04:26:51Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | huanbao-oc | 265,213,352 | U_kgDOD87VqA | User | false |
openclaw/openclaw | 4,009,088,323 | I_kwDOQb6kR87u9dVD | 31,237 | https://github.com/openclaw/openclaw/issues/31237 | https://api.github.com/repos/openclaw/openclaw/issues/31237 | Heartbeat queue stuck - polls queued but never drain (queued=1 permanent) | ## Heartbeat queue stuck - polls never drain
**Version:** 2026.2.26 (beta channel)
**OS:** macOS 15.3 (arm64, Mac Mini M4)
**Node:** v25.6.0
### Description
Heartbeat polls get queued but never execute. The diagnostic health-monitor shows `queued=1` persistently, even when no sessions are active (`active=0`).
### St... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T03:39:23Z | 2026-04-15T04:21:16Z | 2026-04-12T04:25:43Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Sulto4 | 188,591,084 | U_kgDOCz2r7A | User | false |
openclaw/openclaw | 4,009,117,604 | I_kwDOQb6kR87u9kek | 31,247 | https://github.com/openclaw/openclaw/issues/31247 | https://api.github.com/repos/openclaw/openclaw/issues/31247 | Dashboard: Channels schema errors, Usage page empty, Telegram health mismatch | ### Bug type
Regression (worked before, now fails)
### Summary
Three related dashboard UI issues on 2026.2.26:
1. **Channels page**: Accounts and Capabilities sections show `Unsupported schema node. Use Raw mode.` in Form mode for Telegram config
2. **Usage page**: Opens with 0 data / empty charts. Requires manual ... | closed | not_planned | true | 3 | [
"bug",
"stale",
"regression"
] | [] | 2026-03-02T03:51:35Z | 2026-04-15T04:21:15Z | 2026-04-12T04:25:47Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | nasacontrol-hq | 261,863,211 | U_kgDOD5u3Kw | User | false |
openclaw/openclaw | 4,009,113,700 | I_kwDOQb6kR87u9jhk | 31,246 | https://github.com/openclaw/openclaw/issues/31246 | https://api.github.com/repos/openclaw/openclaw/issues/31246 | [Bug]: message tool (action: send) prevents Discord slash command interaction from resolving | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When a user-invocable skill uses the message tool (action: send) during a Discord native slash command turn, the interaction never resolves — Discord permanently shows "Bot is responding..." even though the message content is delivered to t... | closed | not_planned | true | 2 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-02T03:49:54Z | 2026-04-15T04:21:15Z | 2026-04-12T04:25:45Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | xueqino1 | 20,064,641 | MDQ6VXNlcjIwMDY0NjQx | User | false |
openclaw/openclaw | 4,009,120,404 | I_kwDOQb6kR87u9lKU | 31,249 | https://github.com/openclaw/openclaw/issues/31249 | https://api.github.com/repos/openclaw/openclaw/issues/31249 | Feature: Session lifecycle hooks for skills/plugins (onPreCompact, onSessionStart, onSessionEnd) | ## Problem
The current `compaction.memoryFlush.prompt` approach for pre-compaction context persistence is fragile — it relies on the agent correctly executing shell commands *during* the flush window. This is non-deterministic: if the LLM output is malformed, the flush silently fails and context is lost.
There is cur... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T03:52:59Z | 2026-04-15T04:21:14Z | 2026-04-12T04:25:48Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | AlexChen31337 | 259,467,130 | U_kgDOD3cneg | User | false |
openclaw/openclaw | 4,009,128,703 | I_kwDOQb6kR87u9nL_ | 31,251 | https://github.com/openclaw/openclaw/issues/31251 | https://api.github.com/repos/openclaw/openclaw/issues/31251 | [Bug]: plugins install/uninstall can clobber unrelated openclaw.json keys (dmPolicy reverts) | # [Bug]: plugins install/uninstall can clobber unrelated openclaw.json keys (dmPolicy reverts)
## Summary
Plugin install/uninstall rewrites `openclaw.json` in a way that can overwrite unrelated keys (e.g., `channels.telegram.dmPolicy` reverting from `"open"` to `"pairing"`).
## Steps to reproduce
1. Set `channels.tel... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T03:56:53Z | 2026-04-15T04:21:13Z | 2026-04-12T04:25:50Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | solstead | 168,413,654 | U_kgDOCgnJ1g | User | false |
openclaw/openclaw | 4,009,145,494 | I_kwDOQb6kR87u9rSW | 31,260 | https://github.com/openclaw/openclaw/issues/31260 | https://api.github.com/repos/openclaw/openclaw/issues/31260 | Discord bot shows as offline despite being functional | ## Description
Discord bot appears offline (gray status) even though messages are being sent and received successfully.
## Environment
- OpenClaw version: 2026.2.3-1
- OS: macOS (Darwin 25.2.0 arm64)
## Steps to Reproduce
1. Configure Discord channel with a bot token
2. Gateway starts and Discord plugin connects
3. B... | closed | not_planned | true | 4 | [
"stale"
] | [] | 2026-03-02T04:03:24Z | 2026-04-15T04:21:12Z | 2026-04-12T04:25:53Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | niMgnoeSeeL | 13,412,318 | MDQ6VXNlcjEzNDEyMzE4 | User | false |
openclaw/openclaw | 4,009,145,241 | I_kwDOQb6kR87u9rOZ | 31,259 | https://github.com/openclaw/openclaw/issues/31259 | https://api.github.com/repos/openclaw/openclaw/issues/31259 | [Feature]: Extract shared helpers from session-cost-usage.ts to enable multi-agent and deleted-session fixes | ### Summary
Extract canonical file-private helpers from the 1016-line `session-cost-usage.ts` god file so that multi-agent aggregation, deleted-session scanning, and duplicate-agent fixes (#24793, #20558, #25461) can be implemented without navigating three tightly-coupled monolithic functions.
### Problem to solve
`... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T04:03:21Z | 2026-04-15T04:21:12Z | 2026-04-12T04:25:52Z | MEMBER | null | 20260417T163222Z | 2026-04-17T16:32:22Z | mcaxtr | 7,562,095 | MDQ6VXNlcjc1NjIwOTU= | User | false |
openclaw/openclaw | 4,009,171,123 | I_kwDOQb6kR87u9xiz | 31,266 | https://github.com/openclaw/openclaw/issues/31266 | https://api.github.com/repos/openclaw/openclaw/issues/31266 | session-memory hook should support session:end events | The bundled session-memory hook only listens to command:new and command:reset events. It should also support session:end to auto-save conversation context when sessions timeout or end naturally.
Currently:
- Hook only saves on /new or /reset commands (when starting NEW session)
- Sessions that end via timeout are not ... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T04:12:22Z | 2026-04-15T04:21:11Z | 2026-04-12T04:25:55Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | sj-unit72 | 19,424,303 | MDQ6VXNlcjE5NDI0MzAz | User | false |
openclaw/openclaw | 4,009,208,464 | I_kwDOQb6kR87u96qQ | 31,273 | https://github.com/openclaw/openclaw/issues/31273 | https://api.github.com/repos/openclaw/openclaw/issues/31273 | ACPX plugin: claude prompt --session fails with NO_SESSION (never creates session) | ## Bug
ACP sessions via `sessions_spawn(runtime: 'acp', agentId: 'claude')` always fail with `ACP_TURN_FAILED`.
## Root Cause
The ACPX plugin runtime calls:
```
acpx claude prompt --session <name> --file -
```
But `acpx claude prompt --session` requires a **pre-existing** session. It never calls `acpx claude sessio... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T04:26:51Z | 2026-04-15T04:21:10Z | 2026-04-12T04:25:57Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | sparkus | 2,079,663 | MDQ6VXNlcjIwNzk2NjM= | User | false |
openclaw/openclaw | 4,009,210,126 | I_kwDOQb6kR87u97EO | 31,275 | https://github.com/openclaw/openclaw/issues/31275 | https://api.github.com/repos/openclaw/openclaw/issues/31275 | [Bug]: Session-memory hook not triggered when using /new command in Feishu | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
The session-memory hook is not triggered when using the `/new` command in Feishu channel, but it works correctly when using the same command in WebUI. This results in session context not being saved to memory files with the expected date-sl... | closed | not_planned | true | 3 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-02T04:27:31Z | 2026-04-15T04:21:09Z | 2026-04-12T04:25:59Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Spirit597 | 38,465,227 | MDQ6VXNlcjM4NDY1MjI3 | User | false |
openclaw/openclaw | 4,009,242,536 | I_kwDOQb6kR87u-C-o | 31,280 | https://github.com/openclaw/openclaw/issues/31280 | https://api.github.com/repos/openclaw/openclaw/issues/31280 | gateway restart CLI fails on macOS LaunchAgent (kickstart service not found) | ## Description
`openclaw gateway restart` fails on macOS with LaunchAgent service management. The CLI attempts to use `launchctl kickstart` but cannot find the service, causing the gateway process to die without a clean restart.
## Steps to Reproduce
1. Install OpenClaw with LaunchAgent (`ai.openclaw.gateway` plist)... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T04:40:03Z | 2026-04-15T04:21:08Z | 2026-04-12T04:26:01Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | pavelsukhachev | 161,184,245 | U_kgDOCZt59Q | User | false |
openclaw/openclaw | 4,009,249,154 | I_kwDOQb6kR87u-EmC | 31,281 | https://github.com/openclaw/openclaw/issues/31281 | https://api.github.com/repos/openclaw/openclaw/issues/31281 | Feature: per-channel and per-account blockStreamingBreak configuration | ## Problem
Currently `blockStreamingBreak` (`text_end` / `message_end`) is only configurable at the global level (`agents.defaults.blockStreamingBreak`). Similarly, `blockStreaming` (on/off) is only configurable at the Discord adapter level or per-account, but not per-guild or per-guild-channel.
This creates a proble... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T04:42:32Z | 2026-04-15T04:21:07Z | 2026-04-12T04:26:03Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | MingruiZhang | 5,669,963 | MDQ6VXNlcjU2Njk5NjM= | User | false |
openclaw/openclaw | 4,016,474,360 | I_kwDOQb6kR87vZoj4 | 32,998 | https://github.com/openclaw/openclaw/issues/32998 | https://api.github.com/repos/openclaw/openclaw/issues/32998 | [Bug]: TUI 无法正常响应消息 | ### Bug type
Regression (worked before, now fails)
### Summary
更新到 2026.3.2 后,Feishu WebSocket 一直连接失败 (400错误),导致 Gateway 不稳定,TUI 无法正常响应
### Steps to reproduce
1. 启用 Feishu 插件
2. 运行 openclaw gateway start
3. 查看日志发现 [ws] Request failed with status code 400
4. 运行 openclaw tui 尝试发消息
### Expected behavior
Feishu 正常连接... | closed | not_planned | true | 6 | [
"bug",
"stale",
"regression"
] | [] | 2026-03-03T12:43:07Z | 2026-04-22T04:30:22Z | 2026-04-20T04:26:48Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | sajdijjsid | 150,767,141 | U_kgDOCPyGJQ | User | false |
openclaw/openclaw | 4,009,310,426 | I_kwDOQb6kR87u-Tja | 31,303 | https://github.com/openclaw/openclaw/issues/31303 | https://api.github.com/repos/openclaw/openclaw/issues/31303 | WhatsApp reactions broken — messageId not passed to agent context | ## Bug Description
WhatsApp message reactions fail with `messageId required` error. The agent cannot react to inbound WhatsApp messages because the `messageId` is never available in the tool call params.
## Steps to Reproduce
1. Receive an inbound WhatsApp message
2. Agent attempts to react with an emoji (e.g. 👍) u... | closed | not_planned | true | 4 | [
"stale"
] | [] | 2026-03-02T05:06:37Z | 2026-04-15T04:21:06Z | 2026-04-12T04:26:06Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | VCSMO | 214,151,035 | U_kgDODMOvew | User | false |
openclaw/openclaw | 4,009,283,115 | I_kwDOQb6kR87u-M4r | 31,293 | https://github.com/openclaw/openclaw/issues/31293 | https://api.github.com/repos/openclaw/openclaw/issues/31293 | [Bug]: `message:sent` internal hook never fires for agent replies (all channels) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Internal hooks registered for `message:sent` via HOOK.md never fire for normal agent replies on any channel (Telegram, Discord, Signal, Slack, iMessage). They only fire for `message` tool sends (`action=send`). Each channel's delivery path ... | closed | not_planned | true | 5 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-02T04:56:06Z | 2026-04-15T04:21:06Z | 2026-04-12T04:26:04Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | kingrubic | 116,256,161 | U_kgDOBu3toQ | User | false |
openclaw/openclaw | 4,009,392,075 | I_kwDOQb6kR87u-nfL | 31,322 | https://github.com/openclaw/openclaw/issues/31322 | https://api.github.com/repos/openclaw/openclaw/issues/31322 | Bug: Silent Daily Session Resets and Data Loss After Upgrade to v2026.2.26 | ### 1. Description of the Bug
Upgrading OpenClaw to `v2026.2.26` caused a critical regression where our existing group chat sessions began resetting daily, leading to silent and unexpected loss of conversation history.
This happened even though **our configuration file (`openclaw.json`) did not change**. The issue st... | closed | not_planned | true | 5 | [
"stale"
] | [] | 2026-03-02T05:37:11Z | 2026-04-15T04:21:05Z | 2026-04-12T04:26:08Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | RadonX | 944,393 | MDQ6VXNlcjk0NDM5Mw== | User | false |
openclaw/openclaw | 4,009,446,735 | I_kwDOQb6kR87u-01P | 31,340 | https://github.com/openclaw/openclaw/issues/31340 | https://api.github.com/repos/openclaw/openclaw/issues/31340 | [Feature]: Support OAuth 2.0 flow for Slack to enable cross-workspace installation | ### Summary
Support Slack's OAuth 2.0 flow so that a single OpenClaw instance can be installed into external Slack workspaces via the standard "Add to Slack" button, without requiring manual App creation and token sharing per workspace.
### Problem to solve
The current Slack setup requires creating a Slack App manua... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T05:57:12Z | 2026-04-15T04:21:04Z | 2026-04-12T04:26:11Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | carrotRakko | 24,588,751 | MDQ6VXNlcjI0NTg4NzUx | User | false |
openclaw/openclaw | 4,009,514,514 | I_kwDOQb6kR87u_FYS | 31,354 | https://github.com/openclaw/openclaw/issues/31354 | https://api.github.com/repos/openclaw/openclaw/issues/31354 | Session-event run failures should retry instead of dropping immediate wake | ## Summary
`requestSessionEventRun` currently logs and drops when `triggerSessionEventRun` throws. For transient failures (dispatch/model errors), queued exec system events may remain unprocessed until another unrelated inbound message/event triggers a run.
## Why this matters
Exec completion wake is intended to produ... | closed | not_planned | true | 5 | [
"stale"
] | [] | 2026-03-02T06:19:07Z | 2026-04-15T04:21:03Z | 2026-04-12T04:26:14Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | clay-relay | 257,248,983 | U_kgDOD1VO1w | User | false |
openclaw/openclaw | 4,009,490,388 | I_kwDOQb6kR87u-_fU | 31,348 | https://github.com/openclaw/openclaw/issues/31348 | https://api.github.com/repos/openclaw/openclaw/issues/31348 | Session event runs do not route outbound replies for webchat sessions | ## Summary
Session-event-triggered agent runs (for example, source=exec-event) can execute successfully, but outbound replies may be dropped when the session's delivery context resolves to internal webchat.
## Why this matters
The new exec completion wake path uses requestSessionEventRun to trigger immediate runs afte... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T06:11:41Z | 2026-04-15T04:21:03Z | 2026-04-12T04:26:12Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | clay-relay | 257,248,983 | U_kgDOD1VO1w | User | false |
openclaw/openclaw | 4,009,562,858 | I_kwDOQb6kR87u_RLq | 31,368 | https://github.com/openclaw/openclaw/issues/31368 | https://api.github.com/repos/openclaw/openclaw/issues/31368 | Message tool (file attachments) auto-disabled in Telegram forum topics | ## Summary
When running in Telegram forum topics (supergroup with `is_forum: true`), the embedded agent's `message` tool is automatically disabled via the internal `disableMessageTool` flag. This prevents the agent from sending file attachments (documents, images) as responses in topic conversations.
## Environment
... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T06:31:46Z | 2026-04-15T04:21:02Z | 2026-04-12T04:26:16Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | kilhyeonjun | 41,348,539 | MDQ6VXNlcjQxMzQ4NTM5 | User | false |
openclaw/openclaw | 4,009,573,810 | I_kwDOQb6kR87u_T2y | 31,372 | https://github.com/openclaw/openclaw/issues/31372 | https://api.github.com/repos/openclaw/openclaw/issues/31372 | 飞书Drive工具缺少上传功能 | ## 需求描述
飞书 Drive 工具缺少文件上传功能,无法实现自动备份到飞书云盘。
## 当前支持的功能
- list (列出文件)
- info (查看信息)
- create_folder (创建文件夹)
- move/delete (移动/删除)
## 缺失功能
- upload (上传文件)
## 用户场景
用户想要每天自动备份 OpenClaw 到飞书云盘,需要上传功能支持。
## 建议
在 feishu-drive 工具中添加 upload action,支持文件上传到指定文件夹。 | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T06:34:44Z | 2026-04-15T04:21:01Z | 2026-04-12T04:26:18Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | lawrence1210 | 52,460,968 | MDQ6VXNlcjUyNDYwOTY4 | User | false |
openclaw/openclaw | 4,009,614,781 | I_kwDOQb6kR87u_d29 | 31,382 | https://github.com/openclaw/openclaw/issues/31382 | https://api.github.com/repos/openclaw/openclaw/issues/31382 | gateway update.run fails with not-openclaw-root on npm global installs | ## Environment
- OpenClaw: 2026.2.25 → 2026.3.1
- Install method: `npm install -g openclaw`
- OS: Ubuntu 24.04.4 LTS (x86_64)
- Node: v22.22.0
- Gateway managed by: systemd user service
## Bug 1: `update.run` fails with `not-openclaw-root`
When using the gateway tool `update.run` (or `openclaw update` CLI), it fails ... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T06:47:31Z | 2026-04-15T04:21:00Z | 2026-04-12T04:26:19Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | pnaaberi | 31,892,171 | MDQ6VXNlcjMxODkyMTcx | User | false |
openclaw/openclaw | 4,009,710,072 | I_kwDOQb6kR87u_1H4 | 31,412 | https://github.com/openclaw/openclaw/issues/31412 | https://api.github.com/repos/openclaw/openclaw/issues/31412 | [Bug]: New Job section in the Cron Jobs tab in Control UI behaves strangely | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
The New Job section of the Cron Jobs tab has many UI/UX bugs
### Steps to reproduce
1. Open Control UI
2. Click the Cron Jobs tab
3. Scroll and observe
### Expected behavior
The New Job section should scroll with the section on the left... | closed | not_planned | true | 2 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-02T07:16:56Z | 2026-04-15T04:20:59Z | 2026-04-12T04:26:26Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | QQSHI13 | 127,574,818 | U_kgDOB5qjIg | User | false |
openclaw/openclaw | 4,009,673,328 | I_kwDOQb6kR87u_sJw | 31,397 | https://github.com/openclaw/openclaw/issues/31397 | https://api.github.com/repos/openclaw/openclaw/issues/31397 | Webchat/TUI: propagate thread metadata for ACP session-mode context | ## Summary
Webchat and TUI currently send `chat.send` without thread metadata. In ACP session-mode flows, this can drop parent-thread linkage and lose context.
## Problem
When the active session key encodes a thread (for example `agent:<parent-session-key>:<thread-id>`), Webchat/TUI only forward `sessionKey` to the ga... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T07:05:55Z | 2026-04-15T04:20:59Z | 2026-04-12T04:26:23Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | szcascsa | 161,858,038 | U_kgDOCaXB9g | User | false |
openclaw/openclaw | 4,009,718,901 | I_kwDOQb6kR87u_3R1 | 31,416 | https://github.com/openclaw/openclaw/issues/31416 | https://api.github.com/repos/openclaw/openclaw/issues/31416 | Feature: agents.defaults.subagents.allowAgents for inherited sub-agent access | ## Problem
When running multiple agents with shared utility sub-agents (e.g., a `cron-architect` that handles cron job creation), every agent that wants to spawn it needs `cron-architect` explicitly added to its `subagents.allowAgents` list.
This means:
- Adding a new utility agent requires patching **every existing ... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T07:19:24Z | 2026-04-15T04:20:58Z | 2026-04-12T04:26:28Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | sunshineo | 1,072,766 | MDQ6VXNlcjEwNzI3NjY= | User | false |
openclaw/openclaw | 4,009,741,896 | I_kwDOQb6kR87u_85I | 31,422 | https://github.com/openclaw/openclaw/issues/31422 | https://api.github.com/repos/openclaw/openclaw/issues/31422 | session_status model override clobbered by same-session agent turn save | ## Bug / Feature Request
The `POST /v1/chat/completions` endpoint does not process slash commands or directives (`/model`, `/status`, `/reset`, etc.). Messages like `/model gemini` are passed directly to the agent as chat content instead of being intercepted by the gateway's command/directive layer.
## Root Cause
`s... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T07:26:17Z | 2026-04-15T04:20:57Z | 2026-04-12T04:26:30Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | jmann1111 | 84,252,583 | MDQ6VXNlcjg0MjUyNTgz | User | false |
openclaw/openclaw | 4,009,760,065 | I_kwDOQb6kR87vABVB | 31,427 | https://github.com/openclaw/openclaw/issues/31427 | https://api.github.com/repos/openclaw/openclaw/issues/31427 | Log timestamp timezone inconsistency at Gateway startup | ## Bug Description
Gateway启动时日志时间戳时区不一致。部分模块使用UTC时间,部分使用本地时间(GMT+8)。
## Steps to Reproduce
1. 启动OpenClaw Gateway
2. 观察日志输出,发现时间戳有时是UTC(如06:41:35),有时是本地时间(如14:41:34)
## Expected Behavior
所有日志时间戳应使用统一的时区(本地时区)
## Environment
- OS: Windows_NT (China Standard Time, UTC+8)
- OpenClaw: v2026.2.26
- Node: v24.14.0
## ... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T07:31:25Z | 2026-04-15T04:20:56Z | 2026-04-12T04:26:33Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | haibo15 | 56,247,093 | MDQ6VXNlcjU2MjQ3MDkz | User | false |
openclaw/openclaw | 4,009,755,918 | I_kwDOQb6kR87vAAUO | 31,425 | https://github.com/openclaw/openclaw/issues/31425 | https://api.github.com/repos/openclaw/openclaw/issues/31425 | cron edit --light-context flag accepted but not persisted | ### Bug
`openclaw cron edit <id> --light-context` accepts the flag without error but does not persist `lightContext` to the job payload.
### Steps to reproduce
1. Create a cron job: `openclaw cron add --name test --at "1h" --light-context --message "test"`
2. Verify `lightContext: true` appears in `openclaw cron lis... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T07:30:11Z | 2026-04-15T04:20:56Z | 2026-04-12T04:26:32Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | lnxg33k | 1,170,490 | MDQ6VXNlcjExNzA0OTA= | User | false |
openclaw/openclaw | 4,009,801,767 | I_kwDOQb6kR87vALgn | 31,435 | https://github.com/openclaw/openclaw/issues/31435 | https://api.github.com/repos/openclaw/openclaw/issues/31435 | Pre-compaction memoryFlush: larger default threshold, workspace-level flush instructions, and post-compaction audit | ## Problem
The pre-compaction memoryFlush feature exists and works, but the default configuration leads to context loss in long-running agent sessions.
### Root causes
1. **softThresholdTokens default (4000) is too small.** Not enough headroom for a thorough flush that updates multiple workspace files (e.g. session ... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T07:43:27Z | 2026-04-15T04:20:55Z | 2026-04-12T04:26:35Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | prasants | 5,096,602 | MDQ6VXNlcjUwOTY2MDI= | User | false |
openclaw/openclaw | 4,009,836,575 | I_kwDOQb6kR87vAUAf | 31,440 | https://github.com/openclaw/openclaw/issues/31440 | https://api.github.com/repos/openclaw/openclaw/issues/31440 | [Bug]: Session memory not shared between Telegram and Webchat - separate sessions created for same user | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Description:
When using OpenClaw with Telegram and Webchat channels, the same user (same Telegram ID) creates separate sessions for each channel. The session memory/context is not shared between these sessions.
Expected behavior:
All mess... | closed | not_planned | true | 2 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-02T07:53:50Z | 2026-04-15T04:20:54Z | 2026-04-12T04:26:37Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | assistant-marlon | 264,953,047 | U_kgDOD8rc1w | User | false |
openclaw/openclaw | 4,012,544,287 | I_kwDOQb6kR87vKpEf | 31,999 | https://github.com/openclaw/openclaw/issues/31999 | https://api.github.com/repos/openclaw/openclaw/issues/31999 | Telegram: Use sendMessageDraft API for smooth real-time streaming | ## Feature Request
**Bot API 9.5+ (March 1, 2026)** introduced `sendMessageDraft` which enables real-time streaming text in private chats — similar to ChatGPT's typing animation.
### Current behavior
OpenClaw uses `editMessageText` in a loop for streaming (`streaming: "block"` or `"partial"`). This creates chunky upd... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T17:54:52Z | 2026-04-15T04:20:53Z | 2026-04-12T04:26:41Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | jakereadxoutlier | 206,018,079 | U_kgDODEeWHw | User | false |
openclaw/openclaw | 4,012,526,456 | I_kwDOQb6kR87vKkt4 | 31,994 | https://github.com/openclaw/openclaw/issues/31994 | https://api.github.com/repos/openclaw/openclaw/issues/31994 | Feature Request: Support Moonshot Kimi prompt_cache_key for Context Caching | # Feature Request: Support Moonshot Kimi `prompt_cache_key` for Context Caching
## Summary
Add support for Moonshot Kimi's context caching mechanism via the `prompt_cache_key` parameter.
## Background
Moonshot Kimi API supports context caching through the `prompt_cache_key` parameter (similar to Anthropic's prompt ... | closed | not_planned | true | 6 | [
"stale"
] | [] | 2026-03-02T17:50:59Z | 2026-04-15T04:20:53Z | 2026-04-12T04:26:40Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Karlex | 34,488,018 | MDQ6VXNlcjM0NDg4MDE4 | User | false |
openclaw/openclaw | 4,012,612,266 | I_kwDOQb6kR87vK5qq | 32,021 | https://github.com/openclaw/openclaw/issues/32021 | https://api.github.com/repos/openclaw/openclaw/issues/32021 | [Feature Request] Decouple global browser config, support dynamic browser profile allocation for Cron jobs | ### Pain Points
Currently, OpenClaw's browser Profiles (e.g., `cdpPort`) are statically defined in the global configuration `openclaw.json`. When multiple tasks (such as a high-frequency Cron monitoring task and the user's main conversation) share the same Profile, the following issues occur:
* **Physical Deadlock**:... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T18:11:20Z | 2026-04-15T04:20:52Z | 2026-04-12T04:26:43Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | dsxyy | 8,911,760 | MDQ6VXNlcjg5MTE3NjA= | User | false |
openclaw/openclaw | 4,012,625,706 | I_kwDOQb6kR87vK88q | 32,023 | https://github.com/openclaw/openclaw/issues/32023 | https://api.github.com/repos/openclaw/openclaw/issues/32023 | fix(feishu): bot receives '[Interactive Card]' instead of card content when @mentioned in group | ## Problem
When a bot is @mentioned in a Feishu group containing an old-format interactive card message (e.g. alert/monitoring cards from external systems), the bot receives `[Interactive Card]` instead of the actual card content.
## Background
Feishu has two distinct interactive card formats:
- **Old-format cards*... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T18:14:59Z | 2026-04-15T04:20:51Z | 2026-04-12T04:26:45Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Elarwei001 | 168,552,401 | U_kgDOCgvn0Q | User | false |
openclaw/openclaw | 4,012,628,658 | I_kwDOQb6kR87vK9qy | 32,025 | https://github.com/openclaw/openclaw/issues/32025 | https://api.github.com/repos/openclaw/openclaw/issues/32025 | [Bug]: openclaw update fails with node-llama-cpp cmake installation error (macOS) | ## Summary
`openclaw update` fails during npm package update when `node-llama-cpp` tries to install cmake via xpm, blocking all updates.
## Environment
- **OS**: macOS 15.x (Sequoia) Apple Silicon
- **OpenClaw Version**: 2026.2.26
- **Node.js**: v25.6.0
- **npm**: v11.9.0
- **Installation**: Homebrew global npm (`/opt... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T18:15:49Z | 2026-04-15T04:20:50Z | 2026-04-12T04:26:47Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Jackwill213 | 255,768,407 | U_kgDODz63Vw | User | false |
openclaw/openclaw | 4,012,691,830 | I_kwDOQb6kR87vLNF2 | 32,041 | https://github.com/openclaw/openclaw/issues/32041 | https://api.github.com/repos/openclaw/openclaw/issues/32041 | Telegram: migrate draft streaming to native sendMessageDraft API (Bot API 9.5) | ## Summary
Bot API 9.5 (March 1, 2026) officially **opened `sendMessageDraft` to all bots**. This was previously restricted and caused `TEXTDRAFT_PEER_INVALID` errors (see #7803), which led to the current `sendMessage` + `editMessageText` fallback approach.
Now that the API is universally available, OpenClaw should m... | closed | not_planned | true | 6 | [
"stale"
] | [] | 2026-03-02T18:32:11Z | 2026-04-15T04:20:49Z | 2026-04-12T04:26:51Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | wangshipei | 54,164,572 | MDQ6VXNlcjU0MTY0NTcy | User | false |
openclaw/openclaw | 4,012,777,218 | I_kwDOQb6kR87vLh8C | 32,057 | https://github.com/openclaw/openclaw/issues/32057 | https://api.github.com/repos/openclaw/openclaw/issues/32057 | TUI: Pre-compaction memory flush system messages visible to user | ## Bug
When using `openclaw tui`, internal system messages like pre-compaction memory flush prompts are rendered as regular chat messages visible to the user.
**Expected:** System messages (compaction flush, heartbeat prompts, etc.) should be hidden from the TUI display.
**Actual:** Messages like this appear in the ... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T18:55:02Z | 2026-04-15T04:20:48Z | 2026-04-12T04:26:54Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Hovhas | 25,573,180 | MDQ6VXNlcjI1NTczMTgw | User | false |
openclaw/openclaw | 4,012,746,148 | I_kwDOQb6kR87vLaWk | 32,048 | https://github.com/openclaw/openclaw/issues/32048 | https://api.github.com/repos/openclaw/openclaw/issues/32048 | Telegram long-polling silently dies on Android/Termux when multi-agent gateway is under load | ## Environment
- **OpenClaw version:** 2026.2.26 (bc50708)
- **OS:** Android 15 (arm64) via Termux
- **Device:** Samsung Galaxy Z Fold 7
- **Node:** v24.13.0
- **Model:** anthropic/claude-opus-4-6
- **Setup:** Single gateway, 2 agents (main + secondary), Telegram + WhatsApp channels
## Problem
Telegram long-polling fo... | closed | not_planned | true | 4 | [
"stale"
] | [] | 2026-03-02T18:46:29Z | 2026-04-15T04:20:48Z | 2026-04-12T04:26:52Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | eros-plusval | 259,764,026 | U_kgDOD3uvOg | User | false |
openclaw/openclaw | 4,012,793,686 | I_kwDOQb6kR87vLl9W | 32,060 | https://github.com/openclaw/openclaw/issues/32060 | https://api.github.com/repos/openclaw/openclaw/issues/32060 | Agent loses context of owner's outbound messages in DM conversations | ## Re-filing of #12480 (closed by stale bot)
**Version:** 2026.3.1 (confirmed still present)
### Problem
In WhatsApp/Telegram DM conversations, the agent's context history does not include the owner's outbound messages. The agent has no visibility into what the owner has already said in the conversation, leading to r... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T18:59:20Z | 2026-04-15T04:20:47Z | 2026-04-12T04:26:56Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | mactsk | 5,073,360 | MDQ6VXNlcjUwNzMzNjA= | User | false |
openclaw/openclaw | 4,012,797,211 | I_kwDOQb6kR87vLm0b | 32,062 | https://github.com/openclaw/openclaw/issues/32062 | https://api.github.com/repos/openclaw/openclaw/issues/32062 | [Bug]: Exec approval lifecycle outcomes can be delayed by collect/steer queueing; Discord workaround requires operator nudge | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
In a Discord workflow using approval-gated `exec`, terminal approval lifecycle outcomes (`allow`, `deny`, `timeout`) can be delayed or effectively swallowed by queue behavior under collect/steer conditions.
This creates operator uncertaint... | closed | not_planned | true | 2 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-02T19:00:13Z | 2026-04-15T04:20:46Z | 2026-04-12T04:26:57Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | ceo-nada | 12,769,724 | MDQ6VXNlcjEyNzY5NzI0 | User | false |
openclaw/openclaw | 4,012,799,244 | I_kwDOQb6kR87vLnUM | 32,063 | https://github.com/openclaw/openclaw/issues/32063 | https://api.github.com/repos/openclaw/openclaw/issues/32063 | Message Reliability: Durable SQLite Outbox, Recovery Worker, and Unified sendPayload | ## Background
OpenClaw currently implements outbound delivery reliability as a collection of subsystem-specific mechanisms: a file-based delivery queue (`delivery-queue/*.json`), channel-specific retry rules, startup-only orphan replay, and local permanent-error classification. Issue and PR history shows repeated fail... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T19:00:44Z | 2026-04-15T04:20:45Z | 2026-04-12T04:26:59Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | nohat | 838,027 | MDQ6VXNlcjgzODAyNw== | User | false |
openclaw/openclaw | 4,012,930,845 | I_kwDOQb6kR87vMHcd | 32,095 | https://github.com/openclaw/openclaw/issues/32095 | https://api.github.com/repos/openclaw/openclaw/issues/32095 | SQLITE_CANTOPEN crash when plugins.slots.memory is set to openclaw-mem0 (v2026.3.1) | ## Description
Gateway enters a crash loop with `SQLITE_CANTOPEN` when `plugins.slots.memory` is set to `"openclaw-mem0"`. The gateway boots fine and channels connect, but it crashes during the `before_agent_start` hook when processing any inbound message.
The `jiti` module tries to load `node_sqlite3.node` from an i... | closed | not_planned | true | 5 | [
"stale"
] | [] | 2026-03-02T19:34:39Z | 2026-04-15T04:20:44Z | 2026-04-12T04:27:03Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | diogo85 | 21,048,221 | MDQ6VXNlcjIxMDQ4MjIx | User | false |
openclaw/openclaw | 4,012,894,973 | I_kwDOQb6kR87vL-r9 | 32,088 | https://github.com/openclaw/openclaw/issues/32088 | https://api.github.com/repos/openclaw/openclaw/issues/32088 | fix(slack): All Slack channels crash on startup with Bolt 4.6.0 — message.channels is not a valid event type | ## Bug
All Slack channels crash immediately on startup with:
```
Although the document mentions "message.channels", it is not a valid event type.
Use "message" instead. If you want to filter message events, you can use event.channel_type for it.
```
## Root cause
`src/slack/monitor/events/messages.ts` registers `me... | closed | not_planned | true | 7 | [
"stale"
] | [] | 2026-03-02T19:25:42Z | 2026-04-15T04:20:44Z | 2026-04-12T04:27:01Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | cathrynlavery | 50,469,282 | MDQ6VXNlcjUwNDY5Mjgy | User | false |
openclaw/openclaw | 4,012,951,405 | I_kwDOQb6kR87vMMdt | 32,101 | https://github.com/openclaw/openclaw/issues/32101 | https://api.github.com/repos/openclaw/openclaw/issues/32101 | gateway restart: tool result lost due to WebSocket disconnect race condition | ## Summary
When running `openclaw gateway restart` via the agent's `exec` tool, the tool result is never returned. The session receives a synthetic error: `missing tool result in session history; inserted synthetic error result for transcript repair.`
## Root Cause
The `gateway restart` command restarts the gateway ... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T19:39:56Z | 2026-04-15T04:20:43Z | 2026-04-12T04:27:05Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Matesko17 | 86,290,717 | MDQ6VXNlcjg2MjkwNzE3 | User | false |
openclaw/openclaw | 4,012,971,180 | I_kwDOQb6kR87vMRSs | 32,106 | https://github.com/openclaw/openclaw/issues/32106 | https://api.github.com/repos/openclaw/openclaw/issues/32106 | v2026.3.1 regression: aggressive compaction loop with memoryFlush | ## Description
After upgrading from v2026.2.26 to v2026.3.1, all agents experience a compaction loop — compacting every 2-3 minutes regardless of conversation length.
## Steps to Reproduce
1. Upgrade from v2026.2.26 to v2026.3.1
2. Start a conversation with any agent
3. Observe compaction triggering repeatedly after... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T19:44:28Z | 2026-04-15T04:20:42Z | 2026-04-12T04:27:06Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | stephenschoettler | 7,587,303 | MDQ6VXNlcjc1ODczMDM= | User | false |
openclaw/openclaw | 4,013,009,684 | I_kwDOQb6kR87vMasU | 32,113 | https://github.com/openclaw/openclaw/issues/32113 | https://api.github.com/repos/openclaw/openclaw/issues/32113 | [Bug]:Installer Bug: OpenClaw CLI (bin/openclaw.js) Missing from npm Package and GitHub Repo | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Hi OpenClaw Team,
I’m unable to install or update OpenClaw using any standard method (npm global, local, npx, or even a GitHub clone into my Docker environment). The root issue: the core CLI binary (bin/openclaw.js) is missing from both the... | closed | not_planned | true | 7 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-02T19:52:16Z | 2026-04-15T04:20:41Z | 2026-04-12T04:27:10Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | trinmabrker | 264,659,249 | U_kgDOD8ZhMQ | User | false |
openclaw/openclaw | 4,012,994,819 | I_kwDOQb6kR87vMXED | 32,109 | https://github.com/openclaw/openclaw/issues/32109 | https://api.github.com/repos/openclaw/openclaw/issues/32109 | feat: default session reset for group/topic sessions should be idle, not daily | ## Problem
Group chat and forum topic sessions default to `daily` reset mode (reset at 4 AM), which causes conversation context loss every day. This is unintuitive for topic-based workflows where a topic represents an ongoing project/discussion.
### Current Behavior
1. Default `session.reset.mode` = `"daily"` with `... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T19:49:25Z | 2026-04-15T04:20:41Z | 2026-04-12T04:27:08Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | kzai2037 | 260,737,714 | U_kgDOD4qKsg | User | false |
openclaw/openclaw | 4,013,016,387 | I_kwDOQb6kR87vMcVD | 32,115 | https://github.com/openclaw/openclaw/issues/32115 | https://api.github.com/repos/openclaw/openclaw/issues/32115 | Straico provider broken: 422 Missing model or smart llm selector | ## Bug Description
The Straico provider integration fails with HTTP 422 error when attempting to use any Straico model (e.g., ). The error message indicates a mismatch between the API request format and what Straico expects.
## Error Message
## Root Cause Analysis
After reviewing Straico's official API documentati... | closed | not_planned | true | 4 | [
"stale"
] | [] | 2026-03-02T19:53:38Z | 2026-04-15T04:20:40Z | 2026-04-12T04:27:12Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Valkster70 | 129,894,487 | U_kgDOB74IVw | User | false |
openclaw/openclaw | 4,013,048,367 | I_kwDOQb6kR87vMkIv | 32,123 | https://github.com/openclaw/openclaw/issues/32123 | https://api.github.com/repos/openclaw/openclaw/issues/32123 | Feishu file upload corrupts binary files (xlsx/pdf) when sent via message tool filePath | ## Bug Description
Files sent through the `message` tool with `filePath` parameter are corrupted when delivered via Feishu channel. Recipients cannot open xlsx/pdf files in the Feishu client.
## Environment
- OpenClaw version: 2026.3.1
- Channel: Feishu (飞书)
- OS: macOS (Darwin 25.4.0 arm64)
- Feishu SDK: @larksuite... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T20:00:17Z | 2026-04-15T04:20:39Z | 2026-04-12T04:27:14Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | russellju16-afk | 245,091,598 | U_kgDODpvNDg | User | false |
openclaw/openclaw | 4,013,179,711 | I_kwDOQb6kR87vNEM_ | 32,149 | https://github.com/openclaw/openclaw/issues/32149 | https://api.github.com/repos/openclaw/openclaw/issues/32149 | [Bug]: weird behavior in chat window, seems to overwrite its own output. Report generated by openclaw | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Webchat UI failed to render a complete response. Part of the assistant's message disappeared during initial rendering but appeared correctly when the same content was re-sent in a follow-up message.
### Steps to reproduce
# Bug Report: W... | closed | not_planned | true | 2 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-02T20:31:18Z | 2026-04-15T04:20:38Z | 2026-04-12T04:27:17Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | johreg | 5,519,467 | MDQ6VXNlcjU1MTk0Njc= | User | false |
openclaw/openclaw | 4,013,053,297 | I_kwDOQb6kR87vMlVx | 32,124 | https://github.com/openclaw/openclaw/issues/32124 | https://api.github.com/repos/openclaw/openclaw/issues/32124 | feat: Add subagent completion delivery mode config (internal vs user) | ## Problem
Sub-agent completion messages ("✅ Subagent main finished...") are always forwarded to the user's chat channel (Telegram/Discord/etc). This has caused 10+ unwanted message leaks over 2 weeks.
## Root Cause
In `src/agents/subagent-announce-dispatch.ts`:
- `expectsCompletionMessage` defaults to `true`
- When t... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T20:01:15Z | 2026-04-15T04:20:38Z | 2026-04-12T04:27:15Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | alecsadl3r-eng | 261,381,127 | U_kgDOD5RcBw | User | false |
openclaw/openclaw | 4,013,299,348 | I_kwDOQb6kR87vNhaU | 32,168 | https://github.com/openclaw/openclaw/issues/32168 | https://api.github.com/repos/openclaw/openclaw/issues/32168 | Streaming partial leak: 'NO' visible before NO_REPLY suppression kicks in (v2026.3.1) | ## Bug
**Version:** 2026.3.1
**Component:** Reply pipeline / streaming partial suppression
## Description
When an agent reply starts with `NO_REPLY` (the silent reply token), the partial text `NO` leaks to webchat as a visible streaming delta before the full suppression logic catches it.
## Root Cause
`handleParti... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T21:04:55Z | 2026-04-15T04:20:37Z | 2026-04-12T04:27:19Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | fuller-stack-dev | 263,060,202 | U_kgDOD6366g | User | false |
openclaw/openclaw | 4,013,307,073 | I_kwDOQb6kR87vNjTB | 32,169 | https://github.com/openclaw/openclaw/issues/32169 | https://api.github.com/repos/openclaw/openclaw/issues/32169 | clawhub publish fails with 'SKILL.md required' when .git directory is present | ## Bug
`clawhub publish .` fails with `Error: SKILL.md required` when the skill folder contains a `.git` directory, even though `SKILL.md` exists and is valid (correct frontmatter, UTF-8).
## Steps to Reproduce
1. Create a skill with a valid `SKILL.md` (YAML frontmatter + content)
2. Initialize git in the skill fold... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T21:06:55Z | 2026-04-15T04:20:36Z | 2026-04-12T04:27:20Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | JonathanJing | 17,068,507 | MDQ6VXNlcjE3MDY4NTA3 | User | false |
openclaw/openclaw | 4,013,386,014 | I_kwDOQb6kR87vN2ke | 32,189 | https://github.com/openclaw/openclaw/issues/32189 | https://api.github.com/repos/openclaw/openclaw/issues/32189 | [Bug]: Discord agent reports wrong model — says Haiku but running Minimax | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
1. Switched Discord agent to Minimax — edited config to set model: minimax-portal/MiniMax-M2.5
2. Fixed broken OAuth creds — injected your direct Minimax API key into both gateway and discord-ops agent configs
3. Fixed config errors — remov... | closed | not_planned | true | 3 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-02T21:28:01Z | 2026-04-15T04:20:35Z | 2026-04-12T04:27:24Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | 0xNull-ops | 6,685,040 | MDQ6VXNlcjY2ODUwNDA= | User | false |
openclaw/openclaw | 4,013,373,520 | I_kwDOQb6kR87vNzhQ | 32,184 | https://github.com/openclaw/openclaw/issues/32184 | https://api.github.com/repos/openclaw/openclaw/issues/32184 | [Bug] Security Scanner: False Positive for 'OpenClaw port exposed on public interface' | **Skill:** openclaw-security-scanner v1.0.0
## Summary
The scanner incorrectly reports `CRITICAL: OpenClaw port 18789 exposed on public interface` when the port is actually bound to localhost (127.0.0.1).
## Root Cause
In `oc-security-scan.sh`, the `lsof` command outputs a header line:
\`\`\`
COMMAND PID USER FD T... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T21:25:01Z | 2026-04-15T04:20:35Z | 2026-04-12T04:27:22Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | drpeterkalmar | 243,632,958 | U_kgDODoWLPg | User | false |
openclaw/openclaw | 4,013,437,383 | I_kwDOQb6kR87vODHH | 32,195 | https://github.com/openclaw/openclaw/issues/32195 | https://api.github.com/repos/openclaw/openclaw/issues/32195 | Bug: Slack replies dropped for custom OpenAI-compatible provider despite successful completion | # Bug: Slack replies dropped for custom OpenAI-compatible provider despite successful completion
## Summary
When using a custom OpenAI-compatible provider chain (`OpenClaw -> LiteLLM -> local adapter -> 1min.ai`), model execution completes successfully, provider responses are valid/non-empty, and local curl requests r... | closed | not_planned | true | 3 | [
"stale"
] | [] | 2026-03-02T21:41:55Z | 2026-04-15T04:20:34Z | 2026-04-12T04:27:25Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Mwhite84 | 5,177,066 | MDQ6VXNlcjUxNzcwNjY= | User | false |
openclaw/openclaw | 4,013,476,951 | I_kwDOQb6kR87vOMxX | 32,204 | https://github.com/openclaw/openclaw/issues/32204 | https://api.github.com/repos/openclaw/openclaw/issues/32204 | [Bug]: Docker build fails — TypeScript errors in agent-components.ts and pairing-store.ts after commit 3a08e69 | ### Summary
Docker build (`pnpm build`) fails during the `build:plugin-sdk:dts` step with TypeScript errors in `src/discord/monitor/agent-components.ts` and `src/pairing/pairing-store.ts` introduced in commit `3a08e69` (_refactor: unify queueing and normalize telegram slack flows_).
### Steps to reproduce
1. Deploy ... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-02T21:52:21Z | 2026-04-15T04:20:33Z | 2026-04-12T04:27:28Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | shawnesquivel | 94,336,773 | U_kgDOBZ93BQ | User | false |
openclaw/openclaw | 4,183,133,020 | I_kwDOQb6kR875VYtc | 58,617 | https://github.com/openclaw/openclaw/issues/58617 | https://api.github.com/repos/openclaw/openclaw/issues/58617 | [Bug]: msteams channel file attachments not downloaded — Graph fallback triggers on non-file HTML attachments | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When users attach files (PDFs, images, etc.) to Teams channel messages, the bot cannot see or access the file content. The LLM only receives the text portion of the message.
Additionally, the Graph API media fallback path fires excessively... | closed | completed | true | 0 | [] | [] | 2026-04-01T00:08:25Z | 2026-04-15T04:20:31Z | 2026-04-12T04:33:09Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | atharva-getboon | 221,907,386 | U_kgDODToJug | User | false |
openclaw/openclaw | 4,113,025,601 | I_kwDOQb6kR871J8pB | 51,749 | https://github.com/openclaw/openclaw/issues/51749 | https://api.github.com/repos/openclaw/openclaw/issues/51749 | MS Teams: graph media fetch empty for channel file attachments despite valid permissions and token | ## Summary
MS Teams channel file attachments (PDFs, documents) are never delivered to the agent. The plugin logs `graph media fetch empty` on every inbound file, despite all Graph API permissions being correctly configured and working when tested manually.
## Environment
- OpenClaw: 2026.3.13 (Docker)
- Node: 24.14.... | closed | completed | true | 3 | [] | [] | 2026-03-21T16:47:17Z | 2026-04-15T04:20:31Z | 2026-04-12T04:33:09Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | staroscott | 230,729,624 | U_kgDODcCnmA | User | false |
openclaw/openclaw | 4,204,828,109 | I_kwDOQb6kR876oJXN | 60,859 | https://github.com/openclaw/openclaw/issues/60859 | https://api.github.com/repos/openclaw/openclaw/issues/60859 | web_fetch fails with "resolves to private/internal/special-use IP address" in proxy environments using fake-ip mode | # Issue Draft: web_fetch blocked by SSRF protection in proxy/fake-ip environments
---
**Title:** `web_fetch` fails with "resolves to private/internal/special-use IP address" in proxy environments using fake-ip mode
## Description
`web_fetch` fails for all URLs when running behind a transparent proxy (Surge, Clash, ... | closed | completed | true | 1 | [] | [] | 2026-04-04T12:36:33Z | 2026-04-15T04:20:29Z | 2026-04-12T06:23:23Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Sheetaa | 4,032,009 | MDQ6VXNlcjQwMzIwMDk= | User | false |
openclaw/openclaw | 4,019,288,271 | I_kwDOQb6kR87vkXjP | 33,639 | https://github.com/openclaw/openclaw/issues/33639 | https://api.github.com/repos/openclaw/openclaw/issues/33639 | WebUI shows wizard.lastRunVersion instead of actual package version | ## Bug Description
The WebUI (Control UI) displays the version from `wizard.lastRunVersion` in the config instead of the actual installed package version.
## Steps to Reproduce
1. Install OpenClaw (e.g., `2026.2.26`)
2. Run `openclaw configure` (wizard)
3. Update OpenClaw to a newer version (e.g., `2026.3.2`) via `n... | closed | completed | true | 3 | [
"stale"
] | [] | 2026-03-03T23:55:58Z | 2026-04-15T04:20:29Z | 2026-04-12T05:07:22Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | nixprosoft | 5,313,478 | MDQ6VXNlcjUzMTM0Nzg= | User | false |
openclaw/openclaw | 4,170,050,270 | I_kwDOQb6kR874jere | 57,672 | https://github.com/openclaw/openclaw/issues/57672 | https://api.github.com/repos/openclaw/openclaw/issues/57672 | [Bug]: Mistral tool call IDs fail validation when routed via OpenRouter | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
OpenClaw version: **2026.3.28**
When using a Mistral model via OpenRouter, any request that involves tool/function calling fails with a **400 error** from the Mistral API. The error indicates that OpenClaw is generating tool ca... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-03-30T11:44:08Z | 2026-04-15T04:20:25Z | 2026-04-12T09:19:32Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Lkrambar | 158,211,079 | U_kgDOCW4cBw | User | false |
openclaw/openclaw | 4,183,906,222 | I_kwDOQb6kR875YVeu | 58,727 | https://github.com/openclaw/openclaw/issues/58727 | https://api.github.com/repos/openclaw/openclaw/issues/58727 | [Bug] Busy queue / announce queue can leak internal prompts into user-visible replies | ## Summary
When OpenClaw is busy, queued follow-up / announce flows can reuse internal prompt text as if it were safe user-visible content. This creates a boundary leak where internal draft text, queue scaffolding, or intermediate system text can later be surfaced back to users.
This is not just a presentation problem... | closed | completed | true | 2 | [] | [] | 2026-04-01T04:05:43Z | 2026-04-15T04:20:19Z | 2026-04-12T11:37:53Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | KimGLee | 150,593,189 | U_kgDOCPnepQ | User | false |
openclaw/openclaw | 4,011,235,906 | I_kwDOQb6kR87vFppC | 31,708 | https://github.com/openclaw/openclaw/issues/31708 | https://api.github.com/repos/openclaw/openclaw/issues/31708 | [Bug]: 5x API costs due to ineffective prompt caching (Re-file of #23715) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
#23715 was misidentified as a duplicate and closed in error without resolution, then locked for discussion despite actively discussed and worke on by several contributors. I'm opening this issue to be able to continue discussion.
Mentions:... | closed | completed | true | 17 | [
"bug",
"bug:behavior"
] | [] | 2026-03-02T13:18:35Z | 2026-04-30T04:51:03Z | 2026-04-25T05:38:04Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | svenssonaxel | 163,858 | MDQ6VXNlcjE2Mzg1OA== | User | false |
openclaw/openclaw | 4,009,857,214 | I_kwDOQb6kR87vAZC- | 31,449 | https://github.com/openclaw/openclaw/issues/31449 | https://api.github.com/repos/openclaw/openclaw/issues/31449 | Enhancement: HTTP API should return full execution details (reasoning, tool I/O, context) - parity with web UI | ## Issue Description
The current HTTP API endpoints (`/v1/chat/completions` and `/v1/responses`) do not return the same level of detail that's visible in the web UI.
**What's missing in HTTP API:**
- Reasoning/thought process (when enabled)
- Tool call inputs and outputs
- Context window usage
- Full execution metada... | closed | not_planned | true | 5 | [
"stale"
] | [] | 2026-03-02T07:59:12Z | 2026-04-22T04:30:23Z | 2026-04-20T04:26:36Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | xiabo0816 | 11,622,590 | MDQ6VXNlcjExNjIyNTkw | User | false |
openclaw/openclaw | 4,009,713,927 | I_kwDOQb6kR87u_2EH | 31,414 | https://github.com/openclaw/openclaw/issues/31414 | https://api.github.com/repos/openclaw/openclaw/issues/31414 | npm update doesn't update systemd service version, causing version mismatch | ### Problem
When updating openclaw via `npm update -g openclaw`, the systemd service configuration file is not updated, causing the gateway to report an outdated version.
### Steps to Reproduce
1. Install openclaw and set up the gateway service via `openclaw daemon install`
2. Update openclaw: `npm update -g openclaw`... | closed | not_planned | true | 4 | [
"stale"
] | [] | 2026-03-02T07:17:58Z | 2026-04-22T04:30:23Z | 2026-04-20T04:26:35Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | LAVOZDELINFIERNO | 57,243,579 | MDQ6VXNlcjU3MjQzNTc5 | User | false |
openclaw/openclaw | 4,009,144,332 | I_kwDOQb6kR87u9rAM | 31,257 | https://github.com/openclaw/openclaw/issues/31257 | https://api.github.com/repos/openclaw/openclaw/issues/31257 | Support Anthropic 1M context window for Amazon Bedrock provider | ## Feature Request
### Problem
The `context1m: true` option currently only works with the direct Anthropic API provider (`provider === "anthropic"`). When using Amazon Bedrock via the Converse API, the 1M context window beta (`context-1m-2025-08-07`) requires injecting `additionalModelRequestFields` with `anthropic_be... | closed | not_planned | true | 5 | [
"stale"
] | [] | 2026-03-02T04:03:03Z | 2026-04-22T04:30:24Z | 2026-04-20T04:26:29Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | jesamkim | 28,912,784 | MDQ6VXNlcjI4OTEyNzg0 | User | false |
openclaw/openclaw | 4,216,683,461 | I_kwDOQb6kR877VXvF | 62,409 | https://github.com/openclaw/openclaw/issues/62409 | https://api.github.com/repos/openclaw/openclaw/issues/62409 | Bug: commentary-phase assistant text leaks to Telegram as user-visible messages | ## Summary
On OpenClaw 2026.4.5, assistant messages tagged internally as `phase: "commentary"` are leaking to Telegram as user-visible messages.
This should never happen. Commentary/tool-chatter should remain internal, while only the final user-facing reply should be delivered.
## Impact
High trust / privacy issue.... | closed | completed | true | 2 | [] | [] | 2026-04-07T09:16:48Z | 2026-04-15T04:20:01Z | 2026-04-12T16:46:02Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | danielt69 | 3,322,065 | MDQ6VXNlcjMzMjIwNjU= | User | false |
openclaw/openclaw | 4,212,194,777 | I_kwDOQb6kR877EP3Z | 61,936 | https://github.com/openclaw/openclaw/issues/61936 | https://api.github.com/repos/openclaw/openclaw/issues/61936 | [Feature]: Allow dreaming with external memory slot plugins | ### Summary
Allow `memory-core` dreaming to run even when `plugins.slots.memory` points at an external memory plugin such as `memory-lancedb-pro`.
### Problem to solve
Right now dreaming is tied to `memory-core`, but `memory-core` is effectively disabled whenever another `kind: "memory"` plugin owns the memory slot.... | closed | completed | true | 3 | [] | [] | 2026-04-06T14:30:40Z | 2026-04-15T04:19:53Z | 2026-04-12T16:47:08Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | VACInc | 3,279,061 | MDQ6VXNlcjMyNzkwNjE= | User | false |
openclaw/openclaw | 4,215,404,623 | I_kwDOQb6kR877QfhP | 62,296 | https://github.com/openclaw/openclaw/issues/62296 | https://api.github.com/repos/openclaw/openclaw/issues/62296 | [Bug] memory-core dreaming cron system-event does not trigger before_agent_reply hook (deliveryStatus: not-requested) | ## Bug Description
The managed dreaming cron job created by memory-core at startup does not actually trigger the dreaming promotion logic. The cron runs but completes in ~4ms with `deliveryStatus: "not-requested"` — it appears to do nothing.
## Steps to Reproduce
1. Install openclaw v2026.4.5
2. Enable dreaming: `op... | closed | completed | true | 3 | [] | [] | 2026-04-07T04:27:58Z | 2026-04-15T04:19:46Z | 2026-04-12T16:53:45Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | zuoxiaojun | 24,306,381 | MDQ6VXNlcjI0MzA2Mzgx | User | false |
openclaw/openclaw | 4,142,661,220 | I_kwDOQb6kR8726_5k | 55,096 | https://github.com/openclaw/openclaw/issues/55096 | https://api.github.com/repos/openclaw/openclaw/issues/55096 | Sub-agents cannot send channel messages via cron jobs — only main agent's isolated session has qqbot/telegram/etc. message tool permission | ### Problem Description
When creating a cron job with `payload.kind = "agentTurn"` for a sub-agent (e.g. `study`, `finance`), the OpenClaw Gateway enforces that `sessionTarget` must be `"isolated"`. However, isolated sessions for sub-agents lack permission to send messages via channel plugins (qqbot, telegram, discord... | closed | completed | true | 0 | [] | [] | 2026-03-26T10:42:50Z | 2026-04-15T04:19:42Z | 2026-04-12T17:03:04Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | CobyLee66 | 16,515,534 | MDQ6VXNlcjE2NTE1NTM0 | User | false |
openclaw/openclaw | 4,207,974,612 | I_kwDOQb6kR8760JjU | 61,412 | https://github.com/openclaw/openclaw/issues/61412 | https://api.github.com/repos/openclaw/openclaw/issues/61412 | [Bug]: MiniMax text provider loads, but TTS shows `minimax (not configured)` / `no_provider_registered` on 2026.4.2 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
## Summary
On OpenClaw `2026.4.2`, MiniMax appears to be configured and loaded correctly for text/image models, but MiniMax TTS is still unavailable in `/tts status`.
`/tts status` shows:
- `Provider: minima... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-04-05T16:04:35Z | 2026-04-15T04:19:41Z | 2026-04-12T17:07:11Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | Nana77mi | 79,355,270 | MDQ6VXNlcjc5MzU1Mjcw | User | false |
openclaw/openclaw | 4,216,442,346 | I_kwDOQb6kR877Uc3q | 62,396 | https://github.com/openclaw/openclaw/issues/62396 | https://api.github.com/repos/openclaw/openclaw/issues/62396 | memory-core config validation always fails — even with valid schema-conformant values | ## memory-core config validation always fails — even with valid schema-conformant values
**OpenClaw version:** 2026.4.5
### Problem
Configuring `plugins.entries.memory-core.config.dreaming` with **any** value causes config validation to fail, even values that strictly conform to the documented and schema-declared pr... | closed | completed | true | 1 | [] | [] | 2026-04-07T08:28:01Z | 2026-04-15T04:19:39Z | 2026-04-12T17:31:06Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | lyubia | 85,536,536 | MDQ6VXNlcjg1NTM2NTM2 | User | false |
openclaw/openclaw | 4,104,141,410 | I_kwDOQb6kR870oDpi | 50,628 | https://github.com/openclaw/openclaw/issues/50628 | https://api.github.com/repos/openclaw/openclaw/issues/50628 | [Bug]: Browser control server installs no authentication when gateway auth mode is trusted-proxy | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 9.0 / 10.0 | 9.3 / 10.0 |
| **Severity** | Critical | Critical |
| **Vector** | CVSS:3.1/AV:L/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:N | CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:N/SC:H/SI:H/SA:N |
| **Calculator** |... | closed | completed | true | 2 | [] | [] | 2026-03-19T20:46:52Z | 2026-04-15T04:19:28Z | 2026-04-12T19:54:26Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 4,104,178,287 | I_kwDOQb6kR870oMpv | 50,644 | https://github.com/openclaw/openclaw/issues/50644 | https://api.github.com/repos/openclaw/openclaw/issues/50644 | [Bug]: gateway.auth.mode=none propagates silently to browser control server, removing all authentication from browser automation APIs | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 9.2 / 10.0 | 9.3 / 10.0 |
| **Severity** | Critical | Critical |
| **Vector** | CVSS:3.1/AV:L/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:L | CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:L/SC:H/SI:H/SA:L |
| **Calculator** |... | closed | completed | true | 4 | [] | [] | 2026-03-19T20:56:08Z | 2026-04-15T04:19:27Z | 2026-04-12T19:54:26Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 3,998,171,451 | I_kwDOQb6kR87uT0E7 | 27,996 | https://github.com/openclaw/openclaw/issues/27996 | https://api.github.com/repos/openclaw/openclaw/issues/27996 | Cron scheduler: nextRunAtMs not recomputed after editing cron expr/timezone | ### Summary
Cron jobs do not reliably recompute `state.nextRunAtMs` after editing the cron expression and/or timezone (even after gateway restart). This can cause jobs to never run at the intended times.
### Environment
- OpenClaw CLI: `2026.2.17`
- Cron store: `/home/robo/.openclaw/cron/jobs.json`
- Gateway running o... | closed | completed | true | 11 | [] | [] | 2026-02-26T22:56:28Z | 2026-04-30T04:36:17Z | 2026-04-27T06:48:04Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | fagnersouza666 | 10,549,588 | MDQ6VXNlcjEwNTQ5NTg4 | User | false |
openclaw/openclaw | 4,140,587,479 | I_kwDOQb6kR872zFnX | 54,903 | https://github.com/openclaw/openclaw/issues/54903 | https://api.github.com/repos/openclaw/openclaw/issues/54903 | [Feature]: Federated credential support for Microsoft Teams (certificate + managed identity) | ### Summary
Add federated authentication support for the Microsoft Teams channel plugin, enabling certificate-based and Azure Managed Identity auth as alternatives to client secrets.
### Problem to solve
The MS Teams plugin currently only supports client secret (`appPassword`) authentication. This is insufficient fo... | closed | completed | true | 2 | [] | [] | 2026-03-26T04:52:47Z | 2026-04-15T04:19:18Z | 2026-04-13T03:16:34Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | HDYA | 14,888,949 | MDQ6VXNlcjE0ODg4OTQ5 | User | false |
openclaw/openclaw | 3,974,389,121 | I_kwDOQb6kR87s5F2B | 23,550 | https://github.com/openclaw/openclaw/issues/23550 | https://api.github.com/repos/openclaw/openclaw/issues/23550 | gateway status falsely detects unrelated user services as 'gateway-like' and suggests destructive cleanup | ## Bug
`openclaw gateway status` detects **any** user-scoped systemd service as "Other gateway-like services" and outputs dangerous cleanup hints that suggest **disabling/removing the actual gateway**.
## Reproduction
1. Create any custom user systemd service (not a gateway), e.g. a voice pipeline on port 8084
2. Ru... | closed | not_planned | true | 8 | [
"stale"
] | [] | 2026-02-22T12:25:30Z | 2026-04-22T04:30:25Z | 2026-04-20T04:26:02Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | balthazar-bot | 258,649,781 | U_kgDOD2qutQ | User | false |
openclaw/openclaw | 3,997,317,462 | I_kwDOQb6kR87uQjlW | 27,858 | https://github.com/openclaw/openclaw/issues/27858 | https://api.github.com/repos/openclaw/openclaw/issues/27858 | sessions.patch model override is racy — subagent starts on default model before patch lands | ## Description
Subagent model override via `sessions.patch` has a race condition. The spawn flow creates the session, then patches the model in a separate call — but the session's first LLM call fires before the patch completes, so the subagent always runs on the default primary model.
This is a follow-up to #19460 (... | closed | not_planned | true | 7 | [
"stale"
] | [] | 2026-02-26T19:18:46Z | 2026-04-15T04:19:07Z | 2026-04-12T04:21:46Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | vishal-hirenami | 98,901,082 | U_kgDOBeUcWg | User | false |
openclaw/openclaw | 4,014,530,854 | I_kwDOQb6kR87vSOEm | 32,499 | https://github.com/openclaw/openclaw/issues/32499 | https://api.github.com/repos/openclaw/openclaw/issues/32499 | [Feature]: Add all slash-commands to Slack JSON manifest (or offer in docs) | ### Summary
[The Slack documentation says](https://docs.openclaw.ai/channels/slack), "When native commands are enabled, register matching slash commands in Slack (/<command> names)…." Who has time to add _all_ of them?
### Problem to solve
Manually adding all the slash-commends is tedious and wasting time. The docs ... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-03-03T03:51:26Z | 2026-04-15T04:19:06Z | 2026-04-12T07:12:32Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | theDanielJLewis | 154,936 | MDQ6VXNlcjE1NDkzNg== | User | false |
openclaw/openclaw | 3,894,118,012 | I_kwDOQb6kR87oG4Z8 | 8,441 | https://github.com/openclaw/openclaw/issues/8441 | https://api.github.com/repos/openclaw/openclaw/issues/8441 | Feature: Add thinking and model config to skills.entries | ## Summary
Add `thinking` and `model` fields to `skills.entries.<skill>` config, allowing per-skill overrides for thinking level and model.
## Motivation
Currently, cron jobs support `thinking` and `model` overrides in their payload, but skills don't have an equivalent config option. This means:
- Users can't set a... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-04T01:56:51Z | 2026-05-05T04:14:48Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | tsukhani | 156,283 | MDQ6VXNlcjE1NjI4Mw== | User | false |
openclaw/openclaw | 4,041,704,984 | I_kwDOQb6kR87w54YY | 40,077 | https://github.com/openclaw/openclaw/issues/40077 | https://api.github.com/repos/openclaw/openclaw/issues/40077 | [Bug]: sessions_spawn thread=true fails on Discord guild channels — "Unable to create or bind a Discord thread" | ## Summary
`sessions_spawn` with `thread: true` and `mode: "session"` always fails with:
```
Unable to create or bind a Discord thread for this subagent session. Session mode is unavailable for this target.
```
...even though:
- Config is correct (`threadBindings.enabled: true`, `spawnSubagentSessions: true`)
- Bot ... | closed | completed | true | 3 | [] | [] | 2026-03-08T16:56:28Z | 2026-04-27T04:54:47Z | 2026-04-24T22:06:21Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Neo-Techpros | 266,390,099 | U_kgDOD-DKUw | User | false |
openclaw/openclaw | 4,056,102,965 | I_kwDOQb6kR87xwzg1 | 42,870 | https://github.com/openclaw/openclaw/issues/42870 | https://api.github.com/repos/openclaw/openclaw/issues/42870 | [Bug]: Custom OpenAI-compatible baseUrl returns Cloudflare 403 when messages contain backticks | ## Summary
When OpenClaw is configured to use a custom OpenAI-compatible `baseUrl` such as `https://<custom-domain>/v1`, requests consistently return a `403 Forbidden` Cloudflare WAF page when any `messages[].content` contains a backtick character (`). The same request succeeds when the backticks are removed.
This do... | closed | not_planned | true | 5 | [] | [] | 2026-03-11T06:06:54Z | 2026-04-27T04:52:00Z | 2026-04-24T22:16:09Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ljh740 | 9,330,229 | MDQ6VXNlcjkzMzAyMjk= | User | false |
openclaw/openclaw | 4,049,814,084 | I_kwDOQb6kR87xY0JE | 41,862 | https://github.com/openclaw/openclaw/issues/41862 | https://api.github.com/repos/openclaw/openclaw/issues/41862 | export-session HTML renders blank — template placeholders broken by formatter | ## Description
`/export-session` produces an HTML file (~2MB with session data) that opens in the browser but displays nothing — blank page with the dark theme background, no messages rendered.
## Environment
- OpenClaw 2026.3.8
- macOS, Discord channel
## Steps to Reproduce
1. Run `/export-session` in any Discord c... | closed | completed | true | 4 | [] | [] | 2026-03-10T06:54:34Z | 2026-05-01T04:45:00Z | 2026-04-28T06:44:47Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | briannewman | 16,906,394 | MDQ6VXNlcjE2OTA2Mzk0 | User | false |
openclaw/openclaw | 4,226,757,210 | I_kwDOQb6kR8777zJa | 63,301 | https://github.com/openclaw/openclaw/issues/63301 | https://api.github.com/repos/openclaw/openclaw/issues/63301 | OpenClaw Model Registry: GPT-4.1 family passes through unmanaged | ## Issue
OpenClaw 2026.4.7 uses the GPT-5.4 model family internally for OpenAI workloads. When a user configures a model ID outside the registry — such as `openai/gpt-4.1-nano`, `openai/gpt-4.1-mini`, or `openai/gpt-4.1` — the ID passes through to OpenAI's API raw, without OpenClaw's exec profile, tool-use optimizatio... | closed | completed | true | 1 | [] | [] | 2026-04-08T18:31:18Z | 2026-04-21T04:21:54Z | 2026-04-18T17:07:08Z | NONE | null | 20260421T060038Z | 2026-04-21T06:00:38Z | perlowja | 162,375,972 | U_kgDOCa2pJA | User | false |
openclaw/openclaw | 4,196,845,787 | I_kwDOQb6kR876Jsjb | 59,900 | https://github.com/openclaw/openclaw/issues/59900 | https://api.github.com/repos/openclaw/openclaw/issues/59900 | [Bug]: xAI grok reasoning models (grok-4-1-fast-reasoning) ignore workspace IDENTITY.md behavioral rules at 0% compliance rate | ## Bug type
Behavior bug (incorrect output/state without crash)
## Summary
When `xai/grok-4-1-fast-reasoning` (and likely other xAI fast-reasoning variants) is used as the primary model in OpenClaw, the agent produces 0% compliance with behavioral rules defined in workspace `IDENTITY.md` — specifically financial adv... | closed | completed | true | 1 | [] | [] | 2026-04-02T20:34:32Z | 2026-04-21T04:21:54Z | 2026-04-18T17:07:08Z | NONE | null | 20260421T060038Z | 2026-04-21T06:00:38Z | perlowja | 162,375,972 | U_kgDOCa2pJA | User | false |
openclaw/openclaw | 4,065,145,977 | I_kwDOQb6kR87yTTR5 | 44,134 | https://github.com/openclaw/openclaw/issues/44134 | https://api.github.com/repos/openclaw/openclaw/issues/44134 | [Google Antigravity Ban] Frequent Tool Schema Reloading Causes False Positive Anti-Abuse Detection | ## Description
When using Google Antigravity as the primary model provider, the account was banned due to 'violation of Terms of Service'. After analyzing the logs, we discovered the root cause.
## Root Cause Analysis
The Google Cloud Code Assist API was triggered multiple times in a short period, causing the accoun... | open | null | false | 4 | [] | [] | 2026-03-12T14:39:20Z | 2026-05-04T12:46:10Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | caimao9539 | 16,041,782 | MDQ6VXNlcjE2MDQxNzgy | User | false |
openclaw/openclaw | 4,030,788,779 | I_kwDOQb6kR87wQPSr | 36,754 | https://github.com/openclaw/openclaw/issues/36754 | https://api.github.com/repos/openclaw/openclaw/issues/36754 | Extension discovery silently skips symlinked plugin directories | ## Bug Summary
`discoverInDirectory` in `src/plugins/discovery.ts` uses `fs.readdirSync()` with `{ withFileTypes: true }`, then checks `entry.isDirectory()` to find plugin directories. However, `Dirent.isDirectory()` returns `false` for symlinks — so symlinked plugins in `~/.openclaw/extensions/` are **silently skippe... | closed | completed | true | 4 | [] | [] | 2026-03-05T21:41:47Z | 2026-04-30T04:34:01Z | 2026-04-27T13:17:56Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | Quackstro | 3,889,956 | MDQ6VXNlcjM4ODk5NTY= | User | false |
openclaw/openclaw | 4,141,627,133 | I_kwDOQb6kR8723Db9 | 55,030 | https://github.com/openclaw/openclaw/issues/55030 | https://api.github.com/repos/openclaw/openclaw/issues/55030 | [Bug]: WhatsApp aggressive 30-minute heartbeat causes cascading reconnects every ~60 seconds | ## Bug Description
WhatsApp gateway exhibits cascading reconnects every ~60 seconds after the first 30-minute idle reconnect. This appears to be the aggressive 30-minute heartbeat reconnect documented in #13544 and #14000, still unresolved on v2026.3.24.
## Root Cause
The gateway intentionally reconnects WhatsApp Web... | closed | duplicate | true | 8 | [
"close:duplicate",
"dedupe:child"
] | [] | 2026-03-26T08:09:45Z | 2026-04-30T04:44:28Z | 2026-04-25T07:38:51Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | clawoneloke | 260,136,245 | U_kgDOD4FdNQ | User | false |
openclaw/openclaw | 4,227,706,177 | I_kwDOQb6kR877_a1B | 63,372 | https://github.com/openclaw/openclaw/issues/63372 | https://api.github.com/repos/openclaw/openclaw/issues/63372 | [Feature]: Allow memory-core dreaming to use a configurable model override | ### Summary
Add an optional `model` field to the dreaming config so dreaming phases can use a different model than the agent's primary.
### Problem to solve
Dreaming inherits the agent's primary model with no way to override it. Users who want a cheap/free model (e.g. `google/gemini-2.5-flash`) for background dreami... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-04-08T21:16:23Z | 2026-05-02T04:29:05Z | 2026-04-29T06:39:34Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | marcelocompucorp | 43,748,427 | MDQ6VXNlcjQzNzQ4NDI3 | User | false |
openclaw/openclaw | 3,952,830,721 | I_kwDOQb6kR87rm2kB | 19,176 | https://github.com/openclaw/openclaw/issues/19176 | https://api.github.com/repos/openclaw/openclaw/issues/19176 | [bluebubbles] Add maxMessageAge filter to drop stale re-delivered webhooks | ### Problem
BlueBubbles occasionally re-delivers old message webhooks as if they are new (known BB behavior). This causes OpenClaw to process and reply to messages that are hours or days old, confusing the user.
Additionally, when the agent is overloaded (529/rate-limit), the error text ("The AI service is temporaril... | closed | completed | true | 4 | [
"stale"
] | [] | 2026-02-17T13:41:01Z | 2026-04-14T22:45:07Z | 2026-03-29T04:16:48Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | iranitehrani | 217,807,139 | U_kgDODPt5Iw | User | false |
openclaw/openclaw | 4,213,157,922 | I_kwDOQb6kR877H7Ai | 62,049 | https://github.com/openclaw/openclaw/issues/62049 | https://api.github.com/repos/openclaw/openclaw/issues/62049 | [Bug]: False-positive "plugins.allow is empty" warning in openclaw status (2026.4.5) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
### Summary
`openclaw status` reports:
`[plugins] plugins.allow is empty; discovered non-bundled plugins may auto-load...`
even though `plugins.allow` is explicitly configured and enforced.
This appears to b... | closed | completed | true | 5 | [
"bug"
] | [] | 2026-04-06T17:49:32Z | 2026-04-28T05:07:15Z | 2026-04-25T14:23:43Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | MWiegand133 | 28,255,019 | MDQ6VXNlcjI4MjU1MDE5 | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.