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 | 4,329,439,774 | I_kwDOQb6kR88AAAABAg4CHg | 71,812 | https://github.com/openclaw/openclaw/issues/71812 | https://api.github.com/repos/openclaw/openclaw/issues/71812 | openclaw update leaves plugin-runtime-deps/<version>/.../plugin-sdk in ENOTEMPTY state — every channel plugin silently fails to load | # `openclaw update` leaves `plugin-runtime-deps/<version>/.../plugin-sdk` in ENOTEMPTY state — every channel plugin silently fails to load
## Summary
After running `openclaw update` from `2026.4.23` → `2026.4.24`, every bundled channel plugin failed to load with `ENOTEMPTY, Directory not empty: .../plugin-sdk`. The g... | closed | completed | true | 1 | [] | [] | 2026-04-25T22:54:22Z | 2026-04-28T04:45:21Z | 2026-04-26T00:18:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | vksvikas | 11,882,897 | MDQ6VXNlcjExODgyODk3 | User | false |
openclaw/openclaw | 4,329,441,156 | I_kwDOQb6kR88AAAABAg4HhA | 71,813 | https://github.com/openclaw/openclaw/issues/71813 | https://api.github.com/repos/openclaw/openclaw/issues/71813 | Telegram channel: keep sendChatAction("typing") alive while a turn is in flight | # [feature] Telegram channel: keep `sendChatAction("typing")` alive while a turn is in flight
## Problem
Today the Telegram channel plugin sends `sendChatAction("typing")` once on inbound message receipt. Telegram only honors that for ~5 seconds, after which the "… typing" indicator disappears in the user's chat.
Fo... | closed | not_planned | true | 2 | [] | [] | 2026-04-25T22:55:15Z | 2026-04-29T04:37:27Z | 2026-04-26T10:42:00Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | joeywrightphoto | 114,463,115 | U_kgDOBtKRiw | User | false |
openclaw/openclaw | 4,329,444,912 | I_kwDOQb6kR88AAAABAg4WMA | 71,814 | https://github.com/openclaw/openclaw/issues/71814 | https://api.github.com/repos/openclaw/openclaw/issues/71814 | Bonjour crash loop STILL breaking gateway in 2026.4.24 — stop shipping broken defaults | ## The Situation
I just spent an afternoon debugging gateway crashes on **two separate servers** (Debian 13/PVE and a VPS) after upgrading to 2026.4.24. Root cause: the bonjour plugin throws `Unhandled promise rejection: CIAO ANNOUNCEMENT CANCELLED` and kills the process. The systemd service restarts, it crashes again... | closed | completed | true | 1 | [] | [] | 2026-04-25T22:57:29Z | 2026-04-28T04:45:30Z | 2026-04-26T00:02:08Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | amdhelper | 232,188,533 | U_kgDODdbqdQ | User | false |
openclaw/openclaw | 4,140,225,447 | I_kwDOQb6kR872xtOn | 54,855 | https://github.com/openclaw/openclaw/issues/54855 | https://api.github.com/repos/openclaw/openclaw/issues/54855 | [Bug]: Control UI exec approvals fail in multi-channel setups because exec.approval.resolve is sent without channel/sourceChannel | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Control UI exec approvals fail in multi-channel setups; clicking Allow does not approve the command, and gateway logs require an explicit channel.
### Steps to reproduce
Enable both feishu and openclaw-weixin on OpenClaw 2026.3.24.
Open C... | closed | not_planned | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-03-26T03:32:02Z | 2026-04-28T04:45:46Z | 2026-04-25T23:18:40Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zhx1012 | 103,342,885 | U_kgDOBijjJQ | User | false |
openclaw/openclaw | 4,329,488,830 | I_kwDOQb6kR88AAAABAg7Bvg | 71,818 | https://github.com/openclaw/openclaw/issues/71818 | https://api.github.com/repos/openclaw/openclaw/issues/71818 | Plugin runtime-deps install loop in 2026.4.24 — gateway event loop starved, Telegram polling stalls | ## Summary
On OpenClaw **2026.4.24** the gateway enters an infinite `npm install` loop for plugin runtime-deps. The deps install reports success, but the runtime check on the next cycle still sees them as "missing" and reinstalls. This starves the gateway event loop, causing intermittent gateway HTTP/WS timeouts and T... | closed | completed | true | 7 | [] | [] | 2026-04-25T23:21:18Z | 2026-04-28T04:38:19Z | 2026-04-26T03:16:28Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | TARSDrakon | 262,872,503 | U_kgDOD6sdtw | User | false |
openclaw/openclaw | 4,204,808,341 | I_kwDOQb6kR876oEiV | 60,856 | https://github.com/openclaw/openclaw/issues/60856 | https://api.github.com/repos/openclaw/openclaw/issues/60856 | [Bug]: Memory index with Ollama embeddings times out after 120s even though embeddings are ready and batch is disabled | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
Memory indexing fails with `memory embeddings batch timed out after 120s` when `memorySearch.provider` is set to `ollama`, even though `openclaw memory status --deep` reports `Embeddings: ready` and direct Ollama embedding requ... | closed | completed | true | 4 | [
"bug",
"bug:crash"
] | [] | 2026-04-04T12:24:13Z | 2026-04-28T04:45:46Z | 2026-04-25T23:21:54Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | awirtz | 3,907,935 | MDQ6VXNlcjM5MDc5MzU= | User | false |
openclaw/openclaw | 4,329,241,930 | I_kwDOQb6kR88AAAABAgr9Sg | 71,773 | https://github.com/openclaw/openclaw/issues/71773 | https://api.github.com/repos/openclaw/openclaw/issues/71773 | ACPX extension declares acpx 0.5.3; acpx 0.6.0 needed for reliable ACP thread sessions after upgrade | ## Summary
After upgrading OpenClaw to `2026.4.24`, ACP thread sessions only verified successfully after reinstalling `acpx@0.6.0` plugin-locally under the bundled ACPX extension.
The bundled extension still appears to declare `acpx: 0.5.3`, and the OpenClaw upgrade wiped the previous plugin-local `acpx@0.6.0` overri... | closed | completed | true | 2 | [] | [] | 2026-04-25T21:16:58Z | 2026-04-28T04:45:43Z | 2026-04-25T23:26:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | marcbal77 | 8,789,169 | MDQ6VXNlcjg3ODkxNjk= | User | false |
openclaw/openclaw | 4,229,478,808 | I_kwDOQb6kR878GLmY | 63,548 | https://github.com/openclaw/openclaw/issues/63548 | https://api.github.com/repos/openclaw/openclaw/issues/63548 | sessions_spawn fails with pairing required on local loopback with token auth | # Issue: sessions_spawn fails with "pairing required" on local loopback with token auth
## Environment
- **OpenClaw version**: 2026.4.8 (upgraded from 2026.4.2)
- **OS**: Windows 10 (10.0.26200, x64)
- **Node**: v24.14.1
- **Gateway config**: `bind: loopback`, `auth.mode: token`
- **Channel**: Feishu (direct chat)
##... | closed | completed | true | 2 | [] | [] | 2026-04-09T04:46:20Z | 2026-04-28T04:45:42Z | 2026-04-25T23:26:50Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | yuren19810323-pixel | 247,734,546 | U_kgDODsQhEg | User | false |
openclaw/openclaw | 4,329,504,585 | I_kwDOQb6kR88AAAABAg7_SQ | 71,821 | https://github.com/openclaw/openclaw/issues/71821 | https://api.github.com/repos/openclaw/openclaw/issues/71821 | [Bug]: OpenClaw 2026.4.24: gateway crash loop caused by bonjour/mDNS discovery (CIAO PROBING/ANNOUNCEMENT CANCELLED) | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
After upgrading to OpenClaw 2026.4.24, the gateway entered a crash loop shortly after startup.
While the loop was active:
- the local web dashboard was unavailable
- Telegram connectivity was unavailable
- systemd kept restart... | closed | completed | true | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-25T23:29:38Z | 2026-04-28T04:45:29Z | 2026-04-26T00:02:10Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Kepar1337 | 209,847,643 | U_kgDODIIFWw | User | false |
openclaw/openclaw | 4,132,273,120 | I_kwDOQb6kR872TXvg | 54,220 | https://github.com/openclaw/openclaw/issues/54220 | https://api.github.com/repos/openclaw/openclaw/issues/54220 | `[Bug]: Discord inbound reply delivery bypasses message_sending plugin hook` | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When a lifecycle plugin registers a `message_sending` hook, the hook is never called for Discord inbound replies because `deliverDiscordReply` sends directly to the Discord API without invoking `runMessageSending`, causing raw unfiltered LL... | closed | completed | true | 1 | [
"bug"
] | [
"vincentkoc"
] | 2026-03-25T04:06:50Z | 2026-04-28T04:45:41Z | 2026-04-25T23:30:47Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Rehte | 65,696,030 | MDQ6VXNlcjY1Njk2MDMw | User | false |
openclaw/openclaw | 4,329,508,985 | I_kwDOQb6kR88AAAABAg8QeQ | 71,823 | https://github.com/openclaw/openclaw/issues/71823 | https://api.github.com/repos/openclaw/openclaw/issues/71823 | [Bug]: fly.toml gateway startup always crashes on first-boot due to missing controlUi.allowedOrigins with non-loopback bind | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 10.0 / 10.0 | 10.0 / 10.0 |
| **Severity** | Critical | Critical |
| **Vector** | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H | CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H |
| **Calculator**... | closed | completed | true | 1 | [] | [] | 2026-04-25T23:31:26Z | 2026-04-28T04:44:40Z | 2026-04-26T00:35:09Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 4,085,694,483 | I_kwDOQb6kR87zhsAT | 48,622 | https://github.com/openclaw/openclaw/issues/48622 | https://api.github.com/repos/openclaw/openclaw/issues/48622 | Hook message:sent event not firing for Feishu channel | ## Description
The `message:sent` hook event is not triggered when sending messages via the Feishu channel.
## Steps to Reproduce
1. Configure a hook that listens to `message:sent` (e.g., andrew-core-memory)
2. Send a message to the agent via Feishu
3. Observe that `message:received` fires but `message:sent` never fir... | closed | completed | true | 2 | [] | [
"vincentkoc"
] | 2026-03-17T01:34:45Z | 2026-04-28T04:45:39Z | 2026-04-25T23:35:09Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | panyu-collab | 264,834,155 | U_kgDOD8kMaw | User | false |
openclaw/openclaw | 4,094,464,185 | I_kwDOQb6kR870DJC5 | 49,765 | https://github.com/openclaw/openclaw/issues/49765 | https://api.github.com/repos/openclaw/openclaw/issues/49765 | [Bug]: message:sent hooks not work | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
follow the hooks guide add message send hook, but not work
### Steps to reproduce
1. open website https://docs.openclaw.ai/automation/hooks#message-events
2. copy the sample code "Example: Message Logger Hook"
```
const isMessageReceived... | closed | completed | true | 6 | [
"bug",
"bug:behavior"
] | [
"vincentkoc"
] | 2026-03-18T11:22:59Z | 2026-04-28T04:45:38Z | 2026-04-25T23:35:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Bugaoxingxx | 40,537,064 | MDQ6VXNlcjQwNTM3MDY0 | User | false |
openclaw/openclaw | 4,115,299,303 | I_kwDOQb6kR871Snvn | 52,144 | https://github.com/openclaw/openclaw/issues/52144 | https://api.github.com/repos/openclaw/openclaw/issues/52144 | [Bug]: why hook handler can not capture message:sent event | ### Bug type
Regression (worked before, now fails)
### Summary
Feishu channel, using hooks can only capture message:received and message:preprocessed, but not message:sent.
### Steps to reproduce
1. openclaw version: OpenClaw 2026.3.13 (61d171a)
2. Add a simple Hook, capture message:receive, message:preprocessed， ... | closed | completed | true | 4 | [
"bug",
"regression"
] | [
"vincentkoc"
] | 2026-03-22T08:11:23Z | 2026-04-28T04:45:36Z | 2026-04-25T23:35:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | inkhare | 58,814,395 | MDQ6VXNlcjU4ODE0Mzk1 | User | false |
openclaw/openclaw | 4,329,515,672 | I_kwDOQb6kR88AAAABAg8qmA | 71,826 | https://github.com/openclaw/openclaw/issues/71826 | https://api.github.com/repos/openclaw/openclaw/issues/71826 | Expose stable cron jobId in plugin hook context | ## Summary
`before_agent_reply` plugin hooks for cron-triggered agent runs receive `runId`, but not the stable cron `jobId`.
For cron workflows, `runId` is the execution id (for example `cron:<jobId>:<startedAt>` or a manual execution id), not the persistent cron job identifier. That makes it hard for plugins to matc... | closed | completed | true | 3 | [] | [] | 2026-04-25T23:35:12Z | 2026-04-28T04:43:02Z | 2026-04-26T01:11:42Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | scottgl9 | 10,623,288 | MDQ6VXNlcjEwNjIzMjg4 | User | false |
openclaw/openclaw | 4,074,400,859 | I_kwDOQb6kR87y2mxb | 45,661 | https://github.com/openclaw/openclaw/issues/45661 | https://api.github.com/repos/openclaw/openclaw/issues/45661 | [Share] Field-tested browser automation routing playbook: Playwright workers first, built-in browser last | ## Summary
This is a field-tested browser automation playbook from a real OpenClaw machine after multiple days of browser breakage, worker repairs, smoke tests, and live use across Telegram + TUI.
Goal: share what is actually working right now, invite others to compare notes, and avoid having every operator rediscove... | closed | not_planned | true | 5 | [] | [] | 2026-03-14T02:36:51Z | 2026-04-28T04:45:36Z | 2026-04-25T23:36:07Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | SatsConvert | 212,145,289 | U_kgDODKUUiQ | User | false |
openclaw/openclaw | 4,116,368,926 | I_kwDOQb6kR871Ws4e | 52,336 | https://github.com/openclaw/openclaw/issues/52336 | https://api.github.com/repos/openclaw/openclaw/issues/52336 | Bug: browser automation intermittently fails with gateway closed (1000 normal closure) | ## Summary
OpenClaw browser automation is unstable in a local/headed Chrome setup: `openclaw browser start` can succeed, and some browser RPCs (for example `browser tabs` or `browser tab new`) may work intermittently, but subsequent browser commands often fail with:
```text
gateway connect failed: Error: gateway clos... | open | null | false | 3 | [] | [] | 2026-03-22T15:45:33Z | 2026-05-04T21:00:27Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | Aniloid2 | 17,166,325 | MDQ6VXNlcjE3MTY2MzI1 | User | false |
openclaw/openclaw | 4,329,522,717 | I_kwDOQb6kR88AAAABAg9GHQ | 71,829 | https://github.com/openclaw/openclaw/issues/71829 | https://api.github.com/repos/openclaw/openclaw/issues/71829 | Discord voice channel: bot joins but audio capture pipeline never starts (no PCM/STT/logs) | ## Bug Report
Discord voice channel bot joins successfully (confirmed via API: `GET /guilds/{guild}/voice-states/@me` shows bot in channel with active session), but the audio capture pipeline never starts. No PCM capture, no WAV temp files, no STT transcription, zero voice-related logs despite speaking in the channel.... | closed | not_planned | true | 0 | [] | [] | 2026-04-25T23:39:15Z | 2026-04-28T04:45:35Z | 2026-04-25T23:40:20Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zcode-apps | 87,060,116 | MDQ6VXNlcjg3MDYwMTE2 | User | false |
openclaw/openclaw | 4,055,171,172 | I_kwDOQb6kR87xtQBk | 42,644 | https://github.com/openclaw/openclaw/issues/42644 | https://api.github.com/repos/openclaw/openclaw/issues/42644 | Race condition: hook runner reinitialization causes message_sending hook to be skipped | ## Bug Description
The `message_sending` hook is occasionally skipped, allowing messages to bypass plugin hooks entirely. This causes inconsistent enforcement of rules and other plugin functionality.
**Testing Environment:** Telegram chat (channels.telegram.streaming: "partial")
## Use Case: Custom Rules Engine Plug... | closed | completed | true | 1 | [] | [
"vincentkoc"
] | 2026-03-11T01:47:04Z | 2026-04-28T04:45:34Z | 2026-04-25T23:46:02Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | mikronn2 | 105,823,039 | U_kgDOBk67Pw | User | false |
openclaw/openclaw | 4,329,538,636 | I_kwDOQb6kR88AAAABAg-ETA | 71,832 | https://github.com/openclaw/openclaw/issues/71832 | https://api.github.com/repos/openclaw/openclaw/issues/71832 | Feature Request: subclaw - Independent Child Agent for Per-Directory Task Processing | ## Feature Request: subclaw - Independent Child Agent for Per-Directory Task Processing
### Summary
I propose adding a `subclaw` feature to OpenClaw that would allow users to launch independent child agents focused on processing tasks within specific directories. This would mirror how Claude Code opens independent TUI... | closed | not_planned | true | 2 | [] | [] | 2026-04-25T23:48:00Z | 2026-04-29T04:37:27Z | 2026-04-26T10:42:03Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | JIRBOY | 49,422,297 | MDQ6VXNlcjQ5NDIyMjk3 | User | false |
openclaw/openclaw | 4,329,545,775 | I_kwDOQb6kR88AAAABAg-gLw | 71,834 | https://github.com/openclaw/openclaw/issues/71834 | https://api.github.com/repos/openclaw/openclaw/issues/71834 | Regression: 2026.4.22 invalid thinking bundle; 2026.4.24 Telegram webhook startup fails on VPS | ## Summary
A production VPS node stopped responding after upgrading OpenClaw. Two upgrade blockers were observed while restoring service:
1. `openclaw@2026.4.22` installed from npm contains an invalid generated bundle and the gateway crash-loops on startup.
2. `openclaw@2026.4.24` fixes the syntax error and the gatewa... | closed | completed | true | 2 | [] | [] | 2026-04-25T23:52:08Z | 2026-04-30T04:32:08Z | 2026-04-28T03:36:43Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | jinon86 | 247,078,695 | U_kgDODrofJw | User | false |
openclaw/openclaw | 4,329,551,146 | I_kwDOQb6kR88AAAABAg-1Kg | 71,835 | https://github.com/openclaw/openclaw/issues/71835 | https://api.github.com/repos/openclaw/openclaw/issues/71835 | OpenClaw 2026.4.24 update can leave macOS gateway in mixed-version/unhealthy plugin-runtime-deps state | ## Summary
After attempting to update OpenClaw to 2026.4.24 on macOS, the gateway can enter an unhealthy/mixed-version startup state.
Observed failure modes:
- `openclaw update` appears to partially apply 2026.4.24, but the LaunchAgent/gateway may still run 2026.4.23.
- Config/plugins can require `>=2026.4.24` while... | closed | completed | true | 5 | [] | [] | 2026-04-25T23:55:29Z | 2026-04-29T04:42:29Z | 2026-04-26T07:06:08Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | StefanAIpel | 130,749,432 | U_kgDOB8sT-A | User | false |
openclaw/openclaw | 4,329,557,587 | I_kwDOQb6kR88AAAABAg_OUw | 71,838 | https://github.com/openclaw/openclaw/issues/71838 | https://api.github.com/repos/openclaw/openclaw/issues/71838 | [Bug] TUI duplicate message rendering after websocket reconnect — regression in 2026.4.24 | ## Description
Messages render twice (duplicated) in TUI after websocket reconnect. This started immediately after updating to 2026.4.24.
## Reproduction
1. Run `openclaw gateway restart` (clears duplication temporarily)
2. Open TUI with `openclaw tui` — messages render cleanly
3. Close TUI / disconnect
4. Reopen TUI ... | closed | duplicate | true | 2 | [] | [] | 2026-04-25T23:59:37Z | 2026-04-28T04:44:10Z | 2026-04-26T01:10:05Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | leetaur | 279,354,240 | U_kgDOEKabgA | User | false |
openclaw/openclaw | 4,202,072,132 | I_kwDOQb6kR876dohE | 60,514 | https://github.com/openclaw/openclaw/issues/60514 | https://api.github.com/repos/openclaw/openclaw/issues/60514 | Feature Request: session:end hook + autoDream equivalent capabilities | ## Feature Request: Session Lifecycle Hooks & Claude Code-style Auto-Dream
### Context
Based on the Claude Code open-source release (npm @anthropic-ai/claude-code), there are several core session lifecycle features that are currently missing from OpenClaw.
### Requested Features
**1. session:end Hook (Critical)**
C... | closed | completed | true | 1 | [] | [
"vincentkoc"
] | 2026-04-03T20:21:41Z | 2026-04-28T04:45:30Z | 2026-04-26T00:02:04Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ybbms777 | 85,210,100 | MDQ6VXNlcjg1MjEwMTAw | User | false |
openclaw/openclaw | 4,223,849,801 | I_kwDOQb6kR877wtVJ | 63,102 | https://github.com/openclaw/openclaw/issues/63102 | https://api.github.com/repos/openclaw/openclaw/issues/63102 | [Bug][P2] Plugin tool name conflict triggered on every message: qqbot_channel_api / qqbot_remind | ## Description
Every time the agent processes a message, two ERROR-level log entries are produced:
\\\
plugin tool name conflict (openclaw-qqbot): qqbot_channel_api
plugin tool name conflict (openclaw-qqbot): qqbot_remind
\\\
This happens consistently on every single message, creating significant log noise.
## Envi... | closed | completed | true | 1 | [] | [] | 2026-04-08T10:31:55Z | 2026-04-28T04:45:28Z | 2026-04-26T00:09:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | allenwywang | 268,795,201 | U_kgDOEAV9QQ | User | false |
openclaw/openclaw | 4,199,732,744 | I_kwDOQb6kR876UtYI | 60,219 | https://github.com/openclaw/openclaw/issues/60219 | https://api.github.com/repos/openclaw/openclaw/issues/60219 | Plugin tools not registered when loaded twice with different cacheKey (preferSetupRuntimeForChannelPlugins) | ## Summary
When Gateway starts, plugins are loaded twice with different `preferSetupRuntimeForChannelPlugins` values, causing different cacheKeys. This results in tool registration being skipped on the second load due to `pluginRegistrationState` not being reset.
## Environment
- OpenClaw version: 2026.3.28
- Plugin... | closed | completed | true | 3 | [] | [] | 2026-04-03T10:05:18Z | 2026-04-28T04:45:27Z | 2026-04-26T00:09:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | highland0971 | 2,759,270 | MDQ6VXNlcjI3NTkyNzA= | User | false |
openclaw/openclaw | 4,092,468,065 | I_kwDOQb6kR87z7hth | 49,412 | https://github.com/openclaw/openclaw/issues/49412 | https://api.github.com/repos/openclaw/openclaw/issues/49412 | [Bug]: Feishu plugin - duplicate plugin id warning on startup | ### Bug type
Regression (worked before, now fails)
### Summary
Feishu plugin shows duplicate plugin id warning on startup despite functioning normally
### Steps to reproduce
1. Configure OpenClaw with feishu plugin enabled
2. Start the OpenClaw gateway
3. Check gateway logs or status output
### Expected behavior
... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-03-18T03:30:04Z | 2026-04-28T04:45:26Z | 2026-04-26T00:09:55Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zengfei88 | 129,354,103 | U_kgDOB7XJdw | User | false |
openclaw/openclaw | 4,224,543,651 | I_kwDOQb6kR877zWuj | 63,153 | https://github.com/openclaw/openclaw/issues/63153 | https://api.github.com/repos/openclaw/openclaw/issues/63153 | [Bug]: Bonjour advertiser stalling triggers Gateway self-healing loop with log flooding | ## Bug type
Behavior bug (incorrect output/state without crash)
## Beta release blocker
No
## Summary
mDNSResponder (Bonjour) occasionally gets stuck in "announcing" state on macOS, causing Gateway to repeatedly restart the advertiser and flood logs with warnings.
## Steps to reproduce
1. Run Gateway on macOS with B... | closed | completed | true | 1 | [] | [] | 2026-04-08T12:37:43Z | 2026-04-28T04:45:25Z | 2026-04-26T00:14:04Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | chenlong220192 | 25,541,841 | MDQ6VXNlcjI1NTQxODQx | User | false |
openclaw/openclaw | 4,226,073,626 | I_kwDOQb6kR8775MQa | 63,248 | https://github.com/openclaw/openclaw/issues/63248 | https://api.github.com/repos/openclaw/openclaw/issues/63248 | [Bug] v2026.4.8 — Bonjour/mDNS service-advertise sidecar pegs 99% CPU in headless Docker, eventually triggers container exit | # Title
[Bug] v2026.4.8 — Bonjour/mDNS service-advertise sidecar pegs 99% CPU in headless Docker (no `--network=host`), eventually triggers container exit
## Body
## Summary
Running `ghcr.io/openclaw/openclaw:2026.4.8` in a standard Docker container (default bridge network, no `--network=host`) results in a child `o... | closed | completed | true | 1 | [] | [] | 2026-04-08T16:33:18Z | 2026-04-28T04:45:24Z | 2026-04-26T00:14:06Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Nbruewer | 271,288,845 | U_kgDOECuKDQ | User | false |
openclaw/openclaw | 3,883,352,420 | I_kwDOQb6kR87nd0Fk | 6,711 | https://github.com/openclaw/openclaw/issues/6711 | https://api.github.com/repos/openclaw/openclaw/issues/6711 | Feature request: gateway:shutdown and gateway:pre-restart hook events | ## Summary
Add lifecycle hook events for gateway shutdown/restart to enable pre-shutdown automations like notifying users.
## Use Case
When running a multi-user BobNet setup with multiple agents (bob, olivia, jarvis, etc.), it's useful to broadcast a notification to all connected users before the gateway restarts, e... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-02T00:02:50Z | 2026-04-28T04:45:19Z | 2026-04-26T00:21:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | buildzerobob | 257,883,014 | U_kgDOD177hg | User | false |
openclaw/openclaw | 3,883,400,667 | I_kwDOQb6kR87nd_3b | 6,722 | https://github.com/openclaw/openclaw/issues/6722 | https://api.github.com/repos/openclaw/openclaw/issues/6722 | WhatsApp: Add link preview support (generateHighQualityLinkPreview) | ## Feature Request
When sending messages with URLs via WhatsApp, the link preview (thumbnail + title) is not generated.
## Technical Details
Baileys already supports this via the `generateHighQualityLinkPreview` option in `makeWASocket`:
```javascript
// Current code in dist/web/session.js (line 75):
const sock = m... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-02T00:26:24Z | 2026-04-28T04:45:18Z | 2026-04-26T00:22:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | noamabta-stellarfox | 246,243,321 | U_kgDODq1f-Q | User | false |
openclaw/openclaw | 3,883,763,459 | I_kwDOQb6kR87nfYcD | 6,842 | https://github.com/openclaw/openclaw/issues/6842 | https://api.github.com/repos/openclaw/openclaw/issues/6842 | Feature Request: Add A2A (Agent-to-Agent) Protocol Support | ## Summary
Add support for the [A2A (Agent-to-Agent) Protocol](https://a2a-protocol.org) to enable Clawdbot instances to communicate and collaborate with other AI agents.
## Motivation
As AI agents become more prevalent, interoperability between different agent systems is increasingly important. The A2A protocol, now ... | closed | not_planned | true | 8 | [
"enhancement"
] | [] | 2026-02-02T03:35:19Z | 2026-04-28T04:30:31Z | 2026-04-26T00:22:13Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ss7lyb | 26,716,704 | MDQ6VXNlcjI2NzE2NzA0 | User | false |
openclaw/openclaw | 3,885,326,371 | I_kwDOQb6kR87nlWAj | 7,080 | https://github.com/openclaw/openclaw/issues/7080 | https://api.github.com/repos/openclaw/openclaw/issues/7080 | Feature Request: Enable Processing of WhatsApp Emotion and Poll Responses by Agents | **Description:**
Currently, OpenClaw agents can successfully send WhatsApp polls to users via the `message.poll` tool. However, there is no documented mechanism or tool available for an agent/skill to *receive, listen for, or programmatically process* a user's responses to these polls.
This limitation prevents the im... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-02-02T11:04:28Z | 2026-04-28T04:45:18Z | 2026-04-26T00:22:25Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tyronet | 8,385,489 | MDQ6VXNlcjgzODU0ODk= | User | false |
openclaw/openclaw | 3,887,165,196 | I_kwDOQb6kR87nsW8M | 7,312 | https://github.com/openclaw/openclaw/issues/7312 | https://api.github.com/repos/openclaw/openclaw/issues/7312 | [Feature]: Support OpenTelemetry GenAI Auto-Instrumentation (OpenLLMetry / IITM) |
## Summary
When building an OpenTelemetry observability plugin for OpenClaw, it is currently **impossible** to use standard GenAI auto-instrumentation libraries (OpenLLMetry / `@traceloop/instrumentation-x`) to produce for example `anthropic.chat` spans with full semantic GenAI attributes. This is due to ESM module ... | closed | completed | true | 6 | [
"enhancement"
] | [] | 2026-02-02T17:39:29Z | 2026-04-28T04:45:17Z | 2026-04-26T00:22:34Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | henrikrexed | 81,800,915 | MDQ6VXNlcjgxODAwOTE1 | User | false |
openclaw/openclaw | 3,888,302,623 | I_kwDOQb6kR87nwsof | 7,579 | https://github.com/openclaw/openclaw/issues/7579 | https://api.github.com/repos/openclaw/openclaw/issues/7579 | [Feature]: Perplexity pplx‑api as a first‑class provider (Not Only Sonar) | ## Summary
I’m a Perplexity Pro user and I’d love to plug my existing Perplexity models into Openclaw the same way I can with Anthropic, Google, Grok, etc. Right now the only Perplexity‑related option I can see is a hidden Sonar model, and it doesn’t show up as a normal provider during setup. That means I’m paying for... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-03T00:34:55Z | 2026-04-28T04:45:16Z | 2026-04-26T00:22:57Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | RealHaywoodJ | 193,286,266 | U_kgDOC4VQeg | User | false |
openclaw/openclaw | 3,888,620,385 | I_kwDOQb6kR87nx6Nh | 7,637 | https://github.com/openclaw/openclaw/issues/7637 | https://api.github.com/repos/openclaw/openclaw/issues/7637 | Show heartbeat interval and model on /status | ## Feature Request
The `/status` command should display:
1. **Heartbeat interval** - Currently configured via `agents.defaults.heartbeat.every` (e.g., `1h`)
2. **Model** - The current model being used for the session
This gives users quick visibility into their agent's configuration without needing to dig into confi... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-03T02:20:25Z | 2026-04-28T04:45:15Z | 2026-04-26T00:23:02Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | popcorn-spec | 258,968,874 | U_kgDOD2-NKg | User | false |
openclaw/openclaw | 3,889,064,930 | I_kwDOQb6kR87nzmvi | 7,720 | https://github.com/openclaw/openclaw/issues/7720 | https://api.github.com/repos/openclaw/openclaw/issues/7720 | Feature Request: Tool Allowlist Config (tools.exec.allowlist) | **Feature:** Tool-level command allowlist configuration
**What happened:**
We attempted to implement OWASP security hardening with this config:
```javascript
{
"tools": {
"exec": {
"security": "allowlist",
"allowlist": ["ls", "cat", "grep", "find", "git", "docker"]
}
}
}
```
**Result:**
```
In... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-03T04:31:33Z | 2026-04-28T04:45:15Z | 2026-04-26T00:23:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | LumenLantern | 258,675,890 | U_kgDOD2sUsg | User | false |
openclaw/openclaw | 3,889,076,307 | I_kwDOQb6kR87nzphT | 7,724 | https://github.com/openclaw/openclaw/issues/7724 | https://api.github.com/repos/openclaw/openclaw/issues/7724 | Feature Request: Agent loop lifecycle hooks (thinking, response, message events) | ## Summary
Request to add hook events for the agent turn lifecycle, enabling hardware/presence integrations that respond to AI cognitive state.
## Context
Building an AI presence device (Unit 8×8×8 LED cube) that needs to reflect the agent's internal state through light. Current hooks only cover commands and session... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-03T04:36:03Z | 2026-04-28T04:45:14Z | 2026-04-26T00:23:20Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | veteze | 7,168,151 | MDQ6VXNlcjcxNjgxNTE= | User | false |
openclaw/openclaw | 3,890,784,570 | I_kwDOQb6kR87n6Kk6 | 7,959 | https://github.com/openclaw/openclaw/issues/7959 | https://api.github.com/repos/openclaw/openclaw/issues/7959 | Feature: Identity file versioning / changelog | # Feature Request: Identity File Versioning / Changelog
## Summary
Built-in automatic versioning and changelog for core identity files (SOUL.md, USER.md, AGENTS.md, etc.) that tracks what changed, when, and optionally why.
## Problem
AI assistants using OpenClaw develop over time. Their identity files (SOUL.md, USE... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-03T11:27:59Z | 2026-04-28T04:45:13Z | 2026-04-26T00:23:30Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jlkirbee | 6,578,190 | MDQ6VXNlcjY1NzgxOTA= | User | false |
openclaw/openclaw | 4,075,792,997 | I_kwDOQb6kR87y76pl | 46,137 | https://github.com/openclaw/openclaw/issues/46137 | https://api.github.com/repos/openclaw/openclaw/issues/46137 | [Bug]: Browser automation fails with "tab not found" after form submission (targetId becomes stale) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When using OpenClaw browser automation (profile=openclaw or chrome-relay), after submitting a form (clicking a button that triggers navigation or AJAX), subsequent commands (focus, screenshot, evaluate) fail with:
`GatewayClientRequestError... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-03-14T13:10:04Z | 2026-04-28T04:45:12Z | 2026-04-26T00:23:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | amaralrj | 3,217,560 | MDQ6VXNlcjMyMTc1NjA= | User | false |
openclaw/openclaw | 3,891,615,440 | I_kwDOQb6kR87n9VbQ | 8,064 | https://github.com/openclaw/openclaw/issues/8064 | https://api.github.com/repos/openclaw/openclaw/issues/8064 | Notion skill: simplify API key setup and first-page bootstrap | ## Problem
Setting up Notion integration for OpenClaw requires multiple manual steps:
1. Create integration at https://notion.so/my-integrations
2. Copy API key
3. Store key at `~/.config/notion/api_key`
4. **Manually share pages with the integration in Notion UI** ← This is the friction point
Internal Notion integr... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-03T14:41:46Z | 2026-04-28T04:45:12Z | 2026-04-26T00:24:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ldraney | 107,153,866 | U_kgDOBmMJyg | User | false |
openclaw/openclaw | 3,891,780,343 | I_kwDOQb6kR87n99r3 | 8,080 | https://github.com/openclaw/openclaw/issues/8080 | https://api.github.com/repos/openclaw/openclaw/issues/8080 | Feature request: Image resize/maxBytes option for Read tool | ## Problem
When an agent uses the `Read` tool on large images, the full image is sent to the model, consuming significant context tokens. This can contribute to context overflow, especially in sessions with multiple image operations.
## Current Workaround
I've created a wrapper script (`view-image`) that:
1. Checks ... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-03T15:16:13Z | 2026-04-28T04:45:11Z | 2026-04-26T00:24:07Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Enashka | 19,220,062 | MDQ6VXNlcjE5MjIwMDYy | User | false |
openclaw/openclaw | 3,892,839,654 | I_kwDOQb6kR87oCATm | 8,234 | https://github.com/openclaw/openclaw/issues/8234 | https://api.github.com/repos/openclaw/openclaw/issues/8234 | [Feature]: Intel AMX support | ## Summary
Intel AMX support acceleration for local models/offline processing acceleration
## Proposed solution
Provide clarity or configuration options for AMX support enablement for local model running
## Additional context
https://openmetal.io/resources/blog/intel-amx-ai-inference-performance/
https://dl.acm.or... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-03T19:32:26Z | 2026-04-28T04:45:10Z | 2026-04-26T00:24:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | gdwnldsKSC | 3,374,148 | MDQ6VXNlcjMzNzQxNDg= | User | false |
openclaw/openclaw | 3,893,461,577 | I_kwDOQb6kR87oEYJJ | 8,336 | https://github.com/openclaw/openclaw/issues/8336 | https://api.github.com/repos/openclaw/openclaw/issues/8336 | Feature Request: Inline Agent Task with Tools (agent_task) | ## Problem
Currently there's a gap between `llm_task` and `sessions_spawn`:
| Tool | Tools Access | Blocking | Message Cost |
|------|-------------|----------|--------------|
| `llm_task` | ❌ No | ✅ Sync | 0 (part of turn) |
| `sessions_spawn` | ✅ Yes | ❌ Async | +1 message |
When I need a cheap model (Haiku) to do ... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-03T22:06:26Z | 2026-04-28T04:45:09Z | 2026-04-26T00:24:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | buildzerobob | 257,883,014 | U_kgDOD177hg | User | false |
openclaw/openclaw | 3,893,552,812 | I_kwDOQb6kR87oEuas | 8,354 | https://github.com/openclaw/openclaw/issues/8354 | https://api.github.com/repos/openclaw/openclaw/issues/8354 | Feature Request: Hook events for turn completion and sub-agent lifecycle | ## Summary
Add hook events that fire on LLM completions and sub-agent spawn/complete, enabling external systems to track model usage, costs, and delegation patterns in real-time.
## Use Case
I'm building a dashboard to track:
- Per-model token usage and costs
- Sub-agent delegation patterns (which agents spawn which... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-03T22:34:06Z | 2026-04-28T04:45:09Z | 2026-04-26T00:24:37Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | TuftyBruno | 72,035,488 | MDQ6VXNlcjcyMDM1NDg4 | User | false |
openclaw/openclaw | 3,893,700,343 | I_kwDOQb6kR87oFSb3 | 8,364 | https://github.com/openclaw/openclaw/issues/8364 | https://api.github.com/repos/openclaw/openclaw/issues/8364 | Feature: Route /reasoning output to a separate Discord channel | ## Problem
When `/reasoning on` is enabled, the reasoning output is sent as a separate message in the same channel as the reply. For extended thinking models (Opus, Sonnet with high thinking), this can be extremely long and pollutes the conversation channel.
## Proposal
Add a config option to redirect reasoning mess... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-03T23:17:19Z | 2026-04-28T04:45:08Z | 2026-04-26T00:24:42Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | DeetBot | 257,852,771 | U_kgDOD16FYw | User | false |
openclaw/openclaw | 3,894,048,914 | I_kwDOQb6kR87oGniS | 8,429 | https://github.com/openclaw/openclaw/issues/8429 | https://api.github.com/repos/openclaw/openclaw/issues/8429 | Add browser action recording and coordinate-based clicking for canvas automation | I've been using OpenClaw to automate some tasks with browsers and ran into a limitation with the browser tool. Right now it works great for clicking buttons and filling forms because it uses the accessibility tree, but it can't interact with canvas elements like charts. I'm trying to draw lines on a chart by clicking ... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-04T01:32:33Z | 2026-04-28T04:45:07Z | 2026-04-26T00:24:46Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | PrincePlatypus | 193,755,325 | U_kgDOC4x4vQ | User | false |
openclaw/openclaw | 3,894,928,897 | I_kwDOQb6kR87oJ-YB | 8,615 | https://github.com/openclaw/openclaw/issues/8615 | https://api.github.com/repos/openclaw/openclaw/issues/8615 | [Feature]: Native multi-API-key support for OpenRouter provider with load balancing and fallback | ## Summary
OpenClaw currently supports only a single `OPENROUTER_API_KEY` (set via environment variable, config, or `openclaw auth set`).
For heavy users or when running many agent tasks in parallel, this becomes limiting because:
- A single key can hit rate limits (429 errors) quickly.
- Quota exhaustion on one ke... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-04T06:30:46Z | 2026-04-28T04:45:06Z | 2026-04-26T00:24:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | hklcf | 410,302 | MDQ6VXNlcjQxMDMwMg== | User | false |
openclaw/openclaw | 3,896,647,984 | I_kwDOQb6kR87oQiEw | 8,843 | https://github.com/openclaw/openclaw/issues/8843 | https://api.github.com/repos/openclaw/openclaw/issues/8843 | Feature: Expose Telegram file_id for audio/document attachments (enable download) | ## Summary
When a user sends an audio file (not a voice note) or document attachment via Telegram, the agent receives `<media:audio>` or `<media:document>` tags but **no `file_id`** is exposed, making it impossible to download the file via Telegram Bot API.
## Use Case
I want to transcribe audio files sent by users ... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-04T13:47:14Z | 2026-04-28T04:45:06Z | 2026-04-26T00:25:11Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | danielo515 | 2,270,425 | MDQ6VXNlcjIyNzA0MjU= | User | false |
openclaw/openclaw | 3,897,588,051 | I_kwDOQb6kR87oUHlT | 8,962 | https://github.com/openclaw/openclaw/issues/8962 | https://api.github.com/repos/openclaw/openclaw/issues/8962 | Feature Request: Custom CSS/Theme Support for Webchat Interface | ### Use Case
Users want to customize the webchat UI to match their preferences and workflows. This includes dark themes, custom color schemes, branding, and better integration with desktop aesthetics.
### Proposed Solution
Add a `gateway.webchat.customCss` configuration option:
```yaml
gateway:
webchat:
customC... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-04T17:20:16Z | 2026-04-28T04:45:05Z | 2026-04-26T00:25:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | prashant-em | 233,186,201 | U_kgDODeYjmQ | User | false |
openclaw/openclaw | 3,897,803,622 | I_kwDOQb6kR87oU8Nm | 8,995 | https://github.com/openclaw/openclaw/issues/8995 | https://api.github.com/repos/openclaw/openclaw/issues/8995 | Enhancement: Spawn Verification & Worker Health Monitoring | # OpenClaw Enhancement Proposal: Spawn Verification & Worker Health
**Status:** Draft for upstream submission
**Author:** Ruby (Peter's agent workspace)
**Date:** 2026-02-04
**Related:** `sessions_spawn`, `sessions_list`, worker management
---
## Problem Statement
The `sessions_spawn` tool operates on a **fir... | closed | completed | true | 4 | [
"enhancement"
] | [] | 2026-02-04T18:13:55Z | 2026-04-28T04:45:03Z | 2026-04-26T00:27:18Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | petersykim | 35,867,020 | MDQ6VXNlcjM1ODY3MDIw | User | false |
openclaw/openclaw | 3,898,741,456 | I_kwDOQb6kR87oYhLQ | 9,142 | https://github.com/openclaw/openclaw/issues/9142 | https://api.github.com/repos/openclaw/openclaw/issues/9142 | Feature Request: Built-in Context & Memory Management | ## Summary
Add built-in context management and memory optimization features to reduce token costs and prevent context bloat.
## Problem
Long-running agents accumulate context through:
- Growing memory files (daily logs, learnings, decisions)
- Skill descriptions loaded every turn
- Cron job payloads and histories
- Se... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-04T22:44:58Z | 2026-04-28T04:45:03Z | 2026-04-26T00:27:25Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | sachah | 36,465,654 | MDQ6VXNlcjM2NDY1NjU0 | User | false |
openclaw/openclaw | 3,902,920,602 | I_kwDOQb6kR87oodea | 9,812 | https://github.com/openclaw/openclaw/issues/9812 | https://api.github.com/repos/openclaw/openclaw/issues/9812 | Feature: Agent-Level Rate Limiting Support | ## Problem Statement
Currently, clawdbot does not support configuring rate limits at the agent level in the configuration file. While global commands settings and other agent-specific configurations exist, there is no way to apply rate limiting restrictions to individual agents.
## Desired Behavior
I would like to add... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-05T18:21:52Z | 2026-04-28T04:45:01Z | 2026-04-26T00:27:47Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | MarcoFPO | 216,275,080 | U_kgDODOQYiA | User | false |
openclaw/openclaw | 3,903,576,558 | I_kwDOQb6kR87oq9nu | 9,897 | https://github.com/openclaw/openclaw/issues/9897 | https://api.github.com/repos/openclaw/openclaw/issues/9897 | Feature: Custom pairing message for DM policy | ## Summary
When `dm.policy: "pairing"` is set, unknown senders receive a hardcoded message with a pairing code and CLI instructions. It would be helpful to allow customizing this message.
## Use Case
For workspace deployments where the bot owner wants users to email the code (or contact them another way) rather than r... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-05T21:01:48Z | 2026-04-28T04:45:01Z | 2026-04-26T00:27:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | adapepper | 259,097,501 | U_kgDOD3GDnQ | User | false |
openclaw/openclaw | 3,903,680,653 | I_kwDOQb6kR87orXCN | 9,908 | https://github.com/openclaw/openclaw/issues/9908 | https://api.github.com/repos/openclaw/openclaw/issues/9908 | feat: Add /skill list command to show available skills | ## Summary
When using `/skill` without a name, or `/skill list`, it should show available skills.
## Current Behavior
`/skill <name> [input]` runs a skill, but there's no way to list available skills via command.
**Current code** (`src/auto-reply/skill-commands.ts:79-82`):
```typescript
if (commandName === "skill") {... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-05T21:30:11Z | 2026-04-28T04:45:00Z | 2026-04-26T00:28:00Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | RonanCodes | 6,930,753 | MDQ6VXNlcjY5MzA3NTM= | User | false |
openclaw/openclaw | 3,903,975,103 | I_kwDOQb6kR87ose6_ | 9,959 | https://github.com/openclaw/openclaw/issues/9959 | https://api.github.com/repos/openclaw/openclaw/issues/9959 | Feature: Session list and switch from chat interface (Telegram/Discord/WhatsApp) | ## Summary
Add `/sessions` and `/session <key>` slash commands to list and switch between sessions directly from chat interfaces (Telegram, Discord, WhatsApp, etc.).
## Motivation
Currently, session management from chat is limited:
- `/new` creates a fresh session
- `/status` shows the current session
But there is ... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-05T22:58:52Z | 2026-04-28T04:44:59Z | 2026-04-26T00:28:05Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | echo931 | 259,437,483 | U_kgDOD3azqw | User | false |
openclaw/openclaw | 3,905,412,342 | I_kwDOQb6kR87ox9z2 | 10,240 | https://github.com/openclaw/openclaw/issues/10240 | https://api.github.com/repos/openclaw/openclaw/issues/10240 | Signal: Add REST API support alongside JSON-RPC for containerized deployments | ## Summary
The current Signal channel implementation works well with directly-installed `signal-cli` using JSON-RPC and SSE endpoints (`/api/v1/rpc` and `/api/v1/events`). However, these endpoints **do not exist in the official `bbernhard/signal-cli-rest-api` Docker image**, which only provides REST + WebSocket endpoi... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-02-06T07:43:47Z | 2026-04-28T04:44:58Z | 2026-04-26T00:28:20Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Hua688 | 5,764,877 | MDQ6VXNlcjU3NjQ4Nzc= | User | false |
openclaw/openclaw | 3,905,425,272 | I_kwDOQb6kR87oyA94 | 10,242 | https://github.com/openclaw/openclaw/issues/10242 | https://api.github.com/repos/openclaw/openclaw/issues/10242 | [Feature Request] Restore "New Thread" capability for Feishu (Lark) Channel in DMs | ### Description
In the latest version of the Feishu plugin (2026.2.3-1), the ability to start a new conversation thread in a 1:1 Direct Message (DM) has been lost. This feature existed in previous versions and is essential for organizing different automated tasks.
### Current Behavior
- All bot-sent messages are force... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-06T07:47:26Z | 2026-04-28T04:44:58Z | 2026-04-26T00:28:24Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Zebra-GPT | 147,117,044 | U_kgDOCMTT9A | User | false |
openclaw/openclaw | 3,906,822,872 | I_kwDOQb6kR87o3WLY | 10,419 | https://github.com/openclaw/openclaw/issues/10419 | https://api.github.com/repos/openclaw/openclaw/issues/10419 | [Feature]: Integrate EU Inference: OVHcloud AI Endpoints | ## Summary
Openclaw is an awesome tool that supports a lot of Inference Provider. Unfortunately, it only has US-based Inference Provider to power it. We would like to suggest a new inference provider, that is an alternative to give users access to sovereign LLMs, with data privacy (ZDR) and GDPR compliance.
## Propos... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-06T13:39:59Z | 2026-04-28T04:44:57Z | 2026-04-26T00:28:32Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | eliasto | 55,650,958 | MDQ6VXNlcjU1NjUwOTU4 | User | false |
openclaw/openclaw | 3,907,109,741 | I_kwDOQb6kR87o4cNt | 10,458 | https://github.com/openclaw/openclaw/issues/10458 | https://api.github.com/repos/openclaw/openclaw/issues/10458 | Docs: README uses deprecated --to for openclaw message send (should be --target) | ### Summary
README currently shows `openclaw message send --to ...`, but current CLI requires `--target` (`-t`).
### Repro
From latest CLI:
```bash
openclaw message send --help
```
Shows:
```text
-t, --target <dest>
```
README example still shows:
```bash
openclaw message send --to +1234567890 --message "Hello fr... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-06T14:46:05Z | 2026-04-28T04:44:56Z | 2026-04-26T00:28:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | quqi1599 | 51,637,638 | MDQ6VXNlcjUxNjM3NjM4 | User | false |
openclaw/openclaw | 3,907,257,233 | I_kwDOQb6kR87o5AOR | 10,485 | https://github.com/openclaw/openclaw/issues/10485 | https://api.github.com/repos/openclaw/openclaw/issues/10485 | Feature: Strip markdown formatting before TTS conversion | ## Problem
When TTS is enabled, markdown formatting characters (asterisks, underscores, backticks, headers) get read aloud literally. "Here is the **bold** text" becomes "Here is the asterisk asterisk bold asterisk asterisk text" which is confusing to listen to.
## Proposed Solution
Add a config option like `message... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-06T15:24:27Z | 2026-04-28T04:44:52Z | 2026-04-26T00:29:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | heybeaux | 174,768,117 | U_kgDOCmq_9Q | User | false |
openclaw/openclaw | 3,909,726,380 | I_kwDOQb6kR87pCbCs | 10,981 | https://github.com/openclaw/openclaw/issues/10981 | https://api.github.com/repos/openclaw/openclaw/issues/10981 | Feature: Programmatic session compact/reset for agent-initiated session management | ## Summary
Agents cannot programmatically compact or reset sessions. Sending `/compact` or `/reset` via the message tool sends them as text rather than executing the command.
## Use Case
COSTOP (cost optimization) workflows need to enforce session hygiene during heartbeats:
- Detect sessions over token threshold (e.... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-02-07T07:50:58Z | 2026-04-28T04:44:52Z | 2026-04-26T00:29:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | duckshrug | 259,036,244 | U_kgDOD3CUVA | User | false |
openclaw/openclaw | 3,909,818,887 | I_kwDOQb6kR87pCxoH | 11,015 | https://github.com/openclaw/openclaw/issues/11015 | https://api.github.com/repos/openclaw/openclaw/issues/11015 | Defensive acquisition of openclaw.co.jp & Japanese localization project | I am writing to report that I have defensively registered the domain **`openclaw.co.jp`** to prevent squatting by bad actors.
### Context
The situation around OpenClaw in Japan is becoming chaotic with misinformation. Also, many Japanese users struggle with English, which is a barrier to adoption.
### Proposal & Ques... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-07T08:49:52Z | 2026-04-28T04:44:51Z | 2026-04-26T00:29:53Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | cro555 | 260,028,807 | U_kgDOD3-5hw | User | false |
openclaw/openclaw | 3,909,831,040 | I_kwDOQb6kR87pC0mA | 11,018 | https://github.com/openclaw/openclaw/issues/11018 | https://api.github.com/repos/openclaw/openclaw/issues/11018 | Feature Request: Native event-driven notification system for proactive agent communication | ## Summary / 概要
OpenClaw currently lacks a native mechanism for agents to **proactively notify users when events occur**. All existing approaches are poll-based (cron, heartbeat), which introduces latency and wastes tokens when nothing has changed.
OpenClaw 目前缺少一种原生机制让 Agent **在事件发生时主动通知用户**。所有现有方法都是基于轮询的（cron、heartb... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-07T08:57:35Z | 2026-04-28T04:44:50Z | 2026-04-26T00:29:57Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Zjianru | 20,548,589 | MDQ6VXNlcjIwNTQ4NTg5 | User | false |
openclaw/openclaw | 3,911,827,225 | I_kwDOQb6kR87pKb8Z | 11,634 | https://github.com/openclaw/openclaw/issues/11634 | https://api.github.com/repos/openclaw/openclaw/issues/11634 | Feature: Add xAI OAuth provider auth flow (currently API-key only) | ## Summary
OpenClaw currently supports xAI via API key (`xai-api-key` / `XAI_API_KEY`), but there is no provider OAuth login flow for xAI in `openclaw models auth login`.
## Why this matters
- OAuth-style provider auth gives users a more consistent onboarding UX across providers.
- Some users prefer not to manually ma... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-08T03:03:19Z | 2026-04-28T04:44:49Z | 2026-04-26T00:30:21Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | quqi1599 | 51,637,638 | MDQ6VXNlcjUxNjM3NjM4 | User | false |
openclaw/openclaw | 3,911,945,760 | I_kwDOQb6kR87pK44g | 11,669 | https://github.com/openclaw/openclaw/issues/11669 | https://api.github.com/repos/openclaw/openclaw/issues/11669 | [Feature]: TUI: Option to preserve intermediate streaming output during multi-tool responses | NOTE: The following was written by my OpenClaw agent, but it succinctly describes the problem.
========================================================
When the assistant produces a multi-stage response (text → tool calls → more text → more tool calls → final text), the TUI's "streaming responses update in place unti... | closed | not_planned | true | 1 | [
"enhancement",
"tui"
] | [] | 2026-02-08T04:14:55Z | 2026-04-28T04:44:49Z | 2026-04-26T00:30:28Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Professa-Dementia | 87,164,087 | MDQ6VXNlcjg3MTY0MDg3 | User | false |
openclaw/openclaw | 3,911,974,332 | I_kwDOQb6kR87pK_28 | 11,677 | https://github.com/openclaw/openclaw/issues/11677 | https://api.github.com/repos/openclaw/openclaw/issues/11677 | [Feature]: Preface slack responses with model emoji | Feature: Model Response Prefixes
Description: Automatically prefix responses with model indicators per channel
Config Example:
channels:
slack:
modelPrefixes:
"anthropic/claude-sonnet-4-20250514": ":anthropic-sonnet:"
"google/gemini-2.0-flash": ":google-flash:"
... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-08T04:33:41Z | 2026-04-28T04:44:48Z | 2026-04-26T00:30:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | dajevu | 2,199,567 | MDQ6VXNlcjIxOTk1Njc= | User | false |
openclaw/openclaw | 3,913,687,711 | I_kwDOQb6kR87pRiKf | 12,135 | https://github.com/openclaw/openclaw/issues/12135 | https://api.github.com/repos/openclaw/openclaw/issues/12135 | Feature: Per-Provider Rate Limiting for Free-Tier API Support | ## Problem
Free-tier AI providers (NVIDIA NIM, potentially others) have rate limits that cause failures when agents make rapid tool calls:
**Example failure:**
- DeepSeek on NVIDIA NIM: 40 requests/min limit
- Agent made 8 tool calls in 30 seconds
- Hit 429 "Too Many Requests" error
- Task failed without completing c... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-08T22:16:17Z | 2026-04-28T04:44:47Z | 2026-04-26T00:30:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | arl1984 | 18,625,690 | MDQ6VXNlcjE4NjI1Njkw | User | false |
openclaw/openclaw | 3,913,995,469 | I_kwDOQb6kR87pStTN | 12,215 | https://github.com/openclaw/openclaw/issues/12215 | https://api.github.com/repos/openclaw/openclaw/issues/12215 | [Feature]: Add a "Model Ping" or "Health Check" feature for all models | ## Summary
When using free or cheap models with openclaw, the application often fails to respond without clear feedback about why. There's no visibility into whether the model is responding, how long it takes, or if the request failed entirely. This makes it difficult to diagnose issues and choose reliable models.
##... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-09T01:17:18Z | 2026-04-28T04:44:47Z | 2026-04-26T00:30:55Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | guimatheus92 | 11,640,754 | MDQ6VXNlcjExNjQwNzU0 | User | false |
openclaw/openclaw | 3,914,138,771 | I_kwDOQb6kR87pTQST | 12,252 | https://github.com/openclaw/openclaw/issues/12252 | https://api.github.com/repos/openclaw/openclaw/issues/12252 | [Feature]: Feature request: we might need a better interaction way from http/https website？ | The current webpage is somehow internal and develop-only, not something that can host to public network.
though I know we have somehow way for mattermost connection it be heavy..
IMO it be good time to implement a set of our own public website toolchains？...including
1. user system.
username，password，email，avatar.
mult... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-09T02:29:18Z | 2026-04-28T04:44:46Z | 2026-04-26T00:31:00Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | XenoAmess | 17,455,337 | MDQ6VXNlcjE3NDU1MzM3 | User | false |
openclaw/openclaw | 3,914,215,182 | I_kwDOQb6kR87pTi8O | 12,297 | https://github.com/openclaw/openclaw/issues/12297 | https://api.github.com/repos/openclaw/openclaw/issues/12297 | Feature Request: sessions_kill and sessions_cleanup tools | ## Problem
When cron jobs or sub-agents spawn isolated sessions that fail, they leave behind "zombie" sessions that:
- Consume memory/context tokens
- Clutter the session list
- Cannot be cleaned up by the agent that created them
## Current Gap
**Existing tools:**
- `sessions_list` — View sessions
- `sessions_send... | closed | not_planned | true | 5 | [
"enhancement"
] | [] | 2026-02-09T03:08:15Z | 2026-04-28T04:44:45Z | 2026-04-26T00:33:45Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | dTardis42 | 117,468,349 | U_kgDOBwBsvQ | User | false |
openclaw/openclaw | 3,914,393,786 | I_kwDOQb6kR87pUOi6 | 12,333 | https://github.com/openclaw/openclaw/issues/12333 | https://api.github.com/repos/openclaw/openclaw/issues/12333 | Project/Context Switching within Workspace | ## Problem
Currently, OpenClaw uses a single workspace directory (`~/.openclaw/workspace`) for all agent context. Developers working on multiple projects need to manually manage context or use separate agent instances.
## Proposed Solution
Add the ability to switch between project contexts within the workspace, acce... | closed | not_planned | true | 6 | [
"enhancement"
] | [] | 2026-02-09T04:33:17Z | 2026-04-28T04:44:45Z | 2026-04-26T00:33:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | shreyanshjain7174 | 43,677,304 | MDQ6VXNlcjQzNjc3MzA0 | User | false |
openclaw/openclaw | 3,914,820,801 | I_kwDOQb6kR87pV2zB | 12,412 | https://github.com/openclaw/openclaw/issues/12412 | https://api.github.com/repos/openclaw/openclaw/issues/12412 | Proposal: Native Token Optimization Dashboard & Pre-configured Optimization Tiers | ## Background\nCurrently, users need to manually configure various optimization settings (context pruning, compaction, local memory search) to manage token costs. While powerful, these settings are scattered.\n\n## Proposal\n1. **Native Optimization Dashboard**: Integrate a visual dashboard (similar to [oneles/openclaw... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-09T07:05:07Z | 2026-04-28T04:44:44Z | 2026-04-26T00:33:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | oneles | 62,010,096 | MDQ6VXNlcjYyMDEwMDk2 | User | false |
openclaw/openclaw | 3,915,934,071 | I_kwDOQb6kR87paGl3 | 12,617 | https://github.com/openclaw/openclaw/issues/12617 | https://api.github.com/repos/openclaw/openclaw/issues/12617 | Feature: Add model output context to before_tool_call hook | ## Summary
The `before_tool_call` plugin hook currently receives `toolName` and `params`, but not the model's output that preceded the tool call. This limits the ability to implement compliance patterns.
## Use Case
I want to enforce a behavioral pattern where the model must output a classification statement (e.g., ... | closed | not_planned | true | 5 | [
"enhancement"
] | [] | 2026-02-09T11:35:42Z | 2026-04-28T04:44:43Z | 2026-04-26T00:34:18Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | duckshrug | 259,036,244 | U_kgDOD3CUVA | User | false |
openclaw/openclaw | 3,919,139,158 | I_kwDOQb6kR87pmVFW | 13,054 | https://github.com/openclaw/openclaw/issues/13054 | https://api.github.com/repos/openclaw/openclaw/issues/13054 | Feature: Agent/model selector dropdown in webchat UI | ## Problem
When running multiple agents (e.g., main + circuit + other workspaces), the webchat interface has no way to switch between agents or models. Users have to manually edit openclaw.json config or use CLI commands. This has caused repeated config mistakes and broken sessions in our setup.
## Proposed Solution
... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-10T02:42:20Z | 2026-04-28T04:44:43Z | 2026-04-26T00:34:34Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | StarkTechIndustries | 67,027,658 | MDQ6VXNlcjY3MDI3NjU4 | User | false |
openclaw/openclaw | 3,919,418,660 | I_kwDOQb6kR87pnZUk | 13,131 | https://github.com/openclaw/openclaw/issues/13131 | https://api.github.com/repos/openclaw/openclaw/issues/13131 | Feature: Gateway-level audit logging for all tool calls | ## Problem
OpenClaw agents can self-log tool calls, but this has major gaps:
- **Cron jobs, sub-agents, and heartbeat sessions** run in isolated contexts and don't share the main session's logging
- Self-reporting is honor-system — the agent could skip entries
- No centralized view of all tool invocations across all s... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-10T04:44:36Z | 2026-04-28T04:44:42Z | 2026-04-26T00:34:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | isurfbecause | 1,476,109 | MDQ6VXNlcjE0NzYxMDk= | User | false |
openclaw/openclaw | 3,922,448,544 | I_kwDOQb6kR87py9Cg | 13,517 | https://github.com/openclaw/openclaw/issues/13517 | https://api.github.com/repos/openclaw/openclaw/issues/13517 | Feature Request: Option to exclude images from Slack message history context | ## Summary
When OpenClaw fetches past Slack messages for context, images are included and accumulate in the session. This can cause errors when the total image size exceeds the LLM's limit (e.g., Anthropic's 5MB per image limit).
## Problem
- Error: `LLM request rejected: messages.X.content.Y.image.source.base64: imag... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T16:37:16Z | 2026-04-28T04:44:40Z | 2026-04-26T00:35:08Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | watari-ai | 258,264,784 | U_kgDOD2TO0A | User | false |
openclaw/openclaw | 3,922,628,302 | I_kwDOQb6kR87pzo7O | 13,535 | https://github.com/openclaw/openclaw/issues/13535 | https://api.github.com/repos/openclaw/openclaw/issues/13535 | Feature: Project-based agent shortcuts in dashboard sidebar | ## Feature Request
**Problem:**
When working on multiple projects, switching context requires the agent to reload project memory each time. There's no way to have persistent project-specific sessions.
**Proposed Solution:**
Add a "Projects" or "Agents" section to the Gateway Dashboard sidebar that allows:
1. **Named... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-10T17:16:53Z | 2026-04-28T04:44:39Z | 2026-04-26T00:35:12Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ibechuckp | 108,746,357 | U_kgDOBntWdQ | User | false |
openclaw/openclaw | 3,923,120,914 | I_kwDOQb6kR87p1hMS | 13,621 | https://github.com/openclaw/openclaw/issues/13621 | https://api.github.com/repos/openclaw/openclaw/issues/13621 | [Feature]: Browser relay cannot interact with cross-origin iframe content (e.g., Salesforce Embedded Messaging) | ## Description
The browser relay (Chrome extension) cannot interact with elements inside cross-origin iframes. The `snapshot` and `act` browser tools only see the parent page's DOM — elements within cross-origin iframes are invisible.
## Steps to Reproduce
1. Attach the browser relay to a tab containing a cross-orig... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-10T19:13:05Z | 2026-04-28T04:44:38Z | 2026-04-26T00:35:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | litlmike | 41,237,956 | MDQ6VXNlcjQxMjM3OTU2 | User | false |
openclaw/openclaw | 4,329,636,801 | I_kwDOQb6kR88AAAABAhEDwQ | 71,847 | https://github.com/openclaw/openclaw/issues/71847 | https://api.github.com/repos/openclaw/openclaw/issues/71847 | [Bug]: vLLM/Nemotron Discord agent leaks metadata and drops thinking-only output | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
A vLLM/Nemotron-backed Discord agent can drop thinking-only responses and later leak OpenClaw inbound metadata/envelopes into visible replies, poisoning session history.
### Steps to reproduce
1. Configure OpenClaw 2026.4.... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-26T00:35:55Z | 2026-04-28T04:39:10Z | 2026-04-26T03:21:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jmystaki-create | 267,217,950 | U_kgDOD-1sHg | User | false |
openclaw/openclaw | 3,924,509,154 | I_kwDOQb6kR87p60Hi | 13,844 | https://github.com/openclaw/openclaw/issues/13844 | https://api.github.com/repos/openclaw/openclaw/issues/13844 | Feature request: inputFilter / ignorePatterns for channel messages | ## Problem
When clearing the Telegram chat window, the app automatically sends `/start` to the bot. This triggers a full agent turn (token cost + unnecessary response) even though it's not a real user message.
There's currently no way to configure the gateway to silently drop or ignore specific message patterns befor... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-11T03:07:02Z | 2026-04-28T04:44:37Z | 2026-04-26T00:37:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | firstmovr | 155,568,852 | U_kgDOCUXK1A | User | false |
openclaw/openclaw | 3,926,390,287 | I_kwDOQb6kR87qB_YP | 14,076 | https://github.com/openclaw/openclaw/issues/14076 | https://api.github.com/repos/openclaw/openclaw/issues/14076 | Feature request: read-only allowlist / suppress auto-reply for WhatsApp DMs | ## Problem
When using WhatsApp with `dmPolicy: "allowlist"`, adding a phone number to `allowFrom` makes that contact a **full conversational partner**. Every agent reply in the session is automatically sent back to the WhatsApp sender via the `web-auto-reply` module, even when the agent is primarily interacting on a d... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-11T12:50:20Z | 2026-04-28T04:44:36Z | 2026-04-26T00:37:51Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | pycckuu | 1,489,583 | MDQ6VXNlcjE0ODk1ODM= | User | false |
openclaw/openclaw | 3,928,662,569 | I_kwDOQb6kR87qKqIp | 14,264 | https://github.com/openclaw/openclaw/issues/14264 | https://api.github.com/repos/openclaw/openclaw/issues/14264 | [Feature Request] Add clawhub uninstall command for clean skill removal | ## Feature Request: Skill Uninstall Command
### Problem
Currently, there is no official way to uninstall/remove a skill via the ClawHub CLI. Users must manually delete skill directories from the workspace, which is:
- Error-prone (risk of deleting wrong files)
- Unclean (may leave orphaned configs or dependencies)
- I... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-11T20:56:04Z | 2026-04-28T04:44:36Z | 2026-04-26T00:38:06Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Donmeusi | 109,555,289 | U_kgDOBoeuWQ | User | false |
openclaw/openclaw | 3,929,246,533 | I_kwDOQb6kR87qM4tF | 14,341 | https://github.com/openclaw/openclaw/issues/14341 | https://api.github.com/repos/openclaw/openclaw/issues/14341 | [Feature]: Currently, the Lark plugin does not support sending images or voice messages | ## Summary
The current Feishu plugin only supports sending plain text messages and cannot handle rich media content such as images, audio, etc. When users attempt to send non-text messages (e.g., images/audio), the plugin returns vague error messages (e.g., "Unsupported message type") or HTTP 400 errors, leading to:
- ... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-12T00:20:57Z | 2026-04-28T04:44:35Z | 2026-04-26T00:38:11Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | sevenclockseven | 49,014,592 | MDQ6VXNlcjQ5MDE0NTky | User | false |
openclaw/openclaw | 3,929,468,646 | I_kwDOQb6kR87qNu7m | 14,389 | https://github.com/openclaw/openclaw/issues/14389 | https://api.github.com/repos/openclaw/openclaw/issues/14389 | Feature Request: Auto-rollback on config change failure | ## Problem\n\nWhen modifying the OpenClaw configuration (via `gateway config.patch` or `openclaw config`), there's no safety net if the gateway fails to restart. This leaves the system in a broken state, requiring manual rollback from a backup.\n\n## Proposed Solution\n\nImplement an **auto-rollback mechanism** for con... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-12T01:59:08Z | 2026-04-28T04:44:34Z | 2026-04-26T00:38:18Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | chily | 5,722,662 | MDQ6VXNlcjU3MjI2NjI= | User | false |
openclaw/openclaw | 3,938,379,210 | I_kwDOQb6kR87qvuXK | 15,647 | https://github.com/openclaw/openclaw/issues/15647 | https://api.github.com/repos/openclaw/openclaw/issues/15647 | [Feature]: Add support for Anthropic's native web search tool (Claude API) | ## Problem / Motivation
OpenClaw currently uses external providers (Brave, Perplexity) for web search via the `web_search` tool. Anthropic introduced native web search functionality for Claude models in May 2025, which is directly available through their API without requiring external services.
## Proposed Solution
... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-13T18:04:11Z | 2026-04-28T04:44:33Z | 2026-04-26T00:38:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | hobbesd | 28,409,605 | MDQ6VXNlcjI4NDA5NjA1 | User | false |
openclaw/openclaw | 3,939,384,804 | I_kwDOQb6kR87qzj3k | 15,779 | https://github.com/openclaw/openclaw/issues/15779 | https://api.github.com/repos/openclaw/openclaw/issues/15779 | [Feature]: Easy and automatically loads the models available from a local source i am running | ## Summary
There are many many many open source solutions that dynamically load in the models that are available form homeassistant hacs, to Open Webui.
## Proposed solution
What would you like Clawdbot to do?
Make the loading of models, agents and themes less complicated.
## Alternatives considered
Honestly a mu... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-13T21:42:33Z | 2026-04-28T04:44:33Z | 2026-04-26T00:38:53Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | cdvankammen | 31,197,020 | MDQ6VXNlcjMxMTk3MDIw | User | false |
openclaw/openclaw | 3,941,115,121 | I_kwDOQb6kR87q6KTx | 16,159 | https://github.com/openclaw/openclaw/issues/16159 | https://api.github.com/repos/openclaw/openclaw/issues/16159 | [Feature]: Add coordinate-based click support to browser tool | # Feature Request: Add coordinate-based click support to browser tool
## Problem / Use Case
Currently, the `browser` tool can only interact with HTML elements via selectors or text matching. This works great for standard web pages, but fails for:
- **Canvas-based applications** (games, drawing apps, remote desktop v... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-14T10:54:25Z | 2026-04-28T04:44:32Z | 2026-04-26T00:39:04Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | VAQYBIN | 63,801,831 | MDQ6VXNlcjYzODAxODMx | User | false |
openclaw/openclaw | 3,943,380,296 | I_kwDOQb6kR87rCzVI | 16,910 | https://github.com/openclaw/openclaw/issues/16910 | https://api.github.com/repos/openclaw/openclaw/issues/16910 | [Feature]: Request to allow openclaw use cursor credits | ### Summary
I was just wondering if there are any plans to allow openclaw use cursor credits directly when onboarding, that way cursor becomes a provider as well.
### Problem to solve
There are a lot of models in cursor and I could switch painlessly between models in cursor without having to pay for another provider... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-15T07:25:08Z | 2026-04-28T04:44:31Z | 2026-04-26T00:39:20Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Jraykekreate | 81,491,677 | MDQ6VXNlcjgxNDkxNjc3 | User | false |
openclaw/openclaw | 3,944,423,573 | I_kwDOQb6kR87rGyCV | 17,298 | https://github.com/openclaw/openclaw/issues/17298 | https://api.github.com/repos/openclaw/openclaw/issues/17298 | Control UI should display configured agent avatars instead of default favicon | ### Summary
The Control UI agent list sidebar should render each agent's configured `identity.avatar` instead of using the default OpenClaw favicon for all agents.
### Problem to solve
When running a multi-agent setup (e.g. 4 agents with distinct personas), each agent has a unique avatar configured via `identity.ava... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-15T16:32:02Z | 2026-04-28T04:44:30Z | 2026-04-26T00:39:31Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | C3PO-assistant | 261,654,174 | U_kgDOD5iGng | User | false |
openclaw/openclaw | 4,329,643,796 | I_kwDOQb6kR88AAAABAhEfFA | 71,848 | https://github.com/openclaw/openclaw/issues/71848 | https://api.github.com/repos/openclaw/openclaw/issues/71848 | [SRE] Gateway SIGABRT (134) on macOS launchd (mini-lobby) | ### Description
On the `mini-lobby` machine (macOS Darwin), the OpenClaw Gateway fails to start when managed by `launchctl`, returning exit code **134 (SIGABRT)**.
### Investigation Findings
- **Foreground Success:** Running the exact same command in an interactive terminal works perfectly.
- **launchd Failure:** The ... | closed | completed | true | 2 | [] | [
"vincentkoc"
] | 2026-04-26T00:39:07Z | 2026-04-29T04:32:40Z | 2026-04-27T02:04:08Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | palladius | 132,802 | MDQ6VXNlcjEzMjgwMg== | User | false |
openclaw/openclaw | 3,949,088,243 | I_kwDOQb6kR87rYk3z | 18,469 | https://github.com/openclaw/openclaw/issues/18469 | https://api.github.com/repos/openclaw/openclaw/issues/18469 | Auto-switch model based on content type | ### Summary
Automatically switch to a multimodal model (e.g., kimi-k2.5:cloud) when receiving image/video content, and switch back to the default model for text-only messages.
### Problem to solve
When users send images or videos, the assistant should automatically use a vision-capable model to process them, then re... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-16T18:39:13Z | 2026-04-28T04:44:28Z | 2026-04-26T00:43:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Souzix76 | 241,643,231 | U_kgDODmcu3w | User | false |
openclaw/openclaw | 3,949,626,095 | I_kwDOQb6kR87raoLv | 18,579 | https://github.com/openclaw/openclaw/issues/18579 | https://api.github.com/repos/openclaw/openclaw/issues/18579 | [Feature]: Feature Proposal: Native Google Cloud TTS Provider Integration | ### Summary
Native Google Cloud Text-to-Speech (TTS) provider integration with WaveNet and Polyglot support.
### Problem to solve
Currently, OpenClaw lacks native support for Google Cloud TTS, limiting users to OpenAI (expensive), ElevenLabs (very expensive), or Edge (variable quality). This integration adds a high-... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-16T21:23:31Z | 2026-04-28T04:44:27Z | 2026-04-26T00:43:46Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | 7SoKKoS7 | 97,355,065 | U_kgDOBc2FOQ | User | false |
openclaw/openclaw | 3,950,344,786 | I_kwDOQb6kR87rdXpS | 18,769 | https://github.com/openclaw/openclaw/issues/18769 | https://api.github.com/repos/openclaw/openclaw/issues/18769 | [Feature]: Mejoras generales y evoluciones estructurales (más allá de seguridad básica) | ### Summary
Implementar mejoras transversales para hacer OpenClaw más eficiente, sostenible y fácil de extender/adoptar.
### Problem to solve
Si bien OpenClaw ya cubre seguridad de prompt, tokens y privilegios, existen aspectos funcionales, de rendimiento, experiencia de usuario y adopción que pueden potenciarse.
#... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-17T01:53:49Z | 2026-04-28T04:44:27Z | 2026-04-26T00:43:51Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | 7damnatio | 131,477,060 | U_kgDOB9YuRA | User | false |
openclaw/openclaw | 3,953,774,306 | I_kwDOQb6kR87rqc7i | 19,301 | https://github.com/openclaw/openclaw/issues/19301 | https://api.github.com/repos/openclaw/openclaw/issues/19301 | [Feature]: feat: Add gpt-5.3-codex to GitHub Copilot provider with xhigh reasoning support | ### Summary
feat: Add gpt-5.3-codex to GitHub Copilot provider with xhigh reasoning support
### Problem to solve
OpenClaw's Copilot provider does not support the new `gpt-5.3-codex` model. This limits user access to advanced reasoning capabilities (including `xhigh`)—users can't select the model via `/model`, nor co... | closed | completed | true | 5 | [
"enhancement"
] | [] | 2026-02-17T17:04:21Z | 2026-04-28T04:44:25Z | 2026-04-26T00:44:06Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | LatencyTDH | 6,758,754 | MDQ6VXNlcjY3NTg3NTQ= | User | false |
openclaw/openclaw | 3,871,094,414 | I_kwDOQb6kR87mvDaO | 4,021 | https://github.com/openclaw/openclaw/issues/4021 | https://api.github.com/repos/openclaw/openclaw/issues/4021 | feat: Comprehensive Azure Provider Support Roadmap | ## Summary
Azure offers a wide range of AI and cloud services that can enhance Moltbot. With PR #3675 adding Azure OpenAI model support, I'd like to discuss a roadmap for comprehensive Azure integration.
## Current Status
| Service | Status | PR/Issue |
|---------|--------|----------|
| Azure OpenAI (Models) | 🟡 In... | closed | completed | true | 6 | [
"enhancement"
] | [] | 2026-01-29T14:51:56Z | 2026-04-26T00:44:10Z | 2026-02-01T00:31:21Z | NONE | null | 20260426T030032Z | 2026-04-26T03:00:32Z | Li-Hongmin | 14,259,647 | MDQ6VXNlcjE0MjU5NjQ3 | User | false |
openclaw/openclaw | 3,953,982,697 | I_kwDOQb6kR87rrPzp | 19,330 | https://github.com/openclaw/openclaw/issues/19330 | https://api.github.com/repos/openclaw/openclaw/issues/19330 | [Feature]: Inbound media shared across all agents — no per-agent isolation | ### Summary
In multi-agent setups, all inbound media files (voice notes, images, documents) are saved to a single shared directory (~/.openclaw/media/inbound/). Any agent with filesystem access can read media intended for a different agent.
### Problem to solve
When a user sends media to Agent A, it's saved to the s... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-17T17:54:33Z | 2026-05-05T02:13:36Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | spdesai | 1,816,427 | MDQ6VXNlcjE4MTY0Mjc= | User | false |
openclaw/openclaw | 3,955,789,874 | I_kwDOQb6kR87ryJAy | 19,723 | https://github.com/openclaw/openclaw/issues/19723 | https://api.github.com/repos/openclaw/openclaw/issues/19723 | [Feature]: Add File Upload Support to feishu_drive Tool | ### Summary
Add `upload` action to the `feishu_drive` tool to enable uploading arbitrary files to Feishu/Lark cloud storage. This would dramatically improve system flexibility by allowing users to control content and format themselves, rather than being limited to Markdown-to-docx conversion.
### Problem to solve
##... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-18T03:53:46Z | 2026-04-28T04:44:23Z | 2026-04-26T00:44:18Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | binglius | 115,341,131 | U_kgDOBt_3Sw | User | false |
openclaw/openclaw | 3,955,866,172 | I_kwDOQb6kR87rybo8 | 19,751 | https://github.com/openclaw/openclaw/issues/19751 | https://api.github.com/repos/openclaw/openclaw/issues/19751 | [Feature]: Allow disabling of deliverWebReply | ### Summary
deliverWebReply automatically sends audio from tmp folder to whatsapp. If the audio is in .mp3 format you cannot listen to it in mobile whatsapp and an .ogg coversion is needed, but this causes two voice messages to appear in whatsapp, one is .mp3 sent by deliverWebReply and another is the ogg conversion d... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-18T04:24:13Z | 2026-04-28T04:44:23Z | 2026-04-26T00:44:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jyashi | 47,436,660 | MDQ6VXNlcjQ3NDM2NjYw | User | false |
openclaw/openclaw | 3,957,159,230 | I_kwDOQb6kR87r3XU- | 19,960 | https://github.com/openclaw/openclaw/issues/19960 | https://api.github.com/repos/openclaw/openclaw/issues/19960 | [Feature]: Webhook ingress: support GET handlers and request header access | ### Summary
Webhook ingress: support GET handlers and request header access for Meta/WhatsApp Cloud API integration
### Problem to solve
We're running an OpenClaw agent that receives incoming WhatsApp messages via Meta's Cloud API. We can't use the native /hooks/agent endpoint because of two missing capabilities:
1... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-18T10:38:39Z | 2026-04-28T04:44:22Z | 2026-04-26T00:44:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | angeliti999 | 104,985,926 | U_kgDOBkH1Rg | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.