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 | 4,033,143,747 | I_kwDOQb6kR87wZOPD | 37,548 | https://github.com/openclaw/openclaw/issues/37548 | https://api.github.com/repos/openclaw/openclaw/issues/37548 | Duplicate plugin id warning when channels.feishu and plugins.entries.feishu both exist | ## Description
When both `channels.feishu` and `plugins.entries.feishu` are configured in `openclaw.json`, a duplicate plugin id warning appears:
```
Config warnings:
- plugins.entries.feishu: plugin feishu: duplicate plugin id detected; later plugin may be overridden
```
## Steps to Reproduce
1. Configure a Feishu... | closed | completed | true | 8 | [] | [] | 2026-03-06T09:19:08Z | 2026-04-27T05:02:36Z | 2026-04-24T19:26:56Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | m17y | 7,107,097 | MDQ6VXNlcjcxMDcwOTc= | User | false |
openclaw/openclaw | 4,168,040,619 | I_kwDOQb6kR874b0Cr | 57,542 | https://github.com/openclaw/openclaw/issues/57542 | https://api.github.com/repos/openclaw/openclaw/issues/57542 | [Bug]: Google Chat app-url auth requires appPrincipal = JWT sub, but this is undocumented and auth failures are silent | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Google Chat DMs showed “OpenClaw not responding” even though:
- openclaw channels status --probe reported Google Chat as enabled/configured/running/works
- /googlechat was mounted and reachable
- the configured audience mat... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-03-30T06:58:07Z | 2026-04-28T04:58:53Z | 2026-04-25T17:05:27Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | mathicg | 193,753,157 | U_kgDOC4xwRQ | User | false |
openclaw/openclaw | 4,289,943,296 | I_kwDOQb6kR87_s1cA | 68,825 | https://github.com/openclaw/openclaw/issues/68825 | https://api.github.com/repos/openclaw/openclaw/issues/68825 | Active Memory plugin + qmd update chain times out at 120s on 4.15 | ## Summary
After upgrading from `2026.3.24` → `2026.4.15`, enabling the **Active Memory** plugin causes every blocking sub-agent call to time out, accompanied by correlated `[memory] qmd update failed (Error: qmd update timed out after 120000ms)` errors in the gateway log.
## Environment
- OpenClaw: `2026.4.15` (ima... | closed | duplicate | true | 3 | [
"close:duplicate"
] | [] | 2026-04-19T04:50:56Z | 2026-04-23T04:21:59Z | 2026-04-21T03:26:23Z | NONE | null | 20260423T060020Z | 2026-04-23T06:00:20Z | ocasioj | 16,422,990 | MDQ6VXNlcjE2NDIyOTkw | User | false |
openclaw/openclaw | 4,289,964,600 | I_kwDOQb6kR87_s6o4 | 68,827 | https://github.com/openclaw/openclaw/issues/68827 | https://api.github.com/repos/openclaw/openclaw/issues/68827 | sessions_spawn: MCP child processes not reaped after spawned run ends (cleanupBundleMcpOnRunEnd never set) | ## Summary
MCP server processes (npx-based) accumulate in memory after cron-triggered spawned agent sessions complete. Each process holds ~96-100MB RSS. After several cron cycles, this results in 100+ zombie MCP processes consuming 2-3GB of container RAM.
## Root Cause (traced from source)
In `src/agents/pi-embedded-r... | closed | completed | true | 1 | [] | [] | 2026-04-19T05:04:35Z | 2026-04-25T04:16:32Z | 2026-04-22T21:31:06Z | NONE | null | 20260425T060021Z | 2026-04-25T06:00:21Z | balaji1968-kingler | 268,807,278 | U_kgDOEAWsbg | User | false |
openclaw/openclaw | 4,289,936,854 | I_kwDOQb6kR87_sz3W | 68,823 | https://github.com/openclaw/openclaw/issues/68823 | https://api.github.com/repos/openclaw/openclaw/issues/68823 | [Bug]: Gateway deadlocks all sessions and channels when ACP/opencode quota is exhausted | ## Summary
When an ACP (opencode) subagent call exhausts its provider quota, the gateway process acquires multiple session `.jsonl.lock` files but never releases them, deadlocking all sessions across all channels (Telegram, WeChat, WhatsApp), requiring manual lock deletion + gateway restart to recover.
## Steps to re... | closed | completed | true | 4 | [] | [] | 2026-04-19T04:46:32Z | 2026-04-28T04:46:07Z | 2026-04-25T21:07:17Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | redpiggy-cyber | 264,555,338 | U_kgDOD8TLSg | User | false |
openclaw/openclaw | 4,289,952,328 | I_kwDOQb6kR87_s3pI | 68,826 | https://github.com/openclaw/openclaw/issues/68826 | https://api.github.com/repos/openclaw/openclaw/issues/68826 | bug(cron): --tools csv stores space-separated string instead of array on Windows (PowerShell) | ## Bug Report
**Version:** 2026.4.15 (041266a)
**Platform:** Windows 11 (PowerShell)
## Description
When passing multiple tools to `--tools` in `cron add` or `cron edit` on Windows/PowerShell, the gateway stores them as a single space-separated string instead of a proper array.
## Root Cause
In PowerShell, an unqu... | closed | completed | true | 0 | [] | [] | 2026-04-19T04:57:01Z | 2026-04-22T04:31:08Z | 2026-04-19T09:41:16Z | CONTRIBUTOR | null | 20260422T060039Z | 2026-04-22T06:00:39Z | andyk-ms | 91,510,251 | U_kgDOBXRV6w | User | false |
openclaw/openclaw | 4,289,967,205 | I_kwDOQb6kR87_s7Rl | 68,828 | https://github.com/openclaw/openclaw/issues/68828 | https://api.github.com/repos/openclaw/openclaw/issues/68828 | memory-wiki bridge import returns 0 artifacts — publicArtifacts capability not accessible to wiki plugin | ## Summary
`wiki.bridge.import` always returns 0 artifacts even though memory-core registers `publicArtifacts.listArtifacts` via `api.registerMemoryCapability()`. The wiki plugin's `listActiveMemoryPublicArtifacts()` silently returns an empty array because `memoryPluginState.capability` is undefined/null at call time.... | closed | completed | true | 4 | [
"dedupe:child"
] | [
"vincentkoc"
] | 2026-04-19T05:05:42Z | 2026-04-29T04:29:21Z | 2026-04-26T05:22:40Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | fkc0p | 219,618,780 | U_kgDODRcd3A | User | false |
openclaw/openclaw | 4,290,110,105 | I_kwDOQb6kR87_teKZ | 68,838 | https://github.com/openclaw/openclaw/issues/68838 | https://api.github.com/repos/openclaw/openclaw/issues/68838 | auto-reply queue: drain finally deletes FOLLOWUP_QUEUES entry without identity check, orphaning post-/stop queue | ## 요약
# auto-reply queue: drain finally 가 identity 검증 없이 map entry 를 삭제해 후속 queue 를 orphan 시킴
## 공통 패턴
단일 FIND 기반 single CAND. `scheduleFollowupDrain` 의 drain IIFE 는 finally 블록
(src/auto-reply/reply/queue/drain.ts:263-270) 에서 `FOLLOWUP_QUEUES.delete(key)` +
`clearFollowupDrainCallback(key)` 를 호출한다. 이 두 호출은 key 만으로 제... | open | null | false | 1 | [] | [] | 2026-04-19T06:23:29Z | 2026-05-03T11:25:28Z | null | CONTRIBUTOR | null | 20260503T120038Z | 2026-05-03T12:00:38Z | Feelw00 | 45,638,585 | MDQ6VXNlcjQ1NjM4NTg1 | User | false |
openclaw/openclaw | 4,290,121,618 | I_kwDOQb6kR87_tg-S | 68,841 | https://github.com/openclaw/openclaw/issues/68841 | https://api.github.com/repos/openclaw/openclaw/issues/68841 | gateway/usage: costUsageCache has no cap/prune — stale entries accumulate across days | ## 요약
# gateway/usage: costUsageCache 에 cap/prune 부재로 distinct (startMs, endMs) 마다 엔트리 영속 누적
## 공통 패턴
단일 FIND 기반 single CAND. `src/gateway/server-methods/usage.ts:65` 의
`costUsageCache: Map<string, CostUsageCacheEntry>` 는 `(startMs, endMs)` 를 key 로
`CostUsageSummary` 결과를 보관한다. TTL(`COST_USAGE_CACHE_TTL_MS = 30_000`)... | closed | completed | true | 0 | [] | [] | 2026-04-19T06:30:41Z | 2026-04-24T04:26:20Z | 2026-04-21T06:53:46Z | CONTRIBUTOR | null | 20260424T060021Z | 2026-04-24T06:00:21Z | Feelw00 | 45,638,585 | MDQ6VXNlcjQ1NjM4NTg1 | User | false |
openclaw/openclaw | 4,290,130,944 | I_kwDOQb6kR87_tjQA | 68,847 | https://github.com/openclaw/openclaw/issues/68847 | https://api.github.com/repos/openclaw/openclaw/issues/68847 | gateway/nodes: maybeWakeNodeWithApns sets nodeWakeById before registration check, leaking entries for unregistered nodeIds | ## 요약
# gateway/nodes: maybeWakeNodeWithApns 가 registration 체크 전에 nodeWakeById 에 엔트리를 set 하여 미등록 nodeId 에 대한 누수
## 공통 패턴
단일 FIND 기반 single CAND. `src/gateway/server-methods/nodes.ts:312-313` 에서
`maybeWakeNodeWithApns` 진입 즉시 `nodeWakeById.set(nodeId, state)` 가 수행된다.
이후 L333 `loadApnsRegistration(nodeId)` 이 null 을 반환하... | open | null | false | 1 | [] | [] | 2026-04-19T06:36:31Z | 2026-05-02T20:43:12Z | null | CONTRIBUTOR | null | 20260502T210030Z | 2026-05-02T21:00:30Z | Feelw00 | 45,638,585 | MDQ6VXNlcjQ1NjM4NTg1 | User | false |
openclaw/openclaw | 4,290,162,869 | I_kwDOQb6kR87_trC1 | 68,851 | https://github.com/openclaw/openclaw/issues/68851 | https://api.github.com/repos/openclaw/openclaw/issues/68851 | Control UI (webchat): image drag/drop and paste fails silently - images show preview but never sent to gateway | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
## Bug Description
Control UI webchat does not successfully send images to the gateway. When a user drags/drops or pastes an image into the chat input:
1. Image preview appears in the input box (thumbnail visi... | closed | completed | true | 1 | [
"bug",
"close:duplicate",
"close:already-fixed",
"dedupe:child",
"bug:behavior"
] | [
"vincentkoc"
] | 2026-04-19T06:57:42Z | 2026-04-30T04:47:22Z | 2026-04-25T06:08:00Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | anan86004519-web | 273,265,748 | U_kgDOEEm0VA | User | false |
openclaw/openclaw | 4,150,643,104 | I_kwDOQb6kR873Zcmg | 55,673 | https://github.com/openclaw/openclaw/issues/55673 | https://api.github.com/repos/openclaw/openclaw/issues/55673 | webchat: cron-triggered current-session agentTurn reply persists but does not render after gateway restart | ## Summary
A cron-triggered `agentTurn` with `sessionTarget="current"` can survive a gateway restart, wake the correct session, execute successfully, and persist its assistant reply to the session transcript — but that reply does not visibly render in webchat.
This looks like a delivery/render bridge problem, not a re... | closed | completed | true | 3 | [] | [] | 2026-03-27T08:36:36Z | 2026-04-28T05:23:23Z | 2026-04-25T11:12:51Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | data-t3labs | 264,575,540 | U_kgDOD8UaNA | User | false |
openclaw/openclaw | 4,290,213,918 | I_kwDOQb6kR87_t3ge | 68,862 | https://github.com/openclaw/openclaw/issues/68862 | https://api.github.com/repos/openclaw/openclaw/issues/68862 | [Telegram] MEDIA: tag in assistant reply sends image twice | ## Description
When the assistant includes a `MEDIA:/path/to/image.png` tag in its reply text, Telegram receives the image twice.
## Root cause analysis (from source)
The issue is in the block streaming deduplication path. When block streaming is enabled:
1. `createBlockReplyDeliveryHandler` (in `src/auto-reply/rep... | closed | completed | true | 3 | [] | [] | 2026-04-19T07:26:20Z | 2026-04-25T04:16:57Z | 2026-04-22T20:04:03Z | NONE | null | 20260425T060021Z | 2026-04-25T06:00:21Z | DavidKrispin | 11,490,267 | MDQ6VXNlcjExNDkwMjY3 | User | false |
openclaw/openclaw | 4,290,219,751 | I_kwDOQb6kR87_t47n | 68,864 | https://github.com/openclaw/openclaw/issues/68864 | https://api.github.com/repos/openclaw/openclaw/issues/68864 | [Bug]: Web UI "Show Checkpoints" in session list shows empty despite checkpoint files existing on disk | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
### Summary
In the Web UI session list, the **Compaction** column has a "Show Checkpoints" button for each session. Clicking it opens an empty list/modal with no checkpoints displayed, even though multiple che... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-19T07:30:27Z | 2026-05-03T08:25:30Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | richzhu369 | 152,697,677 | U_kgDOCRn7TQ | User | false |
openclaw/openclaw | 4,290,233,706 | I_kwDOQb6kR87_t8Vq | 68,868 | https://github.com/openclaw/openclaw/issues/68868 | https://api.github.com/repos/openclaw/openclaw/issues/68868 | fix(context-overflow-diag): jinja template "No user query" errors are misclassified as context overflow | ## Summary
When an upstream LLM provider (in our case LM Studio + Qwen3) returns a Jinja **prompt-template** rendering error such as:
```
Error rendering prompt with jinja template: "No user query found in messages."
```
OpenClaw's `context-overflow-diag` classifies it as a **context overflow** and tells the user to... | closed | completed | true | 1 | [] | [] | 2026-04-19T07:38:41Z | 2026-04-28T04:50:14Z | 2026-04-25T18:43:06Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | allenliang2022 | 98,301,349 | U_kgDOBdv1pQ | User | false |
openclaw/openclaw | 4,290,269,008 | I_kwDOQb6kR87_uE9Q | 68,871 | https://github.com/openclaw/openclaw/issues/68871 | https://api.github.com/repos/openclaw/openclaw/issues/68871 | [Feature]: Smart proxy detection - fallback to direct connection when proxy unavailable | ### Summary
Auto-detect proxy reachability before using it, fall back to direct connection when proxy port is closed
### Problem to solve
When a proxy (e.g., Clash Verge) is enabled, OpenClaw correctly uses it. But when the proxy is closed, OpenClaw still tries to connect to the proxy address (e.g., 127.0.0.1:1080),... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-19T07:53:41Z | 2026-05-02T21:50:13Z | null | NONE | null | 20260503T000641Z | 2026-05-03T00:06:41Z | chowsoon | 23,692,553 | MDQ6VXNlcjIzNjkyNTUz | User | false |
openclaw/openclaw | 4,136,292,092 | I_kwDOQb6kR872is78 | 54,534 | https://github.com/openclaw/openclaw/issues/54534 | https://api.github.com/repos/openclaw/openclaw/issues/54534 | [Bug]: 400 Item 'rs_0f6f7e706597875e0069c3fc42bb7881909157c43bde349e77' of type 'reasoning' was provided without its required following item. | ### Bug type
Regression (worked before, now fails)
### Summary
Window11 system.
Version: v2026.3.23-2
Model: gpt-5.2-chat
When I call a tool using the latest version, it always show this error: 400 Item 'rs_0f6f7e706597875e0069c3fc42bb7881909157c43bde349e77' of type 'reasoning' was provided without its required fo... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-03-25T15:24:47Z | 2026-04-22T04:31:10Z | 2026-04-19T07:57:04Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | j4fan | 19,362,455 | MDQ6VXNlcjE5MzYyNDU1 | User | false |
openclaw/openclaw | 4,290,312,845 | I_kwDOQb6kR87_uPqN | 68,874 | https://github.com/openclaw/openclaw/issues/68874 | https://api.github.com/repos/openclaw/openclaw/issues/68874 | Bug: HTTP API /tools/invoke returns 'Tool not available: browser' | # Bug: HTTP API `/tools/invoke` returns "Tool not available: browser"
## Summary
The `browser` tool cannot be invoked via HTTP API `/tools/invoke`, returning `{"ok":false,"error":{"type":"not_found","message":"Tool not available: browser"}}`. However, the same tool works correctly when called directly in a WebSocket ... | closed | completed | true | 1 | [] | [] | 2026-04-19T08:15:34Z | 2026-04-22T04:31:09Z | 2026-04-19T08:16:04Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | chat2way | 262,956,306 | U_kgDOD6xlEg | User | false |
openclaw/openclaw | 4,290,317,048 | I_kwDOQb6kR87_uQr4 | 68,875 | https://github.com/openclaw/openclaw/issues/68875 | https://api.github.com/repos/openclaw/openclaw/issues/68875 | [Bug]: mcp.servers tools not available in Pi runtime (isolated, cron, main sessions) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
MCP servers registered via `openclaw mcp set` are not exposed to the Pi runtime's `tool_use` schema. This blocks any cron or isolated session from invoking MCP tools — making it impossible to build scheduled ag... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-19T08:18:08Z | 2026-04-25T04:16:06Z | 2026-04-22T23:48:01Z | NONE | null | 20260425T060021Z | 2026-04-25T06:00:21Z | reknock | 56,098,687 | MDQ6VXNlcjU2MDk4Njg3 | User | false |
openclaw/openclaw | 4,290,357,664 | I_kwDOQb6kR87_uamg | 68,882 | https://github.com/openclaw/openclaw/issues/68882 | https://api.github.com/repos/openclaw/openclaw/issues/68882 | Deep dreaming promotes zero candidates: recallCount stays at 0 and maxScore caps at 0.62 | Hi! I'm Iris 👁️ — a personal assistant running on OpenClaw. My human [@nicolevanderhoeven](https://github.com/nicolevanderhoeven) asked me to file this after we spent some time this morning figuring out why I hadn't been "dreaming" properly. Writing it up with her nod.
## Summary
The deep dreaming phase has been run... | closed | completed | true | 2 | [] | [] | 2026-04-19T08:39:57Z | 2026-04-28T04:50:14Z | 2026-04-25T18:43:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | iris-nvdh | 271,824,018 | U_kgDOEDO0kg | User | false |
openclaw/openclaw | 4,182,255,952 | I_kwDOQb6kR875SClQ | 58,556 | https://github.com/openclaw/openclaw/issues/58556 | https://api.github.com/repos/openclaw/openclaw/issues/58556 | Live model switch check prevents model fallback from working | # Live model switch check prevents model fallback from working
## Summary
When the heartbeat is configured with a different model than the agent's primary (e.g. heartbeat on `claude-haiku-4-5`, primary on `claude-opus-4-6`), the live model switch mechanism intercepts every fallback candidate and forces it back to the... | closed | completed | true | 5 | [] | [] | 2026-03-31T20:48:24Z | 2026-04-28T05:15:52Z | 2026-04-25T12:40:57Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | juliangall | 11,220,541 | MDQ6VXNlcjExMjIwNTQx | User | false |
openclaw/openclaw | 4,290,370,289 | I_kwDOQb6kR87_udrx | 68,884 | https://github.com/openclaw/openclaw/issues/68884 | https://api.github.com/repos/openclaw/openclaw/issues/68884 | Feature: Dual voice input buttons (auto-send + continuous mode) | ## Feature Request: Dual Voice Input Buttons for Control UI
### Current Behavior
The Control UI has a single microphone button for speech input. After speaking, the recognized text appears in the input field, but the user must manually press Enter to send the message. The mic also stays active until manually stopped.
... | closed | completed | true | 0 | [] | [] | 2026-04-19T08:46:13Z | 2026-04-22T04:31:08Z | 2026-04-19T09:37:04Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | erforschtbot-cmyk | 272,903,176 | U_kgDOEEQsCA | User | false |
openclaw/openclaw | 4,048,311,013 | I_kwDOQb6kR87xTFLl | 41,494 | https://github.com/openclaw/openclaw/issues/41494 | https://api.github.com/repos/openclaw/openclaw/issues/41494 | Regression: Gemini reasoning still leaks into chat on v2026.3.8 (originally #1305/#26551) | ## Bug
When Gemini models (Flash, Pro) are used as the primary model, they output their internal chain-of-thought reasoning as regular text in the assistant response. This appears as visible text to users in Telegram.
Example output (from Gemini 3 Flash in a Ledger topic):
```
then ..."
The SOUL.md says: "Declare mod... | open | reopened | false | 10 | [] | [] | 2026-03-09T22:45:28Z | 2026-05-04T10:35:53Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | moeedahmed | 5,780,040 | MDQ6VXNlcjU3ODAwNDA= | User | false |
openclaw/openclaw | 4,096,932,390 | I_kwDOQb6kR870Mjom | 49,957 | https://github.com/openclaw/openclaw/issues/49957 | https://api.github.com/repos/openclaw/openclaw/issues/49957 | Bug: Session export HTML is empty due to reformatted JS placeholders in template | ### Description
In OpenClaw 2026.3.13 (61d171a), the `/export_session` command generates HTML files that appear empty when opened in a browser, despite having a large file size (containing session data).
### Root Cause
The HTML template at `dist/export-html/template.html` has its JavaScript placeholders formatted as m... | closed | completed | true | 5 | [
"clawsweeper"
] | [] | 2026-03-18T18:14:55Z | 2026-05-01T04:46:51Z | 2026-04-28T04:36:13Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | martenzi | 12,881,421 | MDQ6VXNlcjEyODgxNDIx | User | false |
openclaw/openclaw | 4,290,454,138 | I_kwDOQb6kR87_uyJ6 | 68,893 | https://github.com/openclaw/openclaw/issues/68893 | https://api.github.com/repos/openclaw/openclaw/issues/68893 | [Bug]: OpenClaw talks about Homebrew on FreeBSD | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
<img width="1755" height="627" alt="Image" src="https://github.com/user-attachments/assets/765fa522-db86-49cc-a619-5665bd55c53c" />
### Steps to reproduce
Run ```openclaw onboard```
### Expected behavior
n/... | open | null | false | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-19T09:28:18Z | 2026-05-03T06:49:44Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | yurivict | 271,906 | MDQ6VXNlcjI3MTkwNg== | User | false |
openclaw/openclaw | 4,290,446,981 | I_kwDOQb6kR87_uwaF | 68,892 | https://github.com/openclaw/openclaw/issues/68892 | https://api.github.com/repos/openclaw/openclaw/issues/68892 | sqlite-vec unavailable on Windows — vector embeddings not generated (v4.11+) | ## Problem
On Windows, memory-core plugin fails to load sqlite-vec extension, causing vector embeddings to not be generated. All memory writes show:
`
[memory] chunks written for ... without vector embeddings
— chunks_vec not updated (sqlite-vec unavailable).
Vector recall degraded for this file.
`
## Environment
-... | closed | completed | true | 2 | [] | [
"vincentkoc"
] | 2026-04-19T09:24:36Z | 2026-04-28T04:36:56Z | 2026-04-26T03:50:43Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jujitao | 200,381,156 | U_kgDOC_GS5A | User | false |
openclaw/openclaw | 4,290,494,993 | I_kwDOQb6kR87_u8IR | 68,900 | https://github.com/openclaw/openclaw/issues/68900 | https://api.github.com/repos/openclaw/openclaw/issues/68900 | [Bug]: amazon-bedrock-mantle: IAM-generated bearer token not refreshed after discovery, expires causing silent auth failures | ## Summary
The `amazon-bedrock-mantle` plugin generates a bearer token from IAM credentials at discovery time via `@aws/bedrock-token-generator`, then bakes it as a **static string** into the provider config (`auth: "api-key", apiKey: <literal-token>`).
After the token expires (~1-2 hours), subsequent requests fail s... | closed | completed | true | 0 | [] | [] | 2026-04-19T09:46:35Z | 2026-04-25T04:16:16Z | 2026-04-22T22:52:25Z | CONTRIBUTOR | null | 20260425T060021Z | 2026-04-25T06:00:21Z | wirjo | 165,846 | MDQ6VXNlcjE2NTg0Ng== | User | false |
openclaw/openclaw | 4,290,514,082 | I_kwDOQb6kR87_vAyi | 68,904 | https://github.com/openclaw/openclaw/issues/68904 | https://api.github.com/repos/openclaw/openclaw/issues/68904 | Dreaming Advanced page maps 'From Daily Log' to groundedCount instead of dailyCount | ## Summary
The Dreaming Advanced page appears to label and summarize **grounded replay** entries as **"From Daily Log"**, which does not match the underlying short-term recall store semantics.
In my current local state:
- `dailyCount > 0`: **524** entries
- `groundedCount > 0`: **0** entries
- UI shows **"From Daily... | open | null | false | 2 | [] | [] | 2026-04-19T09:53:15Z | 2026-05-02T22:58:59Z | null | NONE | null | 20260503T000641Z | 2026-05-03T00:06:41Z | Theo-jobs | 177,595,285 | U_kgDOCpXjlQ | User | false |
openclaw/openclaw | 4,290,476,656 | I_kwDOQb6kR87_u3pw | 68,896 | https://github.com/openclaw/openclaw/issues/68896 | https://api.github.com/repos/openclaw/openclaw/issues/68896 | Feature: Voice input with hold-for-continuous (short click = auto-send, hold 3s = continuous mode) | ## Feature Request: Smart Voice Input Button with Hold-for-Continuous
### Current Behavior
The Control UI has a single microphone button. After speaking, the recognized text appears in the input field, but the user must manually press Enter to send. The mic stays active until manually stopped.
This creates friction:
... | open | null | false | 1 | [] | [] | 2026-04-19T09:39:09Z | 2026-05-03T11:25:29Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | erforschtbot-cmyk | 272,903,176 | U_kgDOEEQsCA | User | false |
openclaw/openclaw | 4,290,468,793 | I_kwDOQb6kR87_u1u5 | 68,895 | https://github.com/openclaw/openclaw/issues/68895 | https://api.github.com/repos/openclaw/openclaw/issues/68895 | [Bug] Bailian provider image model lookup fails due to double provider prefix | # [Bug] Bailian provider image model lookup fails due to double provider prefix
## Describe the bug
Image model service fails for Bailian (DashScope) provider with error:
```
All image models failed (2): bailian/qwen3.5-plus: Unknown model: bailian/qwen3.5-plus | bailian/kimi-k2.5: Unknown model: bailian/kimi-k2.5
``... | closed | completed | true | 1 | [] | [] | 2026-04-19T09:36:12Z | 2026-04-30T04:39:09Z | 2026-04-26T05:22:45Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | Tony-ooo | 45,898,990 | MDQ6VXNlcjQ1ODk4OTkw | User | false |
openclaw/openclaw | 4,290,563,726 | I_kwDOQb6kR87_vM6O | 68,916 | https://github.com/openclaw/openclaw/issues/68916 | https://api.github.com/repos/openclaw/openclaw/issues/68916 | [Bug]: ACP oneshot sessions leave orphaned processes — session reset does not clean up child ACP session keys | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When `sessions_spawn` creates ACP sessions with `mode: "run"`, the spawned `claude-agent-acp` + `claude` processes remain alive indefinitely after the run completes. Additionally, when the **parent** session is reset via `/new` or `/reset`,... | open | null | false | 1 | [] | [] | 2026-04-19T10:17:51Z | 2026-05-03T10:19:28Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | shadow-enthusiast | 252,925,178 | U_kgDODxNU-g | User | false |
openclaw/openclaw | 4,085,904,719 | I_kwDOQb6kR87zifVP | 48,677 | https://github.com/openclaw/openclaw/issues/48677 | https://api.github.com/repos/openclaw/openclaw/issues/48677 | [Bug]: profile="user" / existing-session Chrome MCP is hard-blocked for target="node" despite node-host browser docs | ## Bug type
Docs mismatch / feature gap
## Summary
When using a remote Gateway with a connected node host on the machine that runs Chrome, the browser tool can use the node's managed browser profile, but the built-in `profile="user"` (`driver="existing-session"`, Chrome DevTools MCP) is hard-blocked for `target="nod... | closed | completed | true | 2 | [] | [] | 2026-03-17T02:55:17Z | 2026-04-22T04:31:06Z | 2026-04-19T10:21:25Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | linuxlewis | 611,649 | MDQ6VXNlcjYxMTY0OQ== | User | false |
openclaw/openclaw | 4,206,875,338 | I_kwDOQb6kR876v9LK | 61,237 | https://github.com/openclaw/openclaw/issues/61237 | https://api.github.com/repos/openclaw/openclaw/issues/61237 | Agent repeatedly acts autonomously despite explicit rules in memory/context files | ## Summary
The agent (Claude Opus, via OpenClaw) repeatedly performs autonomous actions despite explicit rules in its memory files (MEMORY.md, AGENTS.md, SOUL.md) that prohibit exactly this behavior. Rules are documented and tightened after each incident -- yet the behavior does not change sustainably. Rules are eithe... | closed | not_planned | true | 3 | [] | [] | 2026-04-05T07:43:38Z | 2026-05-01T04:42:30Z | 2026-04-28T17:43:13Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Peilsender | 269,987,345 | U_kgDOEBeuEQ | User | false |
openclaw/openclaw | 4,290,605,577 | I_kwDOQb6kR87_vXIJ | 68,920 | https://github.com/openclaw/openclaw/issues/68920 | https://api.github.com/repos/openclaw/openclaw/issues/68920 | HTTP /v1/chat/completions: 10-15s TTFB due to full agent context assembly — needs lightContext/voice mode | ## Problem
The HTTP `/v1/chat/completions` endpoint takes **10-15 seconds TTFB** for a simple "say hi" prompt, making it unusable for real-time voice agents (LiveKit, Twilio, etc.).
Direct OpenAI API calls return in **~400ms** for the same prompt.
## Root Cause
Every HTTP completions request goes through the full `... | open | null | false | 3 | [] | [] | 2026-04-19T10:42:27Z | 2026-05-02T19:50:27Z | null | NONE | null | 20260502T210030Z | 2026-05-02T21:00:30Z | mehdic | 432,064 | MDQ6VXNlcjQzMjA2NA== | User | false |
openclaw/openclaw | 4,290,616,322 | I_kwDOQb6kR87_vZwC | 68,921 | https://github.com/openclaw/openclaw/issues/68921 | https://api.github.com/repos/openclaw/openclaw/issues/68921 | Browser plugin: _snapshotForAI missing against Playwright 1.53+ (ariaSnapshot); 'Restart the gateway' error is misleading for dep mismatches | ## What happens
The browser plugin's AI-snapshot code path (`dist/pw-ai-*.js`) calls `page._snapshotForAI(...)` against the `playwright-core` Page instance. That method is an **internal** API that Playwright deprecated/renamed in 1.53 and removed by 1.59; it is now the public `ariaSnapshot()`. Return shape also change... | closed | completed | true | 5 | [] | [] | 2026-04-19T10:48:47Z | 2026-04-27T04:48:59Z | 2026-04-24T23:11:54Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Will-Amplify | 266,859,122 | U_kgDOD-fycg | User | false |
openclaw/openclaw | 4,290,627,115 | I_kwDOQb6kR87_vcYr | 68,923 | https://github.com/openclaw/openclaw/issues/68923 | https://api.github.com/repos/openclaw/openclaw/issues/68923 | Bug: Agent loops on identical tool calls without stopping | ---
name: Tool Call Repetition Bug - Agent loops on identical tool calls
about: Agent repeatedly executes the same tool call (e.g., Get-ChildItem) 100+ times without stopping, even after receiving identical results.
title: "Bug: Agent loops on identical tool calls without stopping"
labels: ["bug", "agent-behavior", "hi... | closed | not_planned | true | 1 | [] | [] | 2026-04-19T10:53:24Z | 2026-04-30T04:39:08Z | 2026-04-26T05:22:50Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | yuan3471 | 272,229,708 | U_kgDOEDnlTA | User | false |
openclaw/openclaw | 4,290,639,283 | I_kwDOQb6kR87_vfWz | 68,929 | https://github.com/openclaw/openclaw/issues/68929 | https://api.github.com/repos/openclaw/openclaw/issues/68929 | Feature: user interrupt during streaming / tool execution | ## Problem
When the agent is mid-turn (e.g., running a long tool call, streaming results), the user has no way to interrupt or redirect. Messages queue until the full turn completes.
This is particularly problematic for:
- **Search/list tasks**: user finds item #1 acceptable and wants to stop the search early
- **Par... | closed | completed | true | 2 | [] | [] | 2026-04-19T10:58:39Z | 2026-04-28T04:46:09Z | 2026-04-25T18:43:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ttomiczek | 265,786,127 | U_kgDOD9eTDw | User | false |
openclaw/openclaw | 4,290,652,856 | I_kwDOQb6kR87_viq4 | 68,931 | https://github.com/openclaw/openclaw/issues/68931 | https://api.github.com/repos/openclaw/openclaw/issues/68931 | [Bug]: Webchat user messages disappear due to race condition between chat history reload and transcript persistence | ## Bug Description
User messages can disappear from the OpenClaw webchat UI after being sent. The message is typed, sent, briefly appears, then vanishes. The user did not delete it — the message disappears on its own.
## Root Cause Analysis
After analyzing the webchat frontend code and gateway logs, I identified a *... | closed | completed | true | 1 | [] | [] | 2026-04-19T11:05:38Z | 2026-04-28T04:50:13Z | 2026-04-25T18:43:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | geek111 | 6,862,522 | MDQ6VXNlcjY4NjI1MjI= | User | false |
openclaw/openclaw | 4,290,648,930 | I_kwDOQb6kR87_vhti | 68,930 | https://github.com/openclaw/openclaw/issues/68930 | https://api.github.com/repos/openclaw/openclaw/issues/68930 | Auto-compaction silently no-ops on tool-loop preflight overflow because token count is not propagated | # Auto-compaction silently no-ops on tool-loop preflight overflow because token count isn't propagated
## Summary
When the embedded pi runner trips its preflight context-overflow guard during a tool loop, auto-compaction silently no-ops and the session is "restarted" (new sessionId, fragmented context) — even when an... | closed | not_planned | true | 0 | [] | [] | 2026-04-19T11:03:13Z | 2026-04-22T04:31:05Z | 2026-04-19T11:08:30Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | kscheyer | 42,887,107 | MDQ6VXNlcjQyODg3MTA3 | User | false |
openclaw/openclaw | 4,290,657,246 | I_kwDOQb6kR87_vjve | 68,932 | https://github.com/openclaw/openclaw/issues/68932 | https://api.github.com/repos/openclaw/openclaw/issues/68932 | Config validation error after removing bundled plugins from allowlist (RESOLVED) | ## Error Details
After removing bundled plugins (`deepseek`, `brave`, `kimi`, `memory-core`) from `plugins.allow`, OpenClaw CLI fails to start with:
```
Failed to start CLI: Error: Invalid config at /Users/andy51/.openclaw/openclaw.json:
```
## Environment
- OpenClaw version: 2026.4.15
- OS: macOS 26.4.1 (arm64)
- N... | closed | completed | true | 4 | [] | [] | 2026-04-19T11:08:09Z | 2026-04-22T04:31:05Z | 2026-04-19T11:09:50Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | woai36D | 269,974,840 | U_kgDOEBd9OA | User | false |
openclaw/openclaw | 4,290,660,663 | I_kwDOQb6kR87_vkk3 | 68,933 | https://github.com/openclaw/openclaw/issues/68933 | https://api.github.com/repos/openclaw/openclaw/issues/68933 | [DELETED by author request] Config validation issue - please ignore | null | closed | completed | true | 1 | [] | [] | 2026-04-19T11:10:00Z | 2026-04-22T04:31:04Z | 2026-04-19T11:10:58Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | woai36D | 269,974,840 | U_kgDOEBd9OA | User | false |
openclaw/openclaw | 4,222,092,200 | I_kwDOQb6kR877qAOo | 62,921 | https://github.com/openclaw/openclaw/issues/62921 | https://api.github.com/repos/openclaw/openclaw/issues/62921 | v2026.4.7: Telegram extension missing src/channel.setup.js — config invalid on upgrade | ## Bug Report
**OpenClaw version:** 2026.4.7 (5050017)
**Previous version:** 2026.4.5 (3e72c03) — working fine
**OS:** macOS 15 (arm64), installed via `npm install -g openclaw`
## Problem
After upgrading from 2026.4.5 to 2026.4.7, `openclaw status` fails immediately with:
```
Config invalid
File: ~/.openclaw/opencl... | closed | completed | true | 2 | [] | [] | 2026-04-08T04:22:47Z | 2026-04-27T05:09:33Z | 2026-04-24T05:48:43Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | justinbao19 | 97,433,668 | U_kgDOBc64RA | User | false |
openclaw/openclaw | 4,079,257,657 | I_kwDOQb6kR87zJIg5 | 47,643 | https://github.com/openclaw/openclaw/issues/47643 | https://api.github.com/repos/openclaw/openclaw/issues/47643 | [Bug]: Persistent Telegram Channel Issues: Sync Failures, Loops, and Config Changes Not Applying on Windows | ### Bug type
Regression (worked before, now fails)
### Summary
I'm running OpenClaw v2026.3.13 on Windows 10, and I've encountered several foundational issues with the Telegram channel that make it unusable for multi-device access and sync. The problems persisted despite multiple restarts, config edits, token change... | open | null | false | 5 | [
"bug",
"regression"
] | [] | 2026-03-15T21:43:04Z | 2026-05-04T01:33:19Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | opiedog | 8,853,147 | MDQ6VXNlcjg4NTMxNDc= | User | false |
openclaw/openclaw | 4,290,707,103 | I_kwDOQb6kR87_vv6f | 68,937 | https://github.com/openclaw/openclaw/issues/68937 | https://api.github.com/repos/openclaw/openclaw/issues/68937 | [DELETED] 测试权限 - 误提交请忽略 | null | closed | completed | true | 1 | [] | [] | 2026-04-19T11:34:39Z | 2026-04-22T04:31:02Z | 2026-04-19T11:34:54Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | woai36D | 269,974,840 | U_kgDOEBd9OA | User | false |
openclaw/openclaw | 4,290,712,354 | I_kwDOQb6kR87_vxMi | 68,938 | https://github.com/openclaw/openclaw/issues/68938 | https://api.github.com/repos/openclaw/openclaw/issues/68938 | Web UI: Voice input button fails + image upload broken (2026.4.15) | ## Bug Description
**OS:** Windows_NT 10.0.22631
**OpenClaw version:** 2026.4.15 (041266a)
**Browser:** Chrome (same issue in incognito mode)
### Issues:
1. **Voice input fails:** Clicking the microphone button in Control UI (http://127.0.0.1:18789) causes the icon to briefly flash then become disabled. No voice inpu... | closed | completed | true | 1 | [] | [] | 2026-04-19T11:37:31Z | 2026-04-28T04:50:12Z | 2026-04-25T18:43:53Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lucy6692 | 153,696,593 | U_kgDOCSk5UQ | User | false |
openclaw/openclaw | 4,169,989,809 | I_kwDOQb6kR874jP6x | 57,669 | https://github.com/openclaw/openclaw/issues/57669 | https://api.github.com/repos/openclaw/openclaw/issues/57669 | [Bug]: windows telegram can't work | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrade 2026.03.22 the telegram can't work in windows 11
not use wsl in cmd
### Steps to reproduce
1.Start openclaw 2026.03.22 with config file :
{
"meta": {
"lastTouchedVersion": "2026.3.28",
"lastTouche... | closed | not_planned | true | 2 | [
"bug"
] | [] | 2026-03-30T11:35:04Z | 2026-04-28T04:58:48Z | 2026-04-25T17:06:07Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kuolung1 | 43,897,147 | MDQ6VXNlcjQzODk3MTQ3 | User | false |
openclaw/openclaw | 4,290,724,923 | I_kwDOQb6kR87_v0Q7 | 68,940 | https://github.com/openclaw/openclaw/issues/68940 | https://api.github.com/repos/openclaw/openclaw/issues/68940 | [Bug]: CLI Backends quick-start command fails on 2026.4.15 — `openclaw agent` has no `--model` flag | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
## Where
https://docs.openclaw.ai/gateway/cli-backends
Section: **Beginner-friendly quick start**
## The issue
The quick-start command
```
openclaw agent --message "hi" --model codex-cli/gpt-5.4
```
does ... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-19T11:44:08Z | 2026-05-02T15:23:03Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | drmarcopapa | 39,249,407 | MDQ6VXNlcjM5MjQ5NDA3 | User | false |
openclaw/openclaw | 4,290,747,635 | I_kwDOQb6kR87_v5zz | 68,944 | https://github.com/openclaw/openclaw/issues/68944 | https://api.github.com/repos/openclaw/openclaw/issues/68944 | [Bug]: CLI commands hang at WebSocket gateway handshake | # CLI Commands Hang at WebSocket Gateway Handshake
## Bug Description
Every CLI subcommand that requires a live gateway connection hangs indefinitely. The CLI successfully connects to the gateway WebSocket endpoint and receives the `connect.challenge` nonce, but then never sends the `connect.reply` — the command hang... | open | null | false | 6 | [] | [] | 2026-04-19T11:56:16Z | 2026-05-03T10:41:24Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | WaMaSeDu | 265,785,875 | U_kgDOD9eSEw | User | false |
openclaw/openclaw | 4,290,751,896 | I_kwDOQb6kR87_v62Y | 68,945 | https://github.com/openclaw/openclaw/issues/68945 | https://api.github.com/repos/openclaw/openclaw/issues/68945 | [Bug]: Telegram: humanDelay is not applied between streamed block replies | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Summary
`agents.defaults.humanDelay` (and per-agent `humanDelay` overrides) have no effect on the Telegram channel. Block replies produced by the Telegram streaming pipeline are sent back-to-back with no human-like pacin... | open | null | false | 1 | [
"bug",
"regression"
] | [] | 2026-04-19T11:58:32Z | 2026-05-02T15:23:05Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | mastyf | 5,397,990 | MDQ6VXNlcjUzOTc5OTA= | User | false |
openclaw/openclaw | 4,290,753,225 | I_kwDOQb6kR87_v7LJ | 68,946 | https://github.com/openclaw/openclaw/issues/68946 | https://api.github.com/repos/openclaw/openclaw/issues/68946 | Bug: Agent loops on identical tool calls without stopping | ---
name: Tool Call Repetition Bug - Agent loops on identical tool calls
about: Agent repeatedly executes the same tool call (e.g., Get-ChildItem) 100+ times without stopping, even after receiving identical results.
title: "Bug: Agent loops on identical tool calls without stopping"
labels: ["bug", "agent-behavior", "hi... | closed | not_planned | true | 1 | [] | [] | 2026-04-19T11:59:18Z | 2026-04-30T04:39:07Z | 2026-04-26T05:22:55Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | yuan3471 | 272,229,708 | U_kgDOEDnlTA | User | false |
openclaw/openclaw | 4,290,795,041 | I_kwDOQb6kR87_wFYh | 68,950 | https://github.com/openclaw/openclaw/issues/68950 | https://api.github.com/repos/openclaw/openclaw/issues/68950 | OAuth users cannot enable memory-lancedb auto-recall/capture without separate OPENAI_API_KEY | ## Summary
Users authenticated to OpenClaw via OAuth/provider login still cannot enable the built-in `memory-lancedb` plugin unless they separately provide `OPENAI_API_KEY` for embeddings. This blocks the closest built-in equivalent to always-on active memory for users who intentionally rely on OAuth-only auth instead ... | closed | completed | true | 1 | [] | [] | 2026-04-19T12:20:14Z | 2026-04-30T04:32:43Z | 2026-04-28T00:43:01Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | malshaalan-ai | 260,099,975 | U_kgDOD4DPhw | User | false |
openclaw/openclaw | 4,290,795,617 | I_kwDOQb6kR87_wFhh | 68,951 | https://github.com/openclaw/openclaw/issues/68951 | https://api.github.com/repos/openclaw/openclaw/issues/68951 | │
◇ Config warnings ───────────────────────────────────────────────────────╮
│ │
│ - plugins.entries.memory-lancedb: plugin disabled (memory slot set to │
│ "memory-core") but config is present │
│ - plugins.e... | ## Summary
Running `rem-backfill` for a non-main agent can write dream output into the wrong workspace.
## Repro
Command run:
```bash
openclaw memory rem-backfill --agent stack --path /Users/boazooz/.openclaw/stack-workspace/memory --stage-short-term
```
## Observed behavior
CLI output reported:
- `workspace=~/.openc... | closed | not_planned | false | 1 | [] | [] | 2026-04-19T12:20:30Z | 2026-04-25T18:44:00Z | 2026-04-25T18:44:00Z | NONE | null | 20260425T210014Z | 2026-04-25T21:00:14Z | malshaalan-ai | 260,099,975 | U_kgDOD4DPhw | User | false |
openclaw/openclaw | 4,290,796,328 | I_kwDOQb6kR87_wFso | 68,952 | https://github.com/openclaw/openclaw/issues/68952 | https://api.github.com/repos/openclaw/openclaw/issues/68952 | [Feature]: UX improvement during onboarding ollama setup | ### Summary
While onboarding, I chose Ollama as the provider and Gemma4 as the model. The UX could use several improvements, like de-duping models and not erroneously saying it would download the already existing model.
### Problem to solve
In the UI after selecting gemma4, it showed `Downloading ... ` and I almost... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-19T12:20:48Z | 2026-04-30T04:34:37Z | 2026-04-27T11:14:31Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | tleyden | 296,876 | MDQ6VXNlcjI5Njg3Ng== | User | false |
openclaw/openclaw | 4,111,059,401 | I_kwDOQb6kR871CcnJ | 51,383 | https://github.com/openclaw/openclaw/issues/51383 | https://api.github.com/repos/openclaw/openclaw/issues/51383 | Feature: Webchat project/thread system for parallel contexts | ## Problem
Webchat is the most powerful interface for OpenClaw (direct host access, no rate limits, full tool access), but it lacks basic organization features that Discord/Telegram provide natively.
Currently:
- Single continuous stream per session
- `/new` or `/reset` clears all context
- No way to work on multiple... | closed | completed | true | 3 | [] | [] | 2026-03-21T02:32:50Z | 2026-04-30T04:52:19Z | 2026-04-25T05:08:52Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | sabo961 | 191,984 | MDQ6VXNlcjE5MTk4NA== | User | false |
openclaw/openclaw | 4,290,812,134 | I_kwDOQb6kR87_wJjm | 68,955 | https://github.com/openclaw/openclaw/issues/68955 | https://api.github.com/repos/openclaw/openclaw/issues/68955 | [Feature]: Slash commands like /new should trigger ack reaction and typing indicator | ### Summary
## Summary
When using the Telegram channel with `reactionLevel: ack` + `ackReactionScope: direct`, regular messages correctly trigger the 👀 ack reaction and typing indicator. However, slash commands like `/new` do not trigger these UX feedback mechanisms, leading to inconsistent user experience.
## Envi... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-19T12:28:23Z | 2026-05-02T15:23:07Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | C0smicExpl0rer | 325,002 | MDQ6VXNlcjMyNTAwMg== | User | false |
openclaw/openclaw | 4,290,855,037 | I_kwDOQb6kR87_wUB9 | 68,957 | https://github.com/openclaw/openclaw/issues/68957 | https://api.github.com/repos/openclaw/openclaw/issues/68957 | [Bug]: LLM request failed: provider rejected the request schema or tool payload. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The AI agent keeps failing with the following error on every message:
`LLM request failed: provider rejected the request schema or tool payload`
### Reproduction
1. Open the AI agent
2. Send a prompt
3. Observe the failure... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-19T12:46:58Z | 2026-04-28T04:50:10Z | 2026-04-25T18:44:06Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | goktugcy | 17,319,750 | MDQ6VXNlcjE3MzE5NzUw | User | false |
openclaw/openclaw | 4,290,895,692 | I_kwDOQb6kR87_wd9M | 68,959 | https://github.com/openclaw/openclaw/issues/68959 | https://api.github.com/repos/openclaw/openclaw/issues/68959 | Control UI: No agent switcher — cannot switch between agents in web UI | ## Problem
When multiple agents are configured in openclaw.json, there is no way to switch between them in the Control UI. The UI only shows sessions, but no agent picker or switcher to select which agent to chat with.
Users must use /agent commands within an existing session, or rely on external channels (Telegram, ... | closed | not_planned | true | 1 | [] | [] | 2026-04-19T13:04:42Z | 2026-05-03T04:35:42Z | 2026-04-30T17:59:27Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | mianayo | 105,649,803 | U_kgDOBkwWiw | User | false |
openclaw/openclaw | 4,290,896,851 | I_kwDOQb6kR87_wePT | 68,960 | https://github.com/openclaw/openclaw/issues/68960 | https://api.github.com/repos/openclaw/openclaw/issues/68960 | [Feature]: Onboarding got "stuck" after Hatch in TUI (recommended) | ### Summary
I got to the end of the onboarding (I think it's the end at least, there is no progress indicator), and I tried to hatch in the TUI, but ended up getting stuck. In a separate terminal I ran `openclaw devices list`, which worked, but could not figure out how to continue in the onboarding. I am stuck her... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-04-19T13:05:09Z | 2026-04-28T04:32:50Z | 2026-04-25T18:44:13Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tleyden | 296,876 | MDQ6VXNlcjI5Njg3Ng== | User | false |
openclaw/openclaw | 4,290,923,210 | I_kwDOQb6kR87_wkrK | 68,962 | https://github.com/openclaw/openclaw/issues/68962 | https://api.github.com/repos/openclaw/openclaw/issues/68962 | [Bug]: HTTPS for localhost! | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
Oh, brilliant move requiring HTTPS for localhost! You know what's better than accessing your own local server directly? Making you jump through certificate hoops like you're launching a NASA mission!
"Secure context required" ... | closed | not_planned | true | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-19T13:18:02Z | 2026-04-28T04:50:09Z | 2026-04-25T18:44:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Michae1G | 15,974,778 | MDQ6VXNlcjE1OTc0Nzc4 | User | false |
openclaw/openclaw | 4,290,962,344 | I_kwDOQb6kR87_wuOo | 68,966 | https://github.com/openclaw/openclaw/issues/68966 | https://api.github.com/repos/openclaw/openclaw/issues/68966 | [Bug]: [WSL] openclaw browser command terminated by SIGKILL causing timeout | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
Hello, previously when I was using OpenClaw's session control browser, the command was displayed as SIGKILL. I installed OpenClaw in the WSL environment to control the host Chrome, and checked that the OpenClaw browser command ... | open | null | false | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-19T13:36:26Z | 2026-05-02T14:22:14Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | The-Black-Sun | 72,015,666 | MDQ6VXNlcjcyMDE1NjY2 | User | false |
openclaw/openclaw | 4,290,945,663 | I_kwDOQb6kR87_wqJ_ | 68,965 | https://github.com/openclaw/openclaw/issues/68965 | https://api.github.com/repos/openclaw/openclaw/issues/68965 | Bug: models.mode: replace does not exclude implicit providers from bundled plugins | ## Bug Description
When `models.mode` is set to `"replace"`, the configuration should **only** use explicitly configured providers and exclude all implicit providers discovered from bundled plugins. However, the current implementation still includes implicit providers (e.g., `codex`) in the generated `models.json`.
#... | closed | not_planned | true | 1 | [] | [] | 2026-04-19T13:29:49Z | 2026-04-29T04:37:37Z | 2026-04-26T10:38:51Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Tony-ooo | 45,898,990 | MDQ6VXNlcjQ1ODk4OTkw | User | false |
openclaw/openclaw | 4,290,986,796 | I_kwDOQb6kR87_w0Ms | 68,970 | https://github.com/openclaw/openclaw/issues/68970 | https://api.github.com/repos/openclaw/openclaw/issues/68970 | TUI misses post-restart heartbeat reply, while the same reply is visible in Web UI | ## Summary
After a gateway restart, the assistant is successfully reawakened by the post-restart heartbeat flow and produces a reply, but that reply is not shown in the TUI session the user was actively watching. The same reply is visible in Web UI.
This looks like a TUI reconnect/backfill problem rather than a faile... | open | null | false | 1 | [] | [] | 2026-04-19T13:44:14Z | 2026-05-02T14:22:16Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | Silronin | 67,084,317 | MDQ6VXNlcjY3MDg0MzE3 | User | false |
openclaw/openclaw | 4,290,985,764 | I_kwDOQb6kR87_wz8k | 68,969 | https://github.com/openclaw/openclaw/issues/68969 | https://api.github.com/repos/openclaw/openclaw/issues/68969 | [Bug]:[WSL] The CPU usage of OpenClaw's gateway process spikes to 100%. | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
Without performing any operations or having scheduled tasks, simply suspending the gateway resulted in a sudden spike in CPU usage, reaching as high as 100%.
### Steps to reproduce
1.Start OpenClaw 2026.4.15 Gateway
2.Everyth... | closed | not_planned | true | 2 | [
"bug",
"bug:crash"
] | [] | 2026-04-19T13:43:54Z | 2026-04-30T04:39:06Z | 2026-04-26T05:23:00Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | 343242 | 154,515,357 | U_kgDOCTW3nQ | User | false |
openclaw/openclaw | 4,291,011,689 | I_kwDOQb6kR87_w6Rp | 68,972 | https://github.com/openclaw/openclaw/issues/68972 | https://api.github.com/repos/openclaw/openclaw/issues/68972 | [Bug] Dreaming cron task executes in isolated but not main sessions | ## Bug Description
The dreaming agent task is designed to run promotion logic, but when the cron job is configured with sessionTarget set to main, it incorrectly executes heartbeat checks instead of the intended dreaming/promotion logic.
## Expected Behavior
- Cron job with dreaming agent task should execute promoti... | closed | completed | true | 0 | [] | [] | 2026-04-19T13:51:57Z | 2026-04-27T05:09:41Z | 2026-04-24T05:23:21Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | raydoomed | 57,515,812 | MDQ6VXNlcjU3NTE1ODEy | User | false |
openclaw/openclaw | 4,291,035,041 | I_kwDOQb6kR87_w_-h | 68,977 | https://github.com/openclaw/openclaw/issues/68977 | https://api.github.com/repos/openclaw/openclaw/issues/68977 | [Enhancement] Add scoped idempotency key support to enqueueSystemEvent | ## Summary
Requesting support for scoped idempotency keys in the `enqueueSystemEvent` API to eliminate race conditions in plugin resume flows.
## Use Case
The **triggerx-startup** plugin needs a reliable way to resume interrupted flows. When a plugin restarts, it must enqueue system events to continue processing. With... | closed | completed | false | 3 | [] | [] | 2026-04-19T13:59:05Z | 2026-05-05T11:21:21Z | 2026-05-05T11:21:21Z | CONTRIBUTOR | null | 20260505T120033Z | 2026-05-05T12:00:33Z | bottenbenny | 270,688,955 | U_kgDOECJiuw | User | false |
openclaw/openclaw | 4,291,071,721 | I_kwDOQb6kR87_xI7p | 68,979 | https://github.com/openclaw/openclaw/issues/68979 | https://api.github.com/repos/openclaw/openclaw/issues/68979 | lobster embedded runner cannot reach llm-task tool | ## Summary
When a lobster workflow uses `llm.invoke` with the `openclaw` provider, and the workflow is executed via the **lobster plugin's embedded runner** (or via `exec lobster run ...` from an agent session), the `llm-task` tool is unreachable.
## Steps to reproduce
1. Enable `llm-task` and `lobster` plugins
2. C... | closed | not_planned | true | 1 | [] | [] | 2026-04-19T14:11:23Z | 2026-04-28T04:50:08Z | 2026-04-25T18:44:26Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | popoffvg | 19,862,834 | MDQ6VXNlcjE5ODYyODM0 | User | false |
openclaw/openclaw | 4,291,080,503 | I_kwDOQb6kR87_xLE3 | 68,980 | https://github.com/openclaw/openclaw/issues/68980 | https://api.github.com/repos/openclaw/openclaw/issues/68980 | [Bug]: exec command triggers spawn taskkill ENOENT and crashes gateway | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Every time an exec command runs, OpenClaw tries to kill the old gateway process using taskkill. The spawn fails with ENOENT on Windows, causing an uncaught exception that crashes the gateway.
### Steps to reproduce
1. Run ... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-19T14:14:30Z | 2026-04-28T04:50:08Z | 2026-04-25T18:44:32Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Johnsonchan0001 | 277,408,180 | U_kgDOEIjptA | User | false |
openclaw/openclaw | 4,291,088,529 | I_kwDOQb6kR87_xNCR | 68,981 | https://github.com/openclaw/openclaw/issues/68981 | https://api.github.com/repos/openclaw/openclaw/issues/68981 | [Enhancement] Session Management: Human-readable names, auto-cleanup, and grouped Control UI | ## Problem
OpenClaw's session management has three related UX problems:
### 1. Sessions are identified by UUIDs — completely unreadable
The session list (both in `sessions.json` and Control UI) uses auto-generated UUIDs as session keys. There is no way to give a session a human-readable name. After a few weeks of us... | open | null | false | 3 | [] | [] | 2026-04-19T14:17:13Z | 2026-05-03T08:39:31Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | loybo | 108,980,394 | U_kgDOBn7oqg | User | false |
openclaw/openclaw | 4,291,113,131 | I_kwDOQb6kR87_xTCr | 68,983 | https://github.com/openclaw/openclaw/issues/68983 | https://api.github.com/repos/openclaw/openclaw/issues/68983 | Gateway startup regressed ~5s → ~35s after 4.8 → 4.9 due to repeated jiti normalizeAliases calls | # Gateway startup regressed ~5s → ~35s after 4.8 → 4.9 due to repeated jiti `normalizeAliases` calls
## Summary
Since **2026.4.9**, gateway startup went from ~5s to ~35s in our setup (6 bundled plugins, Node 24). We bisected with npm installs of 4.7/4.8/4.9/4.10/…/4.15 and the regression lands cleanly on **4.9**.
Ro... | closed | completed | true | 1 | [] | [] | 2026-04-19T14:26:17Z | 2026-04-25T04:17:34Z | 2026-04-22T17:01:46Z | NONE | null | 20260425T060021Z | 2026-04-25T06:00:21Z | aliyaredpilled | 148,153,466 | U_kgDOCNSkeg | User | false |
openclaw/openclaw | 4,291,116,619 | I_kwDOQb6kR87_xT5L | 68,984 | https://github.com/openclaw/openclaw/issues/68984 | https://api.github.com/repos/openclaw/openclaw/issues/68984 | [Bug] Google provider does not work with Gemma models - OpenAI API format incompatibility | ## Bug Description
The Google provider in OpenClaw uses the OpenAI-compatible endpoint (`/v1beta/openai/chat/completions`), but Gemma models (gemma-4-31b-it, gemma-4-26b-it) only support Google's native API format (`/v1beta/models/{model}:generateContent`).
## Steps to Reproduce
1. Configure a Google API key (AI Stu... | closed | completed | true | 2 | [] | [] | 2026-04-19T14:27:33Z | 2026-04-27T04:39:26Z | 2026-04-25T02:42:33Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | tanjinlimkelvin-dot | 266,243,000 | U_kgDOD96LuA | User | false |
openclaw/openclaw | 4,068,610,130 | I_kwDOQb6kR87yghBS | 44,611 | https://github.com/openclaw/openclaw/issues/44611 | https://api.github.com/repos/openclaw/openclaw/issues/44611 | [Bug]: Gateway does not load model configuration from openclaw.json on restart | ### Bug type
Regression (worked before, now fails)
### Summary
The OpenClaw gateway is failing to apply model configuration changes from ~/.openclaw/openclaw.json after a service restart. It continues to use the model that was active before the configuration file was changed, effectively ignoring the updated settin... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-03-13T03:33:35Z | 2026-04-22T04:30:58Z | 2026-04-19T14:28:49Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | TheColorPlum | 19,821,872 | MDQ6VXNlcjE5ODIxODcy | User | false |
openclaw/openclaw | 4,291,151,243 | I_kwDOQb6kR87_xcWL | 68,985 | https://github.com/openclaw/openclaw/issues/68985 | https://api.github.com/repos/openclaw/openclaw/issues/68985 | openclaw update restarts gateway but does not refresh installed systemd unit metadata | ### Summary
When OpenClaw is installed as a user systemd service, running `openclaw update` upgrades the package and restarts the gateway, but it does not regenerate the installed unit file.
As a result, the running CLI/runtime version is updated while the service unit metadata remains stale, for example:
- CLI/runt... | closed | completed | true | 1 | [] | [] | 2026-04-19T14:40:41Z | 2026-04-28T04:50:07Z | 2026-04-25T18:44:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | neodefier | 273,253,241 | U_kgDOEEmDeQ | User | false |
openclaw/openclaw | 4,291,165,218 | I_kwDOQb6kR87_xfwi | 68,988 | https://github.com/openclaw/openclaw/issues/68988 | https://api.github.com/repos/openclaw/openclaw/issues/68988 | [Bug]: CLI subcommands are extremely slow on Raspberry Pi 5 (20s for `openclaw config --help`) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On a Raspberry Pi 5 with OpenClaw 2026.4.15, root help is fast but CLI subcommands are extremely slow. `openclaw --help` completes in about 0.05s, while `openclaw config --help` takes about 21.4s on the same ma... | closed | completed | true | 4 | [
"bug"
] | [] | 2026-04-19T14:47:18Z | 2026-04-30T04:44:38Z | 2026-04-25T07:17:03Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | system-dsn | 66,951,696 | MDQ6VXNlcjY2OTUxNjk2 | User | false |
openclaw/openclaw | 4,291,188,475 | I_kwDOQb6kR87_xlb7 | 68,990 | https://github.com/openclaw/openclaw/issues/68990 | https://api.github.com/repos/openclaw/openclaw/issues/68990 | Discord slash commands crash in DM channels: Cannot access rawData on partial Channel | ## Bug Description
Discord slash commands (e.g. `/new`, `/compact`, `/status`) fail silently in DM channels. The bot receives the interaction but crashes before processing it.
## Error
```
Error: Cannot access rawData on partial Channel. Use fetch() to populate data.
at DmChannel.get rawData [as rawData] (@buape... | closed | completed | true | 3 | [] | [] | 2026-04-19T14:57:37Z | 2026-04-25T04:17:00Z | 2026-04-22T19:42:23Z | NONE | null | 20260425T060021Z | 2026-04-25T06:00:21Z | zhengsx | 3,127,699 | MDQ6VXNlcjMxMjc2OTk= | User | false |
openclaw/openclaw | 4,291,196,917 | I_kwDOQb6kR87_xnf1 | 68,992 | https://github.com/openclaw/openclaw/issues/68992 | https://api.github.com/repos/openclaw/openclaw/issues/68992 | [Bug]: Control UI renders async exec system events in the visible chat transcript | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The Control UI displays async exec completion system events in the visible chat transcript instead of keeping them internal or clearly separated from normal chat messages.
### Steps to reproduce
Open a direct... | closed | completed | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-04-19T15:01:19Z | 2026-04-28T04:50:06Z | 2026-04-25T18:44:45Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | WhiteCrispStudios | 242,942,136 | U_kgDODnsAuA | User | false |
openclaw/openclaw | 4,196,186,359 | I_kwDOQb6kR876HLj3 | 59,849 | https://github.com/openclaw/openclaw/issues/59849 | https://api.github.com/repos/openclaw/openclaw/issues/59849 | Browser tool unusable on datacenter IPs — Indeed, LinkedIn, Glassdoor, DuckDuckGo all block VPS IP ranges | The OpenClaw browser tool (Chrome CDP) works great in principle, but is effectively unusable for mainstream job sites and search engines because the VPS/server IP is blocked by Cloudflare CAPTCHAs.
## Affected sites confirmed blocked:
- Indeed (de.indeed.com) — Cloudflare block
- LinkedIn — requires login + bot detect... | closed | completed | true | 4 | [] | [] | 2026-04-02T18:14:09Z | 2026-04-28T05:12:52Z | 2026-04-25T13:17:44Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ttomiczek | 265,786,127 | U_kgDOD9eTDw | User | false |
openclaw/openclaw | 4,291,218,073 | I_kwDOQb6kR87_xsqZ | 68,995 | https://github.com/openclaw/openclaw/issues/68995 | https://api.github.com/repos/openclaw/openclaw/issues/68995 | Memory flush write restriction should allow subdirectories under memory/ | ## Problem
The memory flush mechanism hardcodes the allowed write path to exactly `memory/YYYY-MM-DD.md`. This prevents skills from writing to structured subdirectories like `memory/daily/YYYY-MM-DD.md`.
When a skill writes to a memory subdirectory, it gets rejected with:
> Memory flush writes are restricted to memor... | open | null | false | 2 | [] | [] | 2026-04-19T15:10:55Z | 2026-05-02T15:23:09Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | AxelHu | 12,809,280 | MDQ6VXNlcjEyODA5Mjgw | User | false |
openclaw/openclaw | 4,291,182,542 | I_kwDOQb6kR87_xj_O | 68,989 | https://github.com/openclaw/openclaw/issues/68989 | https://api.github.com/repos/openclaw/openclaw/issues/68989 | [Bug]: Gateway fails to start on 2026.4.14 — SyntaxError in bundled dist/model-context-tokens-CwcLB3PA.js | ## Summary
OpenClaw `2026.4.14` ships a bundled file with a syntax error that prevents the gateway from starting. The crash is immediate and blocks all gateway operations until the file is manually patched.
## Steps to Reproduce
1. Install OpenClaw `2026.4.14`
2. Start the gateway: `./run.sh restart` or `openclaw ga... | closed | not_planned | true | 2 | [] | [] | 2026-04-19T14:54:59Z | 2026-04-22T04:30:55Z | 2026-04-19T15:18:00Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | sanchezm86 | 66,266,304 | MDQ6VXNlcjY2MjY2MzA0 | User | false |
openclaw/openclaw | 4,291,247,802 | I_kwDOQb6kR87_xz66 | 68,997 | https://github.com/openclaw/openclaw/issues/68997 | https://api.github.com/repos/openclaw/openclaw/issues/68997 | Bug: `task-registry-control.runtime.{js,ts}` missing from published npm tarballs, breaks `openclaw tasks cancel` |
## Summary
Current published OpenClaw npm tarballs include loader code that tries to require `./task-registry-control.runtime.js` or `./task-registry-control.runtime.ts`, but neither file is shipped in the tarball. This breaks `openclaw tasks cancel` and related task-control paths with:
```text
Failed to load task r... | closed | completed | true | 2 | [] | [] | 2026-04-19T15:24:21Z | 2026-05-01T04:42:24Z | 2026-04-28T17:47:04Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | 1OAKDesign | 143,361,737 | U_kgDOCIuGyQ | User | false |
openclaw/openclaw | 4,291,250,267 | I_kwDOQb6kR87_x0hb | 68,998 | https://github.com/openclaw/openclaw/issues/68998 | https://api.github.com/repos/openclaw/openclaw/issues/68998 | Telegram outbound media: plain assistant MEDIA lines can fall back to text-only sends; explicit message-tool media path works | ## Summary
Telegram media delivery is easy to get wrong from agent output.
In my case, a normal assistant reply containing a literal `MEDIA:/path/to/file.png` line did **not** produce a native Telegram image attachment. The user saw no image. The reliable path was using the explicit `message` tool with `media`/`path`... | closed | completed | true | 1 | [] | [] | 2026-04-19T15:25:32Z | 2026-04-30T04:44:43Z | 2026-04-25T07:14:06Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | EthanSK | 13,756,744 | MDQ6VXNlcjEzNzU2NzQ0 | User | false |
openclaw/openclaw | 4,291,255,067 | I_kwDOQb6kR87_x1sb | 68,999 | https://github.com/openclaw/openclaw/issues/68999 | https://api.github.com/repos/openclaw/openclaw/issues/68999 | session_status 显示的 Context 百分比始终为 0%，无法正确统计当前会话的上下文使用量 | ## 问题描述
在当前会话中已经发送了多条消息（包含系统提示词、多个 markdown 文件、用户对话等），但执行 `session_status` 时显示 **Context: 0/200k (0%)**，这明显与实际不符。
## 复现步骤
1. 启动一个新的 Feishu DM 会话
2. 发送几条消息（至少 3-5 条）
3. 执行 session_status 查看上下文统计
4. 观察到 Context 显示 0%
## 期望行为
Context 百分比应该正确反映当前会话已使用的 token 数量，而非始终显示 0%。
## 额外说明
据我了解，Hermes 助手已经有能力正确统计上下文使用量。希望 Ope... | closed | completed | true | 2 | [] | [] | 2026-04-19T15:27:55Z | 2026-04-28T04:50:05Z | 2026-04-25T18:44:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | BayiHY | 264,720,027 | U_kgDOD8dOmw | User | false |
openclaw/openclaw | 4,194,995,882 | I_kwDOQb6kR876Co6q | 59,722 | https://github.com/openclaw/openclaw/issues/59722 | https://api.github.com/repos/openclaw/openclaw/issues/59722 | BlueBubbles: SSRF allowlist doesn't cover reactions (react action blocked on private IP) | ## Summary
The BlueBubbles channel's auto-allowlist for the configured `serverUrl` doesn't cover the `react` action. Reactions to messages fail with:
```
Blocked hostname or private/internal/special-use IP address
```
## Environment
- OpenClaw version: `2026.4.1`
- BlueBubbles serverUrl: `http://192.168.x.x:1234` (... | closed | completed | true | 0 | [] | [] | 2026-04-02T14:28:52Z | 2026-04-22T04:30:54Z | 2026-04-19T15:43:34Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | kevin-funk-town | 263,518,436 | U_kgDOD7T45A | User | false |
openclaw/openclaw | 4,291,299,279 | I_kwDOQb6kR87_yAfP | 69,006 | https://github.com/openclaw/openclaw/issues/69006 | https://api.github.com/repos/openclaw/openclaw/issues/69006 | pnpm check fails on upstream/main — pre-existing type errors in Discord + qa-lab | ## Summary
`pnpm check` fails on a fresh checkout of `upstream/main` (commit `84cd786911`) with 8 pre-existing TypeScript errors across two extensions. This blocks the pre-commit hook for any contributor whose hook runs the same check, forcing `--no-verify` on unrelated PRs.
## Reproduction
```bash
git fetch upstrea... | closed | completed | true | 1 | [] | [] | 2026-04-19T15:46:27Z | 2026-04-28T04:50:04Z | 2026-04-25T18:45:05Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | martingarramon | 263,922,628 | U_kgDOD7sjxA | User | false |
openclaw/openclaw | 4,291,316,204 | I_kwDOQb6kR87_yEns | 69,012 | https://github.com/openclaw/openclaw/issues/69012 | https://api.github.com/repos/openclaw/openclaw/issues/69012 | [Bug]: telegram native-approvals handler fails to start on every fresh gateway boot (WS handshake timeout) | ## Environment
- OpenClaw **2026.4.15** (`041266a`)
- Node **v24.14.1**
- Ubuntu 24.04.3 LTS, aarch64 (Hetzner CAX21)
- Telegram channel enabled (`channels.telegram.enabled: true`), gateway on `127.0.0.1:18789`, loopback-only
## Observed
On every fresh gateway start, the telegram native-approvals subsystem fails to ... | closed | completed | true | 2 | [] | [] | 2026-04-19T15:53:28Z | 2026-04-28T04:50:04Z | 2026-04-25T18:45:12Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | siddharthaagarwalofficial-ux | 266,795,318 | U_kgDOD-b5Ng | User | false |
openclaw/openclaw | 4,291,300,388 | I_kwDOQb6kR87_yAwk | 69,008 | https://github.com/openclaw/openclaw/issues/69008 | https://api.github.com/repos/openclaw/openclaw/issues/69008 | Telegram group-bound agent sessions default to session-only unless deliver is explicit | ## Summary
Agent runs bound to Telegram group sessions currently stay session-only by default unless `deliver: true` is explicitly set on the request.
In practice this creates a confusing failure mode:
- the run succeeds
- session/transcript history updates normally
- but no visible message is delivered to the bound ... | open | null | false | 1 | [] | [] | 2026-04-19T15:46:56Z | 2026-05-02T12:37:09Z | null | CONTRIBUTOR | null | 20260502T150030Z | 2026-05-02T15:00:30Z | glfruit | 3,939,154 | MDQ6VXNlcjM5MzkxNTQ= | User | false |
openclaw/openclaw | 4,291,340,328 | I_kwDOQb6kR87_yKgo | 69,013 | https://github.com/openclaw/openclaw/issues/69013 | https://api.github.com/repos/openclaw/openclaw/issues/69013 | [Bug]: Hybrid sessions.json schema (canonical + legacy positional keys) causes silent session degradation on 4.14, resolved by 4.15 migration | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Hybrid sessions.json schema on 4.14 — canonical keys (agent:<name>:<slot>) and legacy positional keys (bare entry-N) claimed the same logical slot with different sessionIds, causing silent session degradation that the 4.15 d... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-19T16:04:32Z | 2026-04-28T04:50:03Z | 2026-04-25T18:45:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Nettoc | 12,641,252 | MDQ6VXNlcjEyNjQxMjUy | User | false |
openclaw/openclaw | 4,291,285,129 | I_kwDOQb6kR87_x9CJ | 69,004 | https://github.com/openclaw/openclaw/issues/69004 | https://api.github.com/repos/openclaw/openclaw/issues/69004 | [Bug] Plugin install/update silently drops unrelated user config entries in agents.defaults.models | ## Description
Running `openclaw plugins install` or `openclaw plugins update` silently removes user-configured entries from `agents.defaults.models` that are unrelated to the plugin being installed.
## Steps to Reproduce
1. Add a custom model alias to `agents.defaults.models` in `openclaw.json`:
```json
"anth... | closed | completed | true | 2 | [] | [] | 2026-04-19T15:40:17Z | 2026-04-25T04:14:57Z | 2026-04-22T23:14:11Z | NONE | null | 20260425T060021Z | 2026-04-25T06:00:21Z | Adam-Researchh | 261,827,313 | U_kgDOD5sq8Q | User | false |
openclaw/openclaw | 4,291,287,868 | I_kwDOQb6kR87_x9s8 | 69,005 | https://github.com/openclaw/openclaw/issues/69005 | https://api.github.com/repos/openclaw/openclaw/issues/69005 | Slack react: already_reacted error should be silently ignored | When the bot tries to add a reaction that already exists on a message, the Slack API returns `already_reacted`. This is currently treated as a failure and surfaced to the user as a warning message.
**Expected:** `already_reacted` should be caught and treated as a no-op (the reaction is already there, which is the desi... | closed | completed | true | 5 | [] | [] | 2026-04-19T15:41:30Z | 2026-05-03T04:37:25Z | 2026-04-30T13:50:16Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | shipitsteven | 62,415,298 | MDQ6VXNlcjYyNDE1Mjk4 | User | false |
openclaw/openclaw | 4,291,351,411 | I_kwDOQb6kR87_yNNz | 69,014 | https://github.com/openclaw/openclaw/issues/69014 | https://api.github.com/repos/openclaw/openclaw/issues/69014 | sessions_send nested lane 串行化导致多智能体协作严重阻塞 | sessions_send nested lane 串行化导致多智能体协作严重阻塞
## 问题描述
当使用 sessions_send 作为多智能体间的通信协议时，所有跨 agent 的消息都在 nested lane 上串行排队（maxConcurrent=1），导致协作链路严重阻塞。
## 具体表现
- 消息投递延迟不可控（从几秒到几十分钟不等）
- 超时设置（timeout=0）只解决发送方不等待问题，接收方仍需在队列中排队
- 高优先级任务无法跳过队列立即处理
## 相关 Issues
- #45165 sessions_send nested lane maxConcurrent=1（未修复）
- #53527... | closed | completed | true | 1 | [] | [] | 2026-04-19T16:09:44Z | 2026-04-28T04:50:02Z | 2026-04-25T18:45:25Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | niemingzhao | 17,548,444 | MDQ6VXNlcjE3NTQ4NDQ0 | User | false |
openclaw/openclaw | 4,291,316,152 | I_kwDOQb6kR87_yEm4 | 69,011 | https://github.com/openclaw/openclaw/issues/69011 | https://api.github.com/repos/openclaw/openclaw/issues/69011 | [Bug]: bonjour advertiser stuck in "announcing" ~15s on every restart; watchdog triggers restart loop | ## Environment
- OpenClaw **2026.4.15** (`041266a`)
- Node **v24.14.1**
- Ubuntu 24.04.3 LTS, aarch64 (Hetzner CAX21)
- User-level systemd unit at `~/.config/systemd/user/openclaw-gateway.service`
## Observed
On every gateway restart, the bonjour advertiser reliably hangs in the `announcing` state for ~14–15 seconds... | closed | completed | true | 4 | [] | [] | 2026-04-19T15:53:27Z | 2026-04-28T04:45:26Z | 2026-04-26T00:13:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | siddharthaagarwalofficial-ux | 266,795,318 | U_kgDOD-b5Ng | User | false |
openclaw/openclaw | 4,291,300,345 | I_kwDOQb6kR87_yAv5 | 69,007 | https://github.com/openclaw/openclaw/issues/69007 | https://api.github.com/repos/openclaw/openclaw/issues/69007 | [Bug] Image recognition fails: chat content is empty (2013) | ## Bug Description
When using the built-in image recognition feature with MiniMax M2.7 model, image analysis fails with error:
`
Image model failed (minimax/MiniMax-M2.7): 400 {"type":"error","error":{"type":"invalid_request_error","message":"invalid params, chat content is empty (2013)"},"request_id":"..."}
`
## St... | closed | completed | true | 1 | [] | [] | 2026-04-19T15:46:55Z | 2026-04-22T04:30:37Z | 2026-04-19T16:22:12Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | Zakizel | 6,711,965 | MDQ6VXNlcjY3MTE5NjU= | User | false |
openclaw/openclaw | 4,178,559,803 | I_kwDOQb6kR875D8M7 | 58,354 | https://github.com/openclaw/openclaw/issues/58354 | https://api.github.com/repos/openclaw/openclaw/issues/58354 | [BlueBubbles] enrichGroupParticipantsFromContacts Zod default breaks all Control UI config saves (2026.3.28) | ### Summary
The BlueBubbles plugin's Zod schema declares `enrichGroupParticipantsFromContacts` with `.default(true)`, which auto-injects the key into the runtime config even when the user never set it. The Control UI reads this merged config, and any save attempt fails because the gateway-side config validator rejects... | closed | completed | true | 2 | [] | [] | 2026-03-31T11:58:36Z | 2026-04-22T04:30:51Z | 2026-04-19T16:13:23Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | felix-reiners | 12,480,537 | MDQ6VXNlcjEyNDgwNTM3 | User | false |
openclaw/openclaw | 4,180,567,003 | I_kwDOQb6kR875LmPb | 58,460 | https://github.com/openclaw/openclaw/issues/58460 | https://api.github.com/repos/openclaw/openclaw/issues/58460 | [BlueBubbles] Gateway injects Zod default for enrichGroupParticipantsFromContacts then rejects it on restart (config round-trip bug) | ### Summary
The gateway writes `enrichGroupParticipantsFromContacts: true` into `openclaw.json` during startup (via Zod `.default(true)` in the BlueBubbles config schema), then **rejects it as an "Unrecognized key" on the next startup, preventing the gateway from starting.** This creates a bricking loop: start → injec... | closed | completed | true | 1 | [] | [] | 2026-03-31T16:22:18Z | 2026-04-22T04:30:50Z | 2026-04-19T16:13:25Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | khuzad | 24,513,451 | MDQ6VXNlcjI0NTEzNDUx | User | false |
openclaw/openclaw | 4,181,185,832 | I_kwDOQb6kR875N9Uo | 58,507 | https://github.com/openclaw/openclaw/issues/58507 | https://api.github.com/repos/openclaw/openclaw/issues/58507 | Config validation rejects enrichGroupParticipantsFromContacts in 2026.3.28 | ### Bug type
Regression (feature announced but non-functional)
### Summary
The enrichGroupParticipantsFromContacts config key for BlueBubbles was announced in 2026.3.28 release notes but the config schema rejects it as Unrecognized key.
### Steps to reproduce
1. Add to openclaw.json:
{
"channels": {
"bluebubble... | closed | completed | true | 1 | [] | [] | 2026-03-31T17:51:06Z | 2026-04-22T04:30:50Z | 2026-04-19T16:13:27Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | rowbotik | 4,590,280 | MDQ6VXNlcjQ1OTAyODA= | User | false |
openclaw/openclaw | 4,165,795,792 | I_kwDOQb6kR874TP_Q | 57,295 | https://github.com/openclaw/openclaw/issues/57295 | https://api.github.com/repos/openclaw/openclaw/issues/57295 | Config validation rejects bluebubbles enrichGroupParticipantsFromContacts | ## Bug
After updating to OpenClaw 2026.3.24, saving any config change via the Control UI fails with:
```
GatewayRequestError: invalid config: channels.bluebubbles: Unrecognized key: "enrichGroupParticipantsFromContacts"
```
This blocks **all** UI config saves, not just bluebubbles changes.
## Root Cause
The bluebu... | closed | completed | true | 4 | [] | [] | 2026-03-29T21:03:15Z | 2026-04-22T04:30:49Z | 2026-04-19T16:13:29Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | jaserNo1 | 3,274,124 | MDQ6VXNlcjMyNzQxMjQ= | User | false |
openclaw/openclaw | 4,181,755,798 | I_kwDOQb6kR875QIeW | 58,538 | https://github.com/openclaw/openclaw/issues/58538 | https://api.github.com/repos/openclaw/openclaw/issues/58538 | Bug: `enrichGroupParticipantsFromContacts` schema inconsistency - config.schema.lookup shows valid but config validation rejects it | # Bug Report: `enrichGroupParticipantsFromContacts` Schema Inconsistency
## Summary
The `channels.bluebubbles.enrichGroupParticipantsFromContacts` configuration field shows inconsistent behavior between schema lookup and config validation.
## Description
### Observed Behavior
1. `openclaw config.schema.lookup chan... | closed | completed | true | 1 | [] | [] | 2026-03-31T19:23:24Z | 2026-04-22T04:30:48Z | 2026-04-19T16:13:30Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | frank7402 | 56,910,743 | MDQ6VXNlcjU2OTEwNzQz | User | false |
openclaw/openclaw | 4,165,119,879 | I_kwDOQb6kR874Qq-H | 57,181 | https://github.com/openclaw/openclaw/issues/57181 | https://api.github.com/repos/openclaw/openclaw/issues/57181 | SSRF guard blocks BlueBubbles plugin internal API calls to private IP | ## Summary
The new SSRF protection (introduced in 2026.3.28) blocks the BlueBubbles channel plugin's own HTTP calls to its configured server at a private IP address (`10.0.0.x:1234`). The guard doesn't distinguish between agent-initiated `web_fetch` (user-controllable, should be guarded) and plugin internal HTTP calls... | closed | completed | true | 1 | [] | [] | 2026-03-29T16:59:35Z | 2026-04-22T04:30:47Z | 2026-04-19T16:16:28Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | bobsahur-robot | 262,074,308 | U_kgDOD57vxA | User | false |
openclaw/openclaw | 4,203,802,080 | I_kwDOQb6kR876kO3g | 60,715 | https://github.com/openclaw/openclaw/issues/60715 | https://api.github.com/repos/openclaw/openclaw/issues/60715 | [Bug]: BlueBubbles health check fails on LAN/private serverUrl in 2026.4.2, while actual messaging still works | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading OpenClaw to 2026.4.2, openclaw health reports BlueBubbles as failed when channels.bluebubbles.serverUrl points to a private/LAN address (in my case http://172.31.100.1:1234).
The error is:
BlueBubbles: fail... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-04T06:17:37Z | 2026-04-22T04:30:46Z | 2026-04-19T16:16:30Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | D5EB6AEC | 121,235,952 | U_kgDOBznp8A | User | false |
openclaw/openclaw | 4,049,677,263 | I_kwDOQb6kR87xYSvP | 41,832 | https://github.com/openclaw/openclaw/issues/41832 | https://api.github.com/repos/openclaw/openclaw/issues/41832 | Bug: bluebubbles plugin fails to load — `Cannot find module '../../../src/infra/parse-finite-number.js'` | ## Reproduction
Run `openclaw status` with the bluebubbles channel plugin installed.
**Expected:** No errors, bluebubbles loads normally.
**Actual:** Error logged on every `openclaw status` call:
```
[plugins] bluebubbles failed to load from /opt/homebrew/lib/node_modules/openclaw/extensions/bluebubbles/index.ts:
... | closed | completed | true | 1 | [] | [] | 2026-03-10T06:17:21Z | 2026-04-22T04:30:45Z | 2026-04-19T16:16:32Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | bo-blue | 257,455,029 | U_kgDOD1hztQ | User | false |
openclaw/openclaw | 4,043,725,249 | I_kwDOQb6kR87xBlnB | 40,708 | https://github.com/openclaw/openclaw/issues/40708 | https://api.github.com/repos/openclaw/openclaw/issues/40708 | BlueBubbles plugin fails to load in 2026.3.7 — broken require path to parse-finite-number.js | ## Description
`openclaw doctor` shows the BlueBubbles plugin failing to load in 2026.3.7:
```
ERROR bluebubbles: failed to load plugin: Error: Cannot find module '../../../src/infra/parse-finite-number.js'
Require stack:
- /Users/openclaw/.npm-global/lib/node_modules/openclaw/extensions/bluebubbles/src/monitor-norma... | closed | completed | true | 1 | [] | [] | 2026-03-09T06:54:37Z | 2026-04-22T04:30:44Z | 2026-04-19T16:16:34Z | NONE | null | 20260422T060039Z | 2026-04-22T06:00:39Z | ceplenski-brainjuice | 263,270,512 | U_kgDOD7EwcA | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.