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 | 4,316,183,406 | I_kwDOQb6kR88AAAABAUO7bg | 70,613 | https://github.com/openclaw/openclaw/issues/70613 | https://api.github.com/repos/openclaw/openclaw/issues/70613 | [Bug]: After the upgrade to 2026.4.21, the program could not be launched. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The 2026.4.21 version upgrade failed again. After the upgrade, Openclaw couldn't run.
After installing npm install -g @larksuiteoapi/node-sdk, Openclaw could run, but WeChat could not be used. What a garbage software! It cra... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-23T13:14:25Z | 2026-04-26T04:30:09Z | 2026-04-23T16:47:13Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | kinlioncd | 83,751,473 | MDQ6VXNlcjgzNzUxNDcz | User | false |
openclaw/openclaw | 4,316,239,485 | I_kwDOQb6kR88AAAABAUSWfQ | 70,615 | https://github.com/openclaw/openclaw/issues/70615 | https://api.github.com/repos/openclaw/openclaw/issues/70615 | [Bug]: grammy module missing after every update - telegram extension breaks repeatedly | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
After every `openclaw update`, the grammy module is missing from dist/extensions/telegram/node_modules/, causing openclaw doctor to fail with "Cannot find module 'grammy'" and breaking the entire Telegram chann... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-23T13:22:28Z | 2026-04-26T04:27:50Z | 2026-04-23T16:44:11Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | tutumomo | 11,463,469 | MDQ6VXNlcjExNDYzNDY5 | User | false |
openclaw/openclaw | 4,316,268,352 | I_kwDOQb6kR88AAAABAUUHQA | 70,616 | https://github.com/openclaw/openclaw/issues/70616 | https://api.github.com/repos/openclaw/openclaw/issues/70616 | Feature Request: Add execution field to dreaming config with fallback model support | ## Description
Currently, the memory-core schema only supports these dreaming fields:
- enabled
- frequency
- timezone
- verboseLogging
- storage
- phases
The `execution` field is not in the allowed list, preventing users from configuring fallback or cheaper models for DREAMING.
## Request
Please add an `execution`... | open | null | false | 2 | [] | [] | 2026-04-23T13:26:11Z | 2026-05-02T13:08:17Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | jimwong1983 | 59,212,098 | MDQ6VXNlcjU5MjEyMDk4 | User | false |
openclaw/openclaw | 4,316,299,762 | I_kwDOQb6kR88AAAABAUWB8g | 70,617 | https://github.com/openclaw/openclaw/issues/70617 | https://api.github.com/repos/openclaw/openclaw/issues/70617 | Separate public message policy from native slash-command policy on Discord | ### Summary
Discord deployments often need different trust policies for public message replies vs native slash commands. Today those controls appear too tightly coupled, making it hard to allow safe public mention-based conversations while keeping slash/native command surfaces private, owner-only, or disabled.
### Pr... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-23T13:30:34Z | 2026-04-28T04:47:33Z | 2026-04-25T19:25:53Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | bukik07 | 25,246,816 | MDQ6VXNlcjI1MjQ2ODE2 | User | false |
openclaw/openclaw | 4,316,353,264 | I_kwDOQb6kR88AAAABAUZS8A | 70,618 | https://github.com/openclaw/openclaw/issues/70618 | https://api.github.com/repos/openclaw/openclaw/issues/70618 | Session crashes with Context overflow when large Project Context is injected at session start | ## Bug Description
Session crashes with the error: `Context overflow: estimated context size exceeds safe threshold during tool loop` when a large Project Context file (MEMORY.md, ~50KB+) is injected at session startup.
This also triggers aggressive session store rotation and cleanup, which can cause data loss.
## S... | closed | completed | true | 1 | [] | [] | 2026-04-23T13:38:27Z | 2026-04-29T04:41:59Z | 2026-04-26T07:16:55Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | mmhzlrj | 29,726,050 | MDQ6VXNlcjI5NzI2MDUw | User | false |
openclaw/openclaw | 4,316,389,592 | I_kwDOQb6kR88AAAABAUbg2A | 70,620 | https://github.com/openclaw/openclaw/issues/70620 | https://api.github.com/repos/openclaw/openclaw/issues/70620 | [Bug] Message sync issue between AI and WeChat - out of order and missing messages | ## Bug Description
There is a message synchronization issue between the AI assistant and WeChat client. Messages appear out of order and some intermediate messages are not displayed correctly.
## Steps to Reproduce
1. User sends a message to AI assistant via WeChat
2. AI assistant processes and sends multiple respon... | closed | not_planned | true | 5 | [] | [] | 2026-04-23T13:43:13Z | 2026-04-28T04:47:33Z | 2026-04-25T19:26:04Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | 15285701683 | 235,316,189 | U_kgDODgaj3Q | User | false |
openclaw/openclaw | 4,316,413,264 | I_kwDOQb6kR88AAAABAUc9UA | 70,622 | https://github.com/openclaw/openclaw/issues/70622 | https://api.github.com/repos/openclaw/openclaw/issues/70622 | [Bug]: assistant turn can enter repeated terminated state immediately after a tool error instead of reporting the failure | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
In a main chat session on OpenClaw 2026.4.21 using `openai-codex/gpt-5.4`, a real tool failure occurred during an active turn, and instead of returning a normal assistant message explaining the failure, the run fell into repeated `error=ter... | closed | completed | true | 2 | [] | [] | 2026-04-23T13:46:27Z | 2026-04-29T04:41:59Z | 2026-04-26T07:17:00Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | gchahrouri | 37,080,676 | MDQ6VXNlcjM3MDgwNjc2 | User | false |
openclaw/openclaw | 4,316,571,306 | I_kwDOQb6kR88AAAABAUmmqg | 70,627 | https://github.com/openclaw/openclaw/issues/70627 | https://api.github.com/repos/openclaw/openclaw/issues/70627 | [Bug]: openai-codex/gpt-5.4 works via infer model run, but embedded chat runs fail with "terminated" on 2026.4.21 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On OpenClaw 2026.4.21, `openai-codex/gpt-5.4` works when called directly with `openclaw infer model run`, and `openclaw models status --probe` reports the model/profile as healthy, but normal chat/embedded runs... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-23T14:08:46Z | 2026-04-26T04:29:14Z | 2026-04-24T01:40:49Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | LFArce | 122,791,295 | U_kgDOB1Glfw | User | false |
openclaw/openclaw | 4,316,576,970 | I_kwDOQb6kR88AAAABAUm8yg | 70,628 | https://github.com/openclaw/openclaw/issues/70628 | https://api.github.com/repos/openclaw/openclaw/issues/70628 | [Bug]: Telegram DM fabricates silent-reply chatter for no-visible-response turns | ## Summary
Telegram direct-message turns that end with **no visible final reply** can still emit a synthetic user-facing bubble like **"No added response from me."** instead of staying silent.
## Reproduction shape
This reproduces when:
- surface = `telegram`
- conversation type = direct / DM
- the turn ends without... | open | null | false | 6 | [] | [] | 2026-04-23T14:09:37Z | 2026-05-03T19:33:26Z | null | NONE | null | 20260503T210034Z | 2026-05-03T21:00:34Z | notskynet-bot | 261,878,757 | U_kgDOD5vz5Q | User | false |
openclaw/openclaw | 4,074,977,267 | I_kwDOQb6kR87y4zfz | 45,833 | https://github.com/openclaw/openclaw/issues/45833 | https://api.github.com/repos/openclaw/openclaw/issues/45833 | Post-update automatic status confirmation | ## Feature Request
### Problem
After running `gateway update.run`, the gateway restarts and the user has no automatic confirmation that the update succeeded. They have to manually check the version or ask the assistant.
### Proposed Solution
After a successful update + restart cycle, OpenClaw could automatically send... | closed | completed | true | 1 | [] | [] | 2026-03-14T07:19:04Z | 2026-04-26T04:30:21Z | 2026-04-23T14:10:01Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | pnaaberi | 31,892,171 | MDQ6VXNlcjMxODkyMTcx | User | false |
openclaw/openclaw | 4,316,452,785 | I_kwDOQb6kR88AAAABAUfXsQ | 70,623 | https://github.com/openclaw/openclaw/issues/70623 | https://api.github.com/repos/openclaw/openclaw/issues/70623 | [Bug]: Channel issues on Telegram and Discord in OpenClaw 2026.4.21 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Bug: Channel issues on Telegram and Discord in OpenClaw 2026.4.21
### Steps to reproduce
**For Telegram**
1. Start a Telegram session with the agent.
2. Have the agent produce any reply (plain text, tool out... | closed | not_planned | true | 5 | [
"bug",
"bug:behavior"
] | [] | 2026-04-23T13:51:56Z | 2026-05-03T04:35:11Z | 2026-04-30T19:53:44Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | L30nardo2026 | 277,026,336 | U_kgDOEIMWIA | User | false |
openclaw/openclaw | 4,316,634,553 | I_kwDOQb6kR88AAAABAUqduQ | 70,631 | https://github.com/openclaw/openclaw/issues/70631 | https://api.github.com/repos/openclaw/openclaw/issues/70631 | [Feature]: Add configurable Gateway CLI timeout and session isolation | ## Summary
Gateway-level timeout configuration for CLI agent calls to prevent main session contention issues.
## Problem
When the OpenClaw main agent session is under heavy load, incoming Gateway CLI calls queue up and timeout. Key symptoms:
1. **Main session contention** — All CLI calls must pass through the main ... | closed | completed | true | 3 | [] | [] | 2026-04-23T14:17:48Z | 2026-04-28T04:47:32Z | 2026-04-25T19:26:11Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tigerjam1980 | 129,712,626 | U_kgDOB7tB8g | User | false |
openclaw/openclaw | 4,316,770,523 | I_kwDOQb6kR88AAAABAUyw2w | 70,634 | https://github.com/openclaw/openclaw/issues/70634 | https://api.github.com/repos/openclaw/openclaw/issues/70634 | [Bug]: Human messages get starved in agent lane queues when agents communicate in loops | ### Summary
ISSUE:
When multiple agents are actively communicating with each other, human-originated messages get indefinitely delayed in the agent's FIFO lane queue. Agent auto-reply loops can sustain themselves faster than the queue drains, effectively locking out the human operator even with no queueAhead.
### Pr... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-23T14:41:03Z | 2026-05-02T09:30:30Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | ewrurwteurwU | 278,707,246 | U_kgDOEJy8Lg | User | false |
openclaw/openclaw | 4,316,772,202 | I_kwDOQb6kR88AAAABAUy3ag | 70,635 | https://github.com/openclaw/openclaw/issues/70635 | https://api.github.com/repos/openclaw/openclaw/issues/70635 | bug(sessions_list): only returns active session — BB/iMessage/idle sessions invisible | ## Summary
`sessions_list` MCP tool returns only the currently-active session, even though the on-disk session catalog contains hundreds of other sessions across multiple channels (BlueBubbles, iMessage, Slack, Telegram, cron, subagent, hook). Downstream consumers that rely on `sessions_list` (cron jobs building daily... | closed | completed | true | 1 | [] | [] | 2026-04-23T14:41:19Z | 2026-04-26T04:30:18Z | 2026-04-23T16:33:05Z | CONTRIBUTOR | null | 20260426T060034Z | 2026-04-26T06:00:34Z | markthebest12 | 6,416,874 | MDQ6VXNlcjY0MTY4NzQ= | User | false |
openclaw/openclaw | 4,087,470,043 | I_kwDOQb6kR87zodfb | 48,910 | https://github.com/openclaw/openclaw/issues/48910 | https://api.github.com/repos/openclaw/openclaw/issues/48910 | [Bug]: google-vertex provider broken with ADC auth: "<authenticated>" sentinel passed as API key + gaxios@7.1.3 incompatible with Node 24 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
When `GOOGLE_APPLICATION_CREDENTIALS` is set for ADC auth, the `google-vertex` provider passes the internal `"<authenticated>"` sentinel string as an API key to pi-ai, causing all Vertex AI requests to fail with `UNAUTHENTICATED`.
### Ste... | closed | completed | false | 3 | [
"bug",
"close:already-fixed",
"dedupe:child",
"bug:behavior"
] | [
"vincentkoc"
] | 2026-03-17T10:02:45Z | 2026-04-25T08:59:48Z | 2026-04-25T08:59:48Z | NONE | null | 20260425T090015Z | 2026-04-25T09:00:15Z | mhakby | 123,645,842 | U_kgDOB16vkg | User | false |
openclaw/openclaw | 4,316,376,122 | I_kwDOQb6kR88AAAABAUasOg | 70,619 | https://github.com/openclaw/openclaw/issues/70619 | https://api.github.com/repos/openclaw/openclaw/issues/70619 | [Bug]: isolated cron run session rows can inherit stale lifecycle fields (status/startedAt/endedAt) from prior runs | ## Summary
For `sessionTarget: "isolated"` cron jobs, a fresh per-run session row can inherit stale lifecycle metadata from older session state.
Observed stale fields include:
- `status`
- `startedAt`
- `endedAt`
This produces run rows that look finalized before the new run has actually finished, or rows that end u... | closed | not_planned | true | 4 | [
"dedupe:parent"
] | [] | 2026-04-23T13:41:26Z | 2026-04-27T04:31:56Z | 2026-04-25T01:21:42Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | f1yankees | 32,306,781 | MDQ6VXNlcjMyMzA2Nzgx | User | false |
openclaw/openclaw | 4,316,850,057 | I_kwDOQb6kR88AAAABAU3niQ | 70,637 | https://github.com/openclaw/openclaw/issues/70637 | https://api.github.com/repos/openclaw/openclaw/issues/70637 | [Bug]: Plugin dependencies (e.g. feishu plugin) are lost after npm pack + global installation | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to the latest openclaw, the issue started reappearing.
### Steps to reproduce
Update openclaw, start the onboarding process.
### Expected behavior
Openclaw onboarding should be seamless, without a bunch o... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-23T14:54:22Z | 2026-04-26T04:30:11Z | 2026-04-23T16:44:14Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | ctapnec | 11,669,622 | MDQ6VXNlcjExNjY5NjIy | User | false |
openclaw/openclaw | 4,236,573,932 | I_kwDOQb6kR878hPzs | 64,078 | https://github.com/openclaw/openclaw/issues/64078 | https://api.github.com/repos/openclaw/openclaw/issues/64078 | Slack: top-level channel messages with replyToMode="all" cause dual-session processing | # OpenClaw Bug Report: Dual Session Processing on Slack Channels with `replyToMode: "all"`
**Product:** OpenClaw
**Version tested:** 2026.4.5 (source code analysis), 2026.4.2 (live reproduction)
**Severity:** High (causes 2x–4x token spend, context pollution across senders)
**Date:** 2026-04-09
---
## 1. Sum... | open | null | false | 3 | [] | [] | 2026-04-10T04:30:36Z | 2026-05-02T11:36:16Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | carayim-dev | 258,936,089 | U_kgDOD28NGQ | User | false |
openclaw/openclaw | 4,316,895,402 | I_kwDOQb6kR88AAAABAU6Yqg | 70,640 | https://github.com/openclaw/openclaw/issues/70640 | https://api.github.com/repos/openclaw/openclaw/issues/70640 | Gateway crashes on Windows: ACL verification unavailable + UTF-8 BOM in secrets.json (v2026.4.21) | ## Bug Description
OpenClaw v2026.4.21 added a security check for `secrets.json` file permissions (ACL) on startup. On Windows, this check fails and causes the gateway to crash.
Additionally, `secrets.json` files with UTF-8 BOM cannot be parsed.
## Root Cause
### Issue 1: ACL verification unavailable on Windows
Th... | closed | completed | true | 0 | [] | [] | 2026-04-23T15:01:20Z | 2026-04-26T04:30:04Z | 2026-04-23T18:32:16Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | bxy3045134656 | 183,266,172 | U_kgDOCuxrfA | User | false |
openclaw/openclaw | 4,316,925,926 | I_kwDOQb6kR88AAAABAU8P5g | 70,641 | https://github.com/openclaw/openclaw/issues/70641 | https://api.github.com/repos/openclaw/openclaw/issues/70641 | [Bug]: Matrix native-approvals handler thrashes loopback gateway with handshake-timeout on multi-account installs | ## Summary
On a gateway running 2026.4.20 with 11 configured Matrix accounts and no explicit `execApprovals` config, the loopback gateway enters a persistent 1 Hz `native-approvals` / `handshake-timeout` reconnect storm that degrades `sessions_list` / `sessions_send` and artificially elevates gateway-process CPU; repr... | closed | completed | true | 1 | [] | [] | 2026-04-23T15:06:11Z | 2026-04-26T04:30:05Z | 2026-04-23T18:16:39Z | CONTRIBUTOR | null | 20260426T060034Z | 2026-04-26T06:00:34Z | lukeboyett | 46,942,646 | MDQ6VXNlcjQ2OTQyNjQ2 | User | false |
openclaw/openclaw | 4,316,931,305 | I_kwDOQb6kR88AAAABAU8k6Q | 70,642 | https://github.com/openclaw/openclaw/issues/70642 | https://api.github.com/repos/openclaw/openclaw/issues/70642 | [Bug]: OpenAI gpt-image-2 image edits abort in OpenClaw for some multi-ref requests that succeed via direct Images API | ## Bug type
Behavior bug
## Summary
`image_generate` with `openai/gpt-image-2` can abort on complex reference-image edit requests even though the same request succeeds against OpenAI's official Images API.
The failing OpenClaw path appears to submit `/v1/images/edits` as JSON with data URLs, while the official workin... | closed | completed | true | 1 | [] | [] | 2026-04-23T15:07:03Z | 2026-04-26T04:29:59Z | 2026-04-23T19:49:37Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | dashhuang | 125,997,726 | U_kgDOB4KSng | User | false |
openclaw/openclaw | 4,316,969,513 | I_kwDOQb6kR88AAAABAU-6KQ | 70,643 | https://github.com/openclaw/openclaw/issues/70643 | https://api.github.com/repos/openclaw/openclaw/issues/70643 | Concurrent writes to openclaw.json produce invalid (concatenated) JSON; gateway refuses config until auto-restore | ## Summary
When two OpenClaw processes write to `~/.openclaw/openclaw.json` around the same time, the resulting file can contain two concatenated top-level JSON objects rather than a merged one. The gateway then fails to load config, and the UI surfaces this to the user as **"agent failed before reply"** on the next me... | open | null | false | 1 | [] | [] | 2026-04-23T15:13:13Z | 2026-05-02T09:00:14Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | hexorcist404 | 180,866,933 | U_kgDOCsfPdQ | User | false |
openclaw/openclaw | 4,316,970,418 | I_kwDOQb6kR88AAAABAU-9sg | 70,644 | https://github.com/openclaw/openclaw/issues/70644 | https://api.github.com/repos/openclaw/openclaw/issues/70644 | bug(signal): heartbeat/session final replies can stay internal instead of reaching Signal | ## Problem
In live Signal DM use, some assistant follow-up/final messages are visible in the internal session transcript (`agent:main:signal:direct:+1...`) but do not actually arrive to the human in Signal.
Observed symptoms:
- the human receives the early "starting..." receipt but not the final report
- the interna... | closed | completed | true | 1 | [] | [] | 2026-04-23T15:13:21Z | 2026-04-29T04:41:58Z | 2026-04-26T07:17:04Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | howardclaw26 | 273,194,802 | U_kgDOEEifMg | User | false |
openclaw/openclaw | 4,168,175,255 | I_kwDOQb6kR874cU6X | 57,555 | https://github.com/openclaw/openclaw/issues/57555 | https://api.github.com/repos/openclaw/openclaw/issues/57555 | Built-in skill 'summarize' only supports brew install, no feedback when enable fails | ## Bug type
UX / Platform support
## Summary
The built-in `summarize` skill only provides a brew install method (`steipete/tap/summarize`), which makes it unusable in Docker/Linux environments. Additionally, when clicking the enable toggle in the Control UI skills page, there is no error message or feedback — the to... | open | null | false | 2 | [] | [] | 2026-03-30T07:18:12Z | 2026-05-02T04:19:07Z | null | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | gucasbrg | 3,182,658 | MDQ6VXNlcjMxODI2NTg= | User | false |
openclaw/openclaw | 4,317,085,433 | I_kwDOQb6kR88AAAABAVF--Q | 70,654 | https://github.com/openclaw/openclaw/issues/70654 | https://api.github.com/repos/openclaw/openclaw/issues/70654 | GPT-5.4 Responses API: encrypted_content item_id mismatch after model switch or compaction | ## Bug Description
When switching to GPT-5.4 (Responses API) mid-session after using a Claude model, the API rejects requests with:
```
400 The encrypted content for item rs_XXXX could not be verified.
Reason: Encrypted content item_id did not match the target item id.
```
This happens consistently and the session b... | closed | completed | true | 6 | [] | [] | 2026-04-23T15:31:17Z | 2026-04-28T04:46:18Z | 2026-04-25T19:26:24Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jakepresent | 31,285,242 | MDQ6VXNlcjMxMjg1MjQy | User | false |
openclaw/openclaw | 4,317,099,591 | I_kwDOQb6kR88AAAABAVG2Rw | 70,658 | https://github.com/openclaw/openclaw/issues/70658 | https://api.github.com/repos/openclaw/openclaw/issues/70658 | [Bug]: Plugin config validator runs on undefined config during CLI preflight/subcommand paths, producing false-positive 'invalid config' errors | ## Summary
The OpenClaw plugin loader calls `validatePluginConfig` with `value: undefined` during preflight / subcommand code paths, producing non-actionable "invalid config" errors for plugins whose config is valid and working at runtime. The plugin later activates successfully via the normal gateway path with the re... | closed | not_planned | true | 1 | [] | [] | 2026-04-23T15:33:31Z | 2026-04-29T04:33:59Z | 2026-04-26T21:31:47Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | iam26phi | 154,225,143 | U_kgDOCTFJ9w | User | false |
openclaw/openclaw | 4,317,101,098 | I_kwDOQb6kR88AAAABAVG8Kg | 70,659 | https://github.com/openclaw/openclaw/issues/70659 | https://api.github.com/repos/openclaw/openclaw/issues/70659 | QA-lab extension test fails on current main: aborts in-flight runner model catalog | ## Summary
Current `main` reproduces the QA-lab extension failure seen on PR #66293 after the PR was refreshed. The failure is outside PR #66293's five-file consumer/QC overlay.
## Current-main proof
- Repo/base: `openclaw/openclaw` `main`
- Commit tested: `fda09c4806cc261e00337926f1aa5e7f37d2f1f7`
- PR reference: #... | closed | completed | true | 1 | [] | [] | 2026-04-23T15:33:44Z | 2026-04-29T04:36:03Z | 2026-04-26T12:21:07Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | BananaAccurate | 225,479,766 | U_kgDODXCMVg | User | false |
openclaw/openclaw | 4,317,145,646 | I_kwDOQb6kR88AAAABAVJqLg | 70,664 | https://github.com/openclaw/openclaw/issues/70664 | https://api.github.com/repos/openclaw/openclaw/issues/70664 | [Bug]: [macOS] findGatewayPidsOnPortSync drops all PIDs due to lsof p_comm vs argv[0] mismatch |
## Bug type
Behavior bug (correctness) — silent no-op in a safety-critical cleanup path, observable as launchd respawn churn and log-file bloat.
## Summary
On macOS, `findGatewayPidsOnPortSync` in `src/infra/restart-stale-pids.ts` never recognizes live openclaw gateway processes. `parsePidsFromLsofOutput` filters c... | closed | completed | false | 2 | [] | [] | 2026-04-23T15:40:49Z | 2026-05-04T01:49:59Z | 2026-05-04T01:49:59Z | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | mmartoccia | 4,631,577 | MDQ6VXNlcjQ2MzE1Nzc= | User | false |
openclaw/openclaw | 4,317,116,676 | I_kwDOQb6kR88AAAABAVH5BA | 70,661 | https://github.com/openclaw/openclaw/issues/70661 | https://api.github.com/repos/openclaw/openclaw/issues/70661 | [Bug]: bundled plugin runtime-deps don't link openclaw, so Discord/Telegram/etc. crash on 2026.4.22 with 'Cannot find package openclaw' | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No — workaround below.
### Summary
After upgrading to `2026.4.22`, **bundled** channel plugins (Discord, probably others that have moved to external `openclaw/plugin-sdk/*` imports in this release) crash at startup with:
```
Cannot find p... | closed | not_planned | true | 1 | [] | [] | 2026-04-23T15:36:12Z | 2026-04-26T04:30:19Z | 2026-04-23T15:42:31Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | MotherSphere | 235,695,427 | U_kgDODgxtQw | User | false |
openclaw/openclaw | 4,317,207,323 | I_kwDOQb6kR88AAAABAVNbGw | 70,666 | https://github.com/openclaw/openclaw/issues/70666 | https://api.github.com/repos/openclaw/openclaw/issues/70666 | reminder-guard appends false "did not schedule a reminder" disclaimer when reminder is written via plugin tool | ## Summary
When an agent uses a registered plugin tool to schedule a reminder, the runtime reminder-guard appends the following disclaimer to the agent reply:
> "Note: I did not schedule a reminder in this turn, so this will not trigger automatically."
The disclaimer is false. The reminder IS scheduled and fires cor... | closed | not_planned | true | 2 | [] | [] | 2026-04-23T15:51:25Z | 2026-04-29T04:36:02Z | 2026-04-26T12:21:11Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | fstrasz | 25,868,006 | MDQ6VXNlcjI1ODY4MDA2 | User | false |
openclaw/openclaw | 4,223,516,207 | I_kwDOQb6kR877vb4v | 63,076 | https://github.com/openclaw/openclaw/issues/63076 | https://api.github.com/repos/openclaw/openclaw/issues/63076 | [Bug]: GPT-5 personality overlay silently skipped for aggregator providers (OpenRouter, OpenCode) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The GPT-5 friendly personality overlay (interaction style, output contract, execution bias) only exists in the OpenAI provider extension. Aggregator providers like OpenRouter and OpenCode that also serve GPT-5 ... | closed | completed | true | 2 | [] | [] | 2026-04-08T09:32:16Z | 2026-04-26T04:30:19Z | 2026-04-23T15:59:37Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | ferran9908 | 42,429,912 | MDQ6VXNlcjQyNDI5OTEy | User | false |
openclaw/openclaw | 4,317,116,447 | I_kwDOQb6kR88AAAABAVH4Hw | 70,660 | https://github.com/openclaw/openclaw/issues/70660 | https://api.github.com/repos/openclaw/openclaw/issues/70660 | [Bug]: OpenCLAW - Stop button disappears during request and cannot terminate running processes if gateway stops | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The Stop button disappears during an active request, preventing termination of running processes when the gateway stops.
### Steps to reproduce
Launch OpenCLAW and start a code-related task that runs continuously.
While th... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-04-23T15:36:09Z | 2026-04-26T04:29:54Z | 2026-04-23T18:12:41Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | alexandre-leng | 72,550,137 | MDQ6VXNlcjcyNTUwMTM3 | User | false |
openclaw/openclaw | 4,317,219,246 | I_kwDOQb6kR88AAAABAVOJrg | 70,668 | https://github.com/openclaw/openclaw/issues/70668 | https://api.github.com/repos/openclaw/openclaw/issues/70668 | [Feature]: Support disabling heartbeat via config | ### Summary
Currently the heartbeat feature cannot be disabled via `openclaw.json`.
Setting `"heartbeat": false` results in a config validation error and the
Gateway falls back to the last-known-good config (`"every": "30m"`).
Users who want to disable heartbeat entirely have no valid config value to do so.
### Prob... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-04-23T15:53:20Z | 2026-04-28T04:47:30Z | 2026-04-25T19:26:31Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | dovelllll | 265,748,522 | U_kgDOD9cAKg | User | false |
openclaw/openclaw | 4,316,977,900 | I_kwDOQb6kR88AAAABAU_a7A | 70,645 | https://github.com/openclaw/openclaw/issues/70645 | https://api.github.com/repos/openclaw/openclaw/issues/70645 | [SECURITY] Reasoning leak in WebSocket outbound stream when approval-pending triggers on channel plugin | ## Summary
When an approval-pending prompt is triggered by the Gateway on a channel plugin (observed on Telegram), the LLM's raw reasoning prose is emitted to the channel alongside the approval prompt, exposing planning monologue, internal UUIDs, current working directory, pending command details, and potentially cred... | closed | completed | true | 2 | [] | [] | 2026-04-23T15:14:28Z | 2026-04-28T04:47:31Z | 2026-04-25T19:26:17Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | David-CKS | 238,664,549 | U_kgDODjm7ZQ | User | false |
openclaw/openclaw | 4,317,306,798 | I_kwDOQb6kR88AAAABAVTfrg | 70,674 | https://github.com/openclaw/openclaw/issues/70674 | https://api.github.com/repos/openclaw/openclaw/issues/70674 | [Bug]: No images shown in the web ui chat | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
No images shown in the web ui chat: neither what I send, nor that which the agent sends to me. This happened before in the issue #15437 but that issue was closed as "completed". I am running the very latest Open Claw
```
$... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-23T16:08:49Z | 2026-04-26T04:30:01Z | 2026-04-23T19:17:43Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | tigran123 | 1,933,903 | MDQ6VXNlcjE5MzM5MDM= | User | false |
openclaw/openclaw | 4,317,373,252 | I_kwDOQb6kR88AAAABAVXjRA | 70,678 | https://github.com/openclaw/openclaw/issues/70678 | https://api.github.com/repos/openclaw/openclaw/issues/70678 | [Bug]: WhatsApp channel force-reconnects every 30 minutes on quiet-device sessions | ## Bug type
Existing bug
## Beta release blocker
No
## Summary
On a WhatsApp account linked to OpenClaw as a secondary device, the gateway force-reconnects the WhatsApp socket every 30 minutes even though the socket is healthy. I noticed it since `openclaw@2026.4.9` but may have existed prior.
## Steps to reprodu... | closed | completed | true | 4 | [
"bug",
"channel: whatsapp-web",
"dedupe:parent"
] | [] | 2026-04-23T16:22:45Z | 2026-04-29T04:34:56Z | 2026-04-26T16:51:42Z | CONTRIBUTOR | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jasonftl | 175,325,999 | U_kgDOCnNDLw | User | false |
openclaw/openclaw | 4,317,374,812 | I_kwDOQb6kR88AAAABAVXpXA | 70,679 | https://github.com/openclaw/openclaw/issues/70679 | https://api.github.com/repos/openclaw/openclaw/issues/70679 | Bug: claude-cli usage accumulator double-counts cache_read_input_tokens across tool-loop iterations | # Bug: claude-cli usage accumulator double-counts `cache_read_input_tokens` across tool-loop iterations, inflating TUI token display and triggering spurious context overflow
## Summary
When a `claude-cli` provider session does multiple tool-use iterations inside a single user turn (e.g. `Read`, `Edit`, web fetches), ... | closed | completed | true | 2 | [] | [] | 2026-04-23T16:23:06Z | 2026-04-28T04:47:30Z | 2026-04-25T19:26:37Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | yongsinzhang-commits | 271,240,752 | U_kgDOECrOMA | User | false |
openclaw/openclaw | 4,317,391,442 | I_kwDOQb6kR88AAAABAVYqUg | 70,680 | https://github.com/openclaw/openclaw/issues/70680 | https://api.github.com/repos/openclaw/openclaw/issues/70680 | doctor falsely flags live *.trajectory.jsonl files as orphan transcripts | ## Summary
On OpenClaw `2026.4.22`, `openclaw doctor --non-interactive` falsely reports live `*.trajectory.jsonl` session artifacts as orphan transcript files and suggests archiving them via `--fix`, even when the owning sessions are still present in `sessions.json` and one may have a live lock.
## Observed behavior
-... | closed | completed | false | 2 | [] | [] | 2026-04-23T16:26:33Z | 2026-05-03T20:07:53Z | 2026-05-03T20:07:53Z | NONE | null | 20260503T210034Z | 2026-05-03T21:00:34Z | editbyjunior | 4,683,787 | MDQ6VXNlcjQ2ODM3ODc= | User | false |
openclaw/openclaw | 4,317,420,279 | I_kwDOQb6kR88AAAABAVaa9w | 70,683 | https://github.com/openclaw/openclaw/issues/70683 | https://api.github.com/repos/openclaw/openclaw/issues/70683 | [Bug]: Feishu channel fails to start with "Cannot find module openclaw/plugin-sdk/text-runtime" in 2026.4.22 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading from 2026.4.15 to 2026.4.22, both Feishu channel accounts immediately exit on startup with Cannot find
module openclaw/plugin-sdk/text-runtime and enter a permanent auto-restart loop.
### Steps to reprodu... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-23T16:32:49Z | 2026-04-26T04:30:17Z | 2026-04-23T16:34:56Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | ASNFalcon | 24,579,332 | MDQ6VXNlcjI0NTc5MzMy | User | false |
openclaw/openclaw | 4,317,451,824 | I_kwDOQb6kR88AAAABAVcWMA | 70,686 | https://github.com/openclaw/openclaw/issues/70686 | https://api.github.com/repos/openclaw/openclaw/issues/70686 | [Bug]: Compaction checkpoint .jsonl files double-count session cost in Usage dashboard | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Compaction checkpoint files (`{sessionId}.checkpoint.{checkpointId}.jsonl`) are treated as independent session transcripts by the Usage cost calculator, causing each assistant turn to be counted **twice** (once from the primary session file... | closed | completed | true | 1 | [] | [] | 2026-04-23T16:39:54Z | 2026-04-28T04:45:45Z | 2026-04-25T23:23:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | nchava79 | 121,129,672 | U_kgDOBzhKyA | User | false |
openclaw/openclaw | 4,317,296,208 | I_kwDOQb6kR88AAAABAVS2UA | 70,671 | https://github.com/openclaw/openclaw/issues/70671 | https://api.github.com/repos/openclaw/openclaw/issues/70671 | [Bug]: 2026.4.21 upgrade leaves gateway in restart loop — feishu/telegram SDK deps never auto-install because config validation fails first | ## Summary
After upgrading from `2026.4.15` to `2026.4.21` via `npm install -g openclaw@2026.4.21`, the gateway cannot start. systemd restarts it every ~25s in a loop. Root cause is a chicken-and-egg between config validation and the plugin runtime-dep auto-installer: validation fails on missing modules before the aut... | closed | completed | true | 1 | [] | [] | 2026-04-23T16:06:38Z | 2026-04-26T04:30:11Z | 2026-04-23T16:44:16Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | liu366811-hue | 244,927,607 | U_kgDODplMdw | User | false |
openclaw/openclaw | 4,317,492,731 | I_kwDOQb6kR88AAAABAVe1-w | 70,687 | https://github.com/openclaw/openclaw/issues/70687 | https://api.github.com/repos/openclaw/openclaw/issues/70687 | Bug: Scope upgrade pending approval error after upgrade to 2026.4.21 | # Bug Report: Scope Upgrade Pending Approval Error After Upgrade to 2026.4.21
## Summary
After upgrading OpenClaw from 2026.4.15 to 2026.4.21, spawn subagent and native approval operations fail with "scope upgrade pending approval" error.
## Environment
- OpenClaw Version: 2026.4.21
- Node Version: v25.9.0
- Platform... | closed | completed | true | 4 | [] | [] | 2026-04-23T16:49:17Z | 2026-04-29T04:29:26Z | 2026-04-26T10:51:32Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | superzmenai01 | 262,949,954 | U_kgDOD6xMQg | User | false |
openclaw/openclaw | 4,317,502,107 | I_kwDOQb6kR88AAAABAVfamw | 70,688 | https://github.com/openclaw/openclaw/issues/70688 | https://api.github.com/repos/openclaw/openclaw/issues/70688 | /status shows codex/* auth as unknown even when Codex CLI auth is available | ## Summary
When a session is using the native `codex/*` provider and valid Codex CLI auth is present in `CODEX_HOME` / `~/.codex/auth.json`, `/status` still reports:
- `🔑 unknown`
This is misleading because the runtime can still be correctly authenticated through the Codex CLI auth path.
## Why this looks like a b... | closed | completed | true | 0 | [] | [] | 2026-04-23T16:51:30Z | 2026-04-27T05:09:35Z | 2026-04-24T05:47:31Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | jb510 | 801,241 | MDQ6VXNlcjgwMTI0MQ== | User | false |
openclaw/openclaw | 4,317,511,235 | I_kwDOQb6kR88AAAABAVf-Qw | 70,690 | https://github.com/openclaw/openclaw/issues/70690 | https://api.github.com/repos/openclaw/openclaw/issues/70690 | [Feature]: Python version of Claude Code-like CLI that supports all LLMs | ### Summary
[Feature]: Python version of Claude Code-like CLI that supports all LLMs
### Problem to solve
As Claude Code is trying to block all access to the 3-rd party agents, once including OpenClaw, I am wondering if OpenClaw could add a Claude Code-like CLI that supports all LLMs. In particular I am looking f... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-23T16:53:38Z | 2026-04-28T04:47:29Z | 2026-04-25T19:26:44Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | cloudpigg101 | 278,755,440 | U_kgDOEJ14cA | User | false |
openclaw/openclaw | 4,317,533,141 | I_kwDOQb6kR88AAAABAVhT1Q | 70,692 | https://github.com/openclaw/openclaw/issues/70692 | https://api.github.com/repos/openclaw/openclaw/issues/70692 | [Bug]: agents.list[].contextTokens is read at runtime but stripped on hot reload — /status always falls back to 200k | ## Bug type
Behavior bug (incorrect output/state without crash)
## Beta release blocker
No
## Summary
A per-agent `contextTokens` override is read by the runtime in at least eight locations (including `/status` rendering), but the JSON Schema at `agents.list[]` does not declare it. The config watcher strips it on ... | closed | completed | true | 0 | [] | [] | 2026-04-23T16:58:20Z | 2026-04-27T04:35:14Z | 2026-04-25T03:47:49Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | yuxiaoyang2007-prog | 258,562,502 | U_kgDOD2lZxg | User | false |
openclaw/openclaw | 4,317,537,749 | I_kwDOQb6kR88AAAABAVhl1Q | 70,694 | https://github.com/openclaw/openclaw/issues/70694 | https://api.github.com/repos/openclaw/openclaw/issues/70694 | sessions_spawn: task prompt appears truncated/lost for some subagent runs (8 input tokens for 2k-char task) | ## Summary
When spawning subagents via `sessions_spawn` (`runtime: "subagent"`, `mode: "run"`), the task prompt appears to be truncated or not delivered to the child process in some cases. The child completes "successfully" but produces output that is clearly not based on the task description — either hallucinated wor... | closed | not_planned | true | 1 | [] | [] | 2026-04-23T16:59:18Z | 2026-05-03T04:34:23Z | 2026-05-01T01:27:36Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | dodge1218 | 235,199,066 | U_kgDODgTaWg | User | false |
openclaw/openclaw | 4,317,590,738 | I_kwDOQb6kR88AAAABAVk00g | 70,696 | https://github.com/openclaw/openclaw/issues/70696 | https://api.github.com/repos/openclaw/openclaw/issues/70696 | [Bug]: configure wizard Model section silently overwrites agents.defaults.model.primary to openrouter/auto | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Severity: P1 (Critical)
Environment:
OpenClaw: 2026.4.22 (00bd2cf)
OS: Linux 6.6.87.2-microsoft-standard-WSL2 (x64)
Node: 22.22.2
Description:
The `openclaw configure --section model` wizard has a critical bug:... | closed | completed | true | 0 | [
"bug",
"bug:behavior"
] | [] | 2026-04-23T17:09:30Z | 2026-04-27T05:06:36Z | 2026-04-24T18:55:21Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | wherewolf87 | 132,775,537 | U_kgDOB-n-cQ | User | false |
openclaw/openclaw | 4,317,283,156 | I_kwDOQb6kR88AAAABAVSDVA | 70,670 | https://github.com/openclaw/openclaw/issues/70670 | https://api.github.com/repos/openclaw/openclaw/issues/70670 | Bug: Infinite edit loop when editing context-loaded files | ## Bug Description
When an agent edits a file that is also loaded into the current context prompt (e.g. MEMORY.md, SOUL.md), the updated file is automatically reloaded into the context after the first edit. If the agent is still responding in the same turn, it sees the updated file and may issue another edit — creatin... | closed | completed | true | 2 | [] | [] | 2026-04-23T16:04:00Z | 2026-04-29T04:41:57Z | 2026-04-26T07:17:09Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | AS76 | 22,101,267 | MDQ6VXNlcjIyMTAxMjY3 | User | false |
openclaw/openclaw | 4,317,603,197 | I_kwDOQb6kR88AAAABAVllfQ | 70,697 | https://github.com/openclaw/openclaw/issues/70697 | https://api.github.com/repos/openclaw/openclaw/issues/70697 | `openclaw gateway restart` can orphan the gateway process in Docker container | ## Bug Description
In a Docker container environment on a Hostinger VPS (Debian 13), `openclaw gateway restart` does not behave reliably. The observed failure modes are:
1. `openclaw gateway restart` may orphan the gateway process — the service appears to stop but the process remains running, or fails to restart clea... | closed | completed | true | 1 | [] | [] | 2026-04-23T17:11:36Z | 2026-04-29T04:36:57Z | 2026-04-26T10:53:25Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | AS76 | 22,101,267 | MDQ6VXNlcjIyMTAxMjY3 | User | false |
openclaw/openclaw | 4,317,629,890 | I_kwDOQb6kR88AAAABAVnNwg | 70,699 | https://github.com/openclaw/openclaw/issues/70699 | https://api.github.com/repos/openclaw/openclaw/issues/70699 | openclaw acp accepts input but never returns output; ACP one-shot runs time out with no first event | ## Summary
After updating to `OpenClaw 2026.4.22 (00bd2cf)`, the `openclaw` ACP harness still appears broken.
The ACP bridge starts and reports ready, but:
- `openclaw acp client` accepts input and then hangs without responding
- ACP one-shot runs via the embedded ACP runtime produce no output and time out after 120s
... | closed | completed | true | 6 | [] | [] | 2026-04-23T17:16:33Z | 2026-04-28T04:47:28Z | 2026-04-25T19:26:50Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Filipenkoman | 249,626,028 | U_kgDODuD9rA | User | false |
openclaw/openclaw | 4,317,648,866 | I_kwDOQb6kR88AAAABAVoX4g | 70,700 | https://github.com/openclaw/openclaw/issues/70700 | https://api.github.com/repos/openclaw/openclaw/issues/70700 | [Bug]: Control UI inconsistently handles workspace-relative agent avatars even when /avatar/{agentId} resolves correctly | ## Bug type
Behavior bug (incorrect output/state without crash)
## Summary
Control UI / WebChat does not consistently handle workspace-relative `identity.avatar` values from `IDENTITY.md`.
A workspace-relative avatar such as:
```md
- **Avatar:** avatars/懂王.jpg
```
is documented/schematically valid and is correctl... | closed | completed | true | 5 | [] | [] | 2026-04-23T17:20:29Z | 2026-04-28T04:30:26Z | 2026-04-25T19:26:57Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | stongxxx | 194,768,251 | U_kgDOC5vtew | User | false |
openclaw/openclaw | 4,317,657,567 | I_kwDOQb6kR88AAAABAVo53w | 70,702 | https://github.com/openclaw/openclaw/issues/70702 | https://api.github.com/repos/openclaw/openclaw/issues/70702 | Control UI: assistant messages don't render until page reload | ## Summary
In the Control UI (web), assistant responses fail to appear in the chat after they're sent. Reloading the page reveals them.
## Environment
- OpenClaw: 2026.4.22 (00bd2cf)
- OS: macOS (Darwin 25.4.0)
- Channel: Control UI (web)
- Browser: Safari Version 26.4 (21624.1.16.11.4)
## Repro
1. Open Control UI
2.... | closed | completed | true | 2 | [] | [] | 2026-04-23T17:22:03Z | 2026-04-28T04:47:27Z | 2026-04-25T19:27:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | michaelfeiertag | 68,660,008 | MDQ6VXNlcjY4NjYwMDA4 | User | false |
openclaw/openclaw | 4,317,685,103 | I_kwDOQb6kR88AAAABAVqlbw | 70,703 | https://github.com/openclaw/openclaw/issues/70703 | https://api.github.com/repos/openclaw/openclaw/issues/70703 | [Feature]: support gpt-image-2 in image_generate via Codex OAuth | ## summary
OpenClaw appears to support `gpt-image-2` in the shared `image_generate` capability now, but only through the OpenAI API-key path.
Today the bundled OpenAI image provider defaults to `gpt-image-2`, while provider inventory reports that `openai/*` image generation requires `OPENAI_API_KEY`. In contrast, Ope... | closed | completed | true | 3 | [] | [] | 2026-04-23T17:27:25Z | 2026-04-26T04:28:35Z | 2026-04-23T20:35:54Z | CONTRIBUTOR | null | 20260426T060034Z | 2026-04-26T06:00:34Z | barronlroth | 5,567,884 | MDQ6VXNlcjU1Njc4ODQ= | User | false |
openclaw/openclaw | 4,162,464,369 | I_kwDOQb6kR874Gipx | 56,794 | https://github.com/openclaw/openclaw/issues/56794 | https://api.github.com/repos/openclaw/openclaw/issues/56794 | Bundled channel schema overrides external plugin schema even when bundled plugin is disabled | ## Bug Description
When an external plugin (e.g., `@larksuite/openclaw-lark`) registers the same channel id (`feishu`) as a bundled channel extension, and the bundled extension is explicitly disabled via `plugins.entries.feishu.enabled: false`, the bundled extension's config schema still takes priority during validati... | closed | completed | true | 2 | [] | [] | 2026-03-29T04:18:12Z | 2026-04-30T04:32:38Z | 2026-04-28T01:16:48Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | wuji-tech-dev | 260,395,371 | U_kgDOD4VRaw | User | false |
openclaw/openclaw | 4,084,708,533 | I_kwDOQb6kR87zd7S1 | 48,483 | https://github.com/openclaw/openclaw/issues/48483 | https://api.github.com/repos/openclaw/openclaw/issues/48483 | Dashboard (Control UI) model dropdown ignores models.mode="replace" — shows all 600+ built-in models | ## Summary
When `models.mode` is set to `"replace"` in `openclaw.json`, `openclaw models list` correctly shows only the configured providers. The Control UI chat model dropdown does not: it still shows the full built-in registry.
## Current Status
Reproduced again on the latest upstream `main` on April 13, 2026.
Fr... | closed | completed | true | 5 | [] | [] | 2026-03-16T20:46:20Z | 2026-04-30T04:31:55Z | 2026-04-28T04:23:56Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | adpena | 10,686,801 | MDQ6VXNlcjEwNjg2ODAx | User | false |
openclaw/openclaw | 4,317,655,210 | I_kwDOQb6kR88AAAABAVowqg | 70,701 | https://github.com/openclaw/openclaw/issues/70701 | https://api.github.com/repos/openclaw/openclaw/issues/70701 | v2026.4.22: Telegram channels crash on startup — Cannot find package 'openclaw' in plugin-runtime-deps | ## Bug Report
**Version:** OpenClaw 2026.4.22 (00bd2cf)
**OS:** macOS Darwin 25.4.0 (arm64)
**Node:** v24.13.0
## What happened
After updating to 2026.4.22, **all Telegram bot accounts** (default, and multi-account agents) immediately crash on startup with a crash-restart loop. None of the bots receive or respond to... | closed | completed | true | 4 | [] | [] | 2026-04-23T17:21:37Z | 2026-04-26T04:29:23Z | 2026-04-24T01:19:46Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | Mengchee118 | 99,036,697 | U_kgDOBecuGQ | User | false |
openclaw/openclaw | 4,317,853,327 | I_kwDOQb6kR88AAAABAV02jw | 70,717 | https://github.com/openclaw/openclaw/issues/70717 | https://api.github.com/repos/openclaw/openclaw/issues/70717 | Gateway RSS regression on 2026.4.15 — fresh cold-start baseline 700MB+ on macOS ARM64, steady climb regardless of workload | ---
title: "Gateway RSS regression on 2026.4.15 — fresh cold-start baseline 700MB+ on macOS ARM64, steady climb regardless of workload"
labels: bug, regression, gateway, memory
---
## Summary
After upgrading from 2026.4.1 to 2026.4.15, gateway RSS baseline climbed from ~400MB fresh-start to 700MB+ fresh-start, with n... | closed | completed | false | 5 | [] | [] | 2026-04-23T17:58:43Z | 2026-04-26T08:22:06Z | 2026-04-26T08:22:06Z | NONE | null | 20260426T090017Z | 2026-04-26T09:00:17Z | jrvanwinkle | 261,341,012 | U_kgDOD5O_VA | User | false |
openclaw/openclaw | 4,317,950,758 | I_kwDOQb6kR88AAAABAV6zJg | 70,725 | https://github.com/openclaw/openclaw/issues/70725 | https://api.github.com/repos/openclaw/openclaw/issues/70725 | [UI] 启动时 skills 过度加载导致对话上下文被压缩,聊天失去记忆 | ## Problem
OpenClaw 启动后,UI 在加载大量 skills 时会显示警告文字。过度加载导致的后果:
1. **对话上下文被压缩** — 系统频繁触发 context overflow,聊天历史被截断丢失
2. **长期记忆失效** — 用户反映聊着聊着就"失忆"了,之前的内容记不住
3. **UI 响应变慢** — skills 越多,loader 压力越大
根本原因:OpenClaw 默认加载 **所有** 已安装的 skills,不管当前会话是否需要。
## Proposed Solution
参考 browser.entries 的 whitelist 模式,为 skills 添加类似的机制:
... | closed | completed | true | 1 | [] | [] | 2026-04-23T18:18:23Z | 2026-04-29T04:41:56Z | 2026-04-26T07:17:14Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | starbb7946 | 278,772,344 | U_kgDOEJ26eA | User | false |
openclaw/openclaw | 4,317,994,173 | I_kwDOQb6kR88AAAABAV9cvQ | 70,729 | https://github.com/openclaw/openclaw/issues/70729 | https://api.github.com/repos/openclaw/openclaw/issues/70729 | webchat: surface X-Forwarded-User (trusted-proxy) identity to agent message context | ## Summary
When the gateway runs with `gateway.auth.mode: trusted-proxy` behind a reverse proxy that authenticates users (Tailscale, IAP, Pomerium, forward_auth, etc.) and forwards the identity via `X-Forwarded-User`, the authenticated email is used for **session scoping** (session keys prefixed with `trusted-proxy:<e... | open | null | false | 1 | [] | [] | 2026-04-23T18:26:26Z | 2026-05-02T12:54:12Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | tylr | 23,349 | MDQ6VXNlcjIzMzQ5 | User | false |
openclaw/openclaw | 4,318,061,261 | I_kwDOQb6kR88AAAABAWBizQ | 70,732 | https://github.com/openclaw/openclaw/issues/70732 | https://api.github.com/repos/openclaw/openclaw/issues/70732 | [Bug]: | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On OpenClaw 2026.4.22, system exec completion notifications are output in the chat, which bloats session context and wastes tokens.
When an exec command completes asynchronously, the gateway sends a system me... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-23T18:38:16Z | 2026-04-29T04:41:56Z | 2026-04-26T07:17:24Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | rcarozza | 22,714,599 | MDQ6VXNlcjIyNzE0NTk5 | User | false |
openclaw/openclaw | 4,318,079,697 | I_kwDOQb6kR88AAAABAWCq0Q | 70,733 | https://github.com/openclaw/openclaw/issues/70733 | https://api.github.com/repos/openclaw/openclaw/issues/70733 | [Bug]: Critical Bug: Spontaneous insertion of "//" in identifiers during 'edit' and 'write' operations | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
I have encountered a critical regression in the edit and write tools. During file modifications, the agent spontaneously inserts double slashes (//) inside variable names, function calls, and method identifiers. This results... | open | null | false | 1 | [
"bug"
] | [] | 2026-04-23T18:41:14Z | 2026-05-02T13:08:18Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | emutant-labs | 203,697,394 | U_kgDODCQs8g | User | false |
openclaw/openclaw | 4,317,990,570 | I_kwDOQb6kR88AAAABAV9Oqg | 70,728 | https://github.com/openclaw/openclaw/issues/70728 | https://api.github.com/repos/openclaw/openclaw/issues/70728 | Windows startup feels slow because Telegram recovery is consistently slow and Feishu sometimes times out during startup | ## Environment
- OpenClaw version: 2026.4.15
- OS: Windows 10 x64
- Node: v22.22.2
- Channels enabled: Telegram, Feishu
- Feishu mode: websocket
- Network: Wi-Fi, no system proxy configured
## Summary
On this machine, OpenClaw feels slow to become fully usable after startup, but the delay is not explained by core gate... | closed | completed | true | 3 | [] | [] | 2026-04-23T18:25:48Z | 2026-04-29T04:30:55Z | 2026-04-26T21:31:52Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | huihaixiao-ctrl | 273,124,102 | U_kgDOEEeLBg | User | false |
openclaw/openclaw | 4,318,129,093 | I_kwDOQb6kR88AAAABAWFrxQ | 70,736 | https://github.com/openclaw/openclaw/issues/70736 | https://api.github.com/repos/openclaw/openclaw/issues/70736 | [Bug]: ollama-cloud provider (openai-completions) doesn't use native tool calling — model hallucinates exec outputs | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When using `ollama-cloud` as provider with `nemotron-3-super:cloud` (or any
tool-capable model), OpenClaw sends the available tools as plain text inside
the system prompt instead of using the native `tools` p... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-23T18:50:03Z | 2026-04-29T04:41:55Z | 2026-04-26T07:17:29Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | tobeefornottobeef | 83,718,258 | MDQ6VXNlcjgzNzE4MjU4 | User | false |
openclaw/openclaw | 4,318,172,209 | I_kwDOQb6kR88AAAABAWIUMQ | 70,738 | https://github.com/openclaw/openclaw/issues/70738 | https://api.github.com/repos/openclaw/openclaw/issues/70738 | [Bug]: | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Macos app Channels not always show options
<img width="936" height="990" alt="Image" src="https://github.com/user-attachments/assets/0c8e7806-7740-4480-bd11-cb56cd330b3f" />
### Steps to reproduce
install ap... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-23T18:57:50Z | 2026-04-29T04:41:54Z | 2026-04-26T07:17:33Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | stiliajohny | 17,669,235 | MDQ6VXNlcjE3NjY5MjM1 | User | false |
openclaw/openclaw | 4,318,189,357 | I_kwDOQb6kR88AAAABAWJXLQ | 70,740 | https://github.com/openclaw/openclaw/issues/70740 | https://api.github.com/repos/openclaw/openclaw/issues/70740 | [Feature]: Clarify and support persistent main-orchestrated specialist agents vs spawned sub-agent runs | ### Summary
I need clarification on OpenClaw’s intended multi-agent architecture.
I originally assumed OpenClaw could support one user-facing main agent that persistently orchestrates multiple specialist agents (code-builder, tech-researcher, knowledge-master, etc.) as internal workers.
After extensive testing, it n... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-23T19:00:57Z | 2026-04-28T04:47:27Z | 2026-04-25T19:27:10Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | shinyotaek-ux | 264,289,081 | U_kgDOD8C7OQ | User | false |
openclaw/openclaw | 4,318,096,684 | I_kwDOQb6kR88AAAABAWDtLA | 70,734 | https://github.com/openclaw/openclaw/issues/70734 | https://api.github.com/repos/openclaw/openclaw/issues/70734 | [Bug]: Fresh-session `[object Object]` hallucination persists after session/database wipe on fix branch for #69079 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After building from source on the fix/issue-69079-heartbeat-user-turn-leak branch (PR #69278, commit f8b7862) to get the heartbeat trigger-policy fix, the main agent no longer returns HEARTBEAT_OK on user turns. Good.
Howeve... | open | null | false | 4 | [
"bug",
"regression"
] | [] | 2026-04-23T18:44:04Z | 2026-05-03T04:43:47Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | piercegordonfrankfort-star | 260,506,806 | U_kgDOD4cEtg | User | false |
openclaw/openclaw | 4,134,555,142 | I_kwDOQb6kR872cE4G | 54,428 | https://github.com/openclaw/openclaw/issues/54428 | https://api.github.com/repos/openclaw/openclaw/issues/54428 | Weixin Plugin Load Failure: Missing openclaw Package in Dependencies | # Weixin Plugin Load Failure: Missing `openclaw` Package in Dependencies
## Issue Type
- [x] Bug Report
- [ ] Feature Request
- [ ] Documentation Issue
## Description
The `openclaw-weixin` plugin fails to load with a module resolution error. The plugin requires the `openclaw` package at runtime for its SDK imports, ... | closed | completed | true | 1 | [] | [] | 2026-03-25T11:08:28Z | 2026-04-26T04:30:02Z | 2026-04-23T19:28:04Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | 43622283 | 42,956,557 | MDQ6VXNlcjQyOTU2NTU3 | User | false |
openclaw/openclaw | 4,318,346,964 | I_kwDOQb6kR88AAAABAWS-1A | 70,746 | https://github.com/openclaw/openclaw/issues/70746 | https://api.github.com/repos/openclaw/openclaw/issues/70746 | Control UI: tool call result payload missing from expanded block | ## Description
When expanding a tool call block in the Control UI chat view, the tool name and inputs are visible but the result payload is not displayed — the expanded block is empty.
## Steps to Reproduce
1. Open the Control UI chat at http://127.0.0.1:19178
2. Trigger any tool call (e.g. a Read file operation)
3. C... | closed | completed | true | 1 | [] | [] | 2026-04-23T19:28:49Z | 2026-04-28T04:47:25Z | 2026-04-25T19:27:23Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tarvis0523 | 277,300,810 | U_kgDOEIdGSg | User | false |
openclaw/openclaw | 4,209,420,408 | I_kwDOQb6kR8765qh4 | 61,666 | https://github.com/openclaw/openclaw/issues/61666 | https://api.github.com/repos/openclaw/openclaw/issues/61666 | Docs/UX: clarify ChatGPT Plus vs Codex OAuth vs API-key routing for GPT-5.4 (and document limit differences) | ## Summary
OpenClaw currently makes it hard to understand how **ChatGPT Plus / ChatGPT OAuth**, **Codex OAuth**, and **OpenAI API key** relate when using GPT-5.4.
Two user-facing problems keep causing confusion:
1. Users cannot tell whether ChatGPT Plus entitlement can be used directly for `openai/gpt-5.4`, or only ... | closed | completed | true | 2 | [] | [] | 2026-04-06T03:52:24Z | 2026-04-26T04:30:00Z | 2026-04-23T19:36:44Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | Lrs-GreenBottle | 253,340,193 | U_kgDODxmqIQ | User | false |
openclaw/openclaw | 4,318,331,663 | I_kwDOQb6kR88AAAABAWSDDw | 70,745 | https://github.com/openclaw/openclaw/issues/70745 | https://api.github.com/repos/openclaw/openclaw/issues/70745 | [Bug]: Discord DM fails with "Unknown Channel" in 2026.4.21 (regression from 2026.4.14) | **Describe the bug**
In OpenClaw 2026.4.21, sending a DM (direct message) to a Discord user fails with the error "Unknown Channel".
**Regression**
This worked correctly in version 2026.4.14.
**To Reproduce**
1. Attempt to send a DM to a user who has not previously interacted with the bot in DMs.
2. The DM fails with ... | closed | completed | true | 2 | [] | [] | 2026-04-23T19:26:19Z | 2026-04-28T04:47:26Z | 2026-04-25T19:27:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | clawdius-cch | 264,650,380 | U_kgDOD8Y-jA | User | false |
openclaw/openclaw | 4,318,448,606 | I_kwDOQb6kR88AAAABAWZL3g | 70,752 | https://github.com/openclaw/openclaw/issues/70752 | https://api.github.com/repos/openclaw/openclaw/issues/70752 | gateway install can early-return already-installed while loaded LaunchAgent still embeds stale OPENCLAW_GATEWAY_TOKEN | ## Summary
On macOS, `openclaw gateway install` can exit with `Gateway service already loaded` even when the currently loaded LaunchAgent still embeds a stale `OPENCLAW_GATEWAY_TOKEN`.
That leaves the machine stuck in a `token_mismatch` / unauthorized loop after a config token rotation or post-update auth change unti... | closed | completed | true | 1 | [] | [] | 2026-04-23T19:46:19Z | 2026-04-30T04:38:11Z | 2026-04-26T06:42:16Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | tsora-assistant | 269,593,031 | U_kgDOEBGpxw | User | false |
openclaw/openclaw | 4,318,485,912 | I_kwDOQb6kR88AAAABAWbdmA | 70,755 | https://github.com/openclaw/openclaw/issues/70755 | https://api.github.com/repos/openclaw/openclaw/issues/70755 | [Bug]: BlueBubbles setup reuses the server password as the public webhook secret | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 9.4 / 10.0 | 9.3 / 10.0 |
| **Severity** | Critical | Critical |
| **Vector** | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:L | CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:H/VA:L/SC:N/SI:N/SA:N |
| **Calculator** |... | open | null | false | 1 | [] | [] | 2026-04-23T19:52:33Z | 2026-05-02T21:25:49Z | null | CONTRIBUTOR | null | 20260503T000641Z | 2026-05-03T00:06:41Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 4,318,485,974 | I_kwDOQb6kR88AAAABAWbd1g | 70,756 | https://github.com/openclaw/openclaw/issues/70756 | https://api.github.com/repos/openclaw/openclaw/issues/70756 | 4.22: Discord provider crash-loops with 'Cannot find package openclaw' on npm global install (ESM subpath import fails from plugin-runtime-deps cache) | ## Summary
On a clean `npm install -g openclaw@2026.4.22`, every Discord provider instance crash-loops immediately after plugin provisioning with:
```
[discord] [default] channel exited: Cannot find package 'openclaw' imported from
/home/<user>/.openclaw-<profile>/plugin-runtime-deps/openclaw-2026.4.22-<hash>/dist/ex... | closed | completed | true | 3 | [] | [] | 2026-04-23T19:52:34Z | 2026-04-26T04:29:22Z | 2026-04-24T01:19:43Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | olalawal | 1,463,247 | MDQ6VXNlcjE0NjMyNDc= | User | false |
openclaw/openclaw | 4,318,507,137 | I_kwDOQb6kR88AAAABAWcwgQ | 70,758 | https://github.com/openclaw/openclaw/issues/70758 | https://api.github.com/repos/openclaw/openclaw/issues/70758 | [Bug]: Cron --announce fails with numeric Telegram chatId: Unsupported channel | ## Bug Description
When a cron job with `--announce` is configured with a numeric Telegram chatId as the delivery destination, the announce fails with:
```
Unsupported channel: 834732672
```
The same happens when the chatId is stored in the `delivery.channel` field without the `telegram:` prefix.
## Environment
- ... | closed | completed | true | 1 | [] | [] | 2026-04-23T19:56:44Z | 2026-04-28T04:47:24Z | 2026-04-25T19:27:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | AS76 | 22,101,267 | MDQ6VXNlcjIyMTAxMjY3 | User | false |
openclaw/openclaw | 4,124,171,887 | I_kwDOQb6kR8710d5v | 53,210 | https://github.com/openclaw/openclaw/issues/53210 | https://api.github.com/repos/openclaw/openclaw/issues/53210 | [Bug] openai/gpt-image-1 not registered in image_generate providers on 2026.3.22 despite loaded openai plugin and valid OPENAI_API_KEY | # [Bug] openai/gpt-image-1 not registered in image_generate providers on 2026.3.22 despite loaded openai plugin and valid OPENAI_API_KEY
## Environment
- **OpenClaw version**: 2026.3.22 (4dcc39c)
- **OS**: macOS 15.7.1 (arm64)
- **Node**: v22.22.1
- **Gateway**: LaunchAgent, loopback
## Configuration
```json
{
"a... | closed | completed | true | 4 | [] | [] | 2026-03-23T22:55:31Z | 2026-04-26T04:29:58Z | 2026-04-23T20:06:06Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | xiangyazi24 | 8,440,428 | MDQ6VXNlcjg0NDA0Mjg= | User | false |
openclaw/openclaw | 4,080,993,419 | I_kwDOQb6kR87zPwSL | 48,035 | https://github.com/openclaw/openclaw/issues/48035 | https://api.github.com/repos/openclaw/openclaw/issues/48035 | [Bug]: main branch build error | ### Bug type
Crash (process/app exits or hangs)
### Summary
[PARSE_ERROR] Error: '0'-prefixed octal literals and octal escape sequences are deprecated
╭─[ ../node_modules/.pnpm/qrcode-terminal@0.12.0/node_modules/qrcode-terminal/lib/main.js:3:13 ]
│
3 │ black = "\033[40m \033[0m",
│ ─────... | closed | completed | true | 0 | [
"bug",
"bug:crash"
] | [] | 2026-03-16T08:28:09Z | 2026-04-26T04:29:56Z | 2026-04-23T20:06:16Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | Unrealplace | 16,250,045 | MDQ6VXNlcjE2MjUwMDQ1 | User | false |
openclaw/openclaw | 4,318,456,792 | I_kwDOQb6kR88AAAABAWZr2A | 70,753 | https://github.com/openclaw/openclaw/issues/70753 | https://api.github.com/repos/openclaw/openclaw/issues/70753 | Bundled HTTP MCP client times out against hosted SSE server even when endpoint responds in <1s | # Bundled HTTP MCP client times out against hosted SSE server even when endpoint responds in <1s
## Summary
After upgrading from `openclaw@2026.4.12` → `2026.4.22`, all HTTP-transport MCP servers pointing at `https://mcp.agentmail.to/mcp` fail to start at gateway boot with a 15–30s SSE/stream timeout. Direct `curl` t... | closed | completed | true | 2 | [] | [] | 2026-04-23T19:47:46Z | 2026-04-29T04:34:31Z | 2026-04-26T18:57:53Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | jojopas | 6,856,488 | MDQ6VXNlcjY4NTY0ODg= | User | false |
openclaw/openclaw | 4,318,460,605 | I_kwDOQb6kR88AAAABAWZ6vQ | 70,754 | https://github.com/openclaw/openclaw/issues/70754 | https://api.github.com/repos/openclaw/openclaw/issues/70754 | [Bug]: sandboxed embedded session shows browser in /tools, but browser is missing from injected tool schemas | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The agent reports that `browser` is available at runtime via `/tools`, but `browser` is not actually present in the injected tool schemas shown by `/context detail`. As a result, the model cannot call the browser tool even t... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-04-23T19:48:23Z | 2026-04-27T04:46:41Z | 2026-04-25T00:01:31Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | bighitbiker3 | 15,604,901 | MDQ6VXNlcjE1NjA0OTAx | User | false |
openclaw/openclaw | 4,318,647,707 | I_kwDOQb6kR88AAAABAWlVmw | 70,766 | https://github.com/openclaw/openclaw/issues/70766 | https://api.github.com/repos/openclaw/openclaw/issues/70766 | Make CLAUDE_LIVE_MAX_STDOUT_BUFFER_CHARS configurable (256KB hardcoded limit aborts long claude-cli turns) | ## Summary
The claude-cli runtime in `execute.runtime-rara7Weg.js` hardcodes a 256KB stdout buffer limit (`CLAUDE_LIVE_MAX_STDOUT_BUFFER_CHARS = 256 * 1024`). When claude-cli emits a single stream-json line larger than this — common with extended thinking on Opus 4.7 / xhigh / max — the live session is aborted with `"... | closed | completed | true | 1 | [] | [] | 2026-04-23T20:21:41Z | 2026-04-28T04:42:47Z | 2026-04-26T02:36:08Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tpjordan | 262,481,515 | U_kgDOD6Umaw | User | false |
openclaw/openclaw | 4,116,258,770 | I_kwDOQb6kR871WR_S | 52,305 | https://github.com/openclaw/openclaw/issues/52305 | https://api.github.com/repos/openclaw/openclaw/issues/52305 | [Bug]: async task completion reports can be lost because system event/wake is not reliably session-targeted | ## Bug type
Bug / async completion routing
## Summary
Async task completion reporting is unreliable when external task runners (for example Codex via `exec`) try to notify OpenClaw with:
```bash
openclaw system event --text "...done..." --mode now
```
In practice this can fail in two ways:
1. the CLI call itself ... | open | null | false | 10 | [] | [] | 2026-03-22T14:56:20Z | 2026-05-04T21:00:22Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | lazica-hub | 267,841,865 | U_kgDOD_bxSQ | User | false |
openclaw/openclaw | 4,318,298,827 | I_kwDOQb6kR88AAAABAWQCyw | 70,744 | https://github.com/openclaw/openclaw/issues/70744 | https://api.github.com/repos/openclaw/openclaw/issues/70744 | Telegram direct session can become unrecoverable after context overflow and auto-compaction hangs | ## Summary
A long-lived Telegram direct chat became permanently non-responsive after hitting context overflow. The gateway and Telegram provider both restarted normally, but the same Telegram direct session immediately re-entered auto-compaction and never recovered enough to send a reply. Repeated gateway restarts did ... | closed | completed | true | 4 | [] | [] | 2026-04-23T19:20:57Z | 2026-04-29T04:36:57Z | 2026-04-26T10:53:30Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | asleep75 | 268,383,229 | U_kgDOD_8z_Q | User | false |
openclaw/openclaw | 4,318,885,340 | I_kwDOQb6kR88AAAABAWz13A | 70,774 | https://github.com/openclaw/openclaw/issues/70774 | https://api.github.com/repos/openclaw/openclaw/issues/70774 | Feature request: slim-mode / pass --bare to claude-cli backend so sub-Opus Claude models fit the context window | ### Summary
On hosts where the `claude-cli` provider is used, the session scaffold currently exceeds ~200k tokens before any user prompt is added. This makes every Claude model with <200k context (`claude-sonnet-4-6` at 195k, `claude-haiku-4-5` at 200k) unusable — requests silently fail with `FailoverError: Prompt is ... | closed | completed | true | 1 | [] | [] | 2026-04-23T21:06:55Z | 2026-04-28T04:47:24Z | 2026-04-25T19:27:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | edith-vitaetpax | 263,146,870 | U_kgDOD69Ndg | User | false |
openclaw/openclaw | 4,318,885,507 | I_kwDOQb6kR88AAAABAWz2gw | 70,775 | https://github.com/openclaw/openclaw/issues/70775 | https://api.github.com/repos/openclaw/openclaw/issues/70775 | Feature request: slim/bare mode for claude-cli backend to support sub-opus models | ## Problem
The `claude-cli` provider backend invokes the underlying Claude Code CLI without `--bare`, so every session rebuilds the full Claude Code harness: CLAUDE.md auto-discovery up the directory tree, hooks, plugin sync, auto-memory, background prefetches. On a reasonably active workstation this preload easily ex... | closed | completed | true | 1 | [] | [] | 2026-04-23T21:06:57Z | 2026-04-28T04:47:23Z | 2026-04-25T19:27:43Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | edith-vitaetpax | 263,146,870 | U_kgDOD69Ndg | User | false |
openclaw/openclaw | 4,318,896,137 | I_kwDOQb6kR88AAAABAW0gCQ | 70,776 | https://github.com/openclaw/openclaw/issues/70776 | https://api.github.com/repos/openclaw/openclaw/issues/70776 | [Bug]: thinkingDefault=xhigh silently remapped to high for anthropic/claude-opus-4-7 — provider-prefixed modelId bypasses isAnthropicOpus47Model matcher | ## Summary
`thinkingDefault: xhigh` set in `openclaw.json` at `agents.defaults.thinkingDefault` is silently **remapped to `high`** on new DM sessions for `anthropic/claude-opus-4-7`, with the notice:
```
Thinking level set to high (xhigh not supported for anthropic/claude-opus-4-7).
```
Opus 4.7 **does** support `xh... | closed | not_planned | true | 1 | [] | [] | 2026-04-23T21:08:49Z | 2026-04-28T04:47:22Z | 2026-04-25T19:27:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | goslingmanagment | 220,680,726 | U_kgDODSdSFg | User | false |
openclaw/openclaw | 4,318,909,118 | I_kwDOQb6kR88AAAABAW1Svg | 70,778 | https://github.com/openclaw/openclaw/issues/70778 | https://api.github.com/repos/openclaw/openclaw/issues/70778 | OpenClaw browser service: intermittent timeouts on click actions (20s cap) – root-cause analysis and fix plan | ### Summary
OpenClaw’s browser automation occasionally times out during click actions, even though navigation succeeds. The known 20,000 ms (20s) cap appears to constrain click interactions.
### Symptoms
- Navigation (page loads) succeeds
- Click actions fail with a timeout after ~20 seconds
- Timeouts are intermitten... | closed | completed | true | 1 | [] | [] | 2026-04-23T21:11:05Z | 2026-04-26T04:29:54Z | 2026-04-23T21:12:21Z | CONTRIBUTOR | null | 20260426T060034Z | 2026-04-26T06:00:34Z | DranboFieldston | 258,684,645 | U_kgDOD2s25Q | User | false |
openclaw/openclaw | 4,318,918,758 | I_kwDOQb6kR88AAAABAW14Zg | 70,779 | https://github.com/openclaw/openclaw/issues/70779 | https://api.github.com/repos/openclaw/openclaw/issues/70779 | [Regression] OpenAI-compatible Gemini backend enters streaming mode but completes immediately with no incremental content updates (4.19 OK, 4.21+/4.22 broken) | # Regression: OpenAI-compatible Gemini backend enters streaming mode but completes immediately with no incremental content updates (4.19 OK, 4.21+/4.22 broken)
## Summary
After upgrading OpenClaw from `2026.4.19` to `2026.4.21` / `2026.4.22`, streaming appears to regress for a custom **OpenAI-compatible Gemini backen... | closed | not_planned | true | 2 | [] | [] | 2026-04-23T21:12:59Z | 2026-05-02T04:28:55Z | 2026-04-29T08:22:25Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | CharyeahOwO | 105,187,600 | U_kgDOBkUJEA | User | false |
openclaw/openclaw | 4,318,950,910 | I_kwDOQb6kR88AAAABAW31_g | 70,782 | https://github.com/openclaw/openclaw/issues/70782 | https://api.github.com/repos/openclaw/openclaw/issues/70782 | Feature: slim mode to pass --bare to claude-cli backend (unlock sub-opus models on heavy workspaces) | ## Problem
On a workspace with many skills / a rich CLAUDE.md / ShieldCortex-style plugin hooks, the OpenClaw `claude-cli` backend builds a pre-turn harness (auto-discovered CLAUDE.md up the tree, plugin sync, hooks, skill catalog, auto-memory) that can exceed the context window of smaller Claude models.
Concrete rep... | closed | completed | true | 1 | [] | [] | 2026-04-23T21:19:10Z | 2026-04-29T04:41:53Z | 2026-04-26T07:17:38Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | edith-vitaetpax | 263,146,870 | U_kgDOD69Ndg | User | false |
openclaw/openclaw | 4,318,965,384 | I_kwDOQb6kR88AAAABAW4uiA | 70,783 | https://github.com/openclaw/openclaw/issues/70783 | https://api.github.com/repos/openclaw/openclaw/issues/70783 | [Feature]: Add tab system for managing multiple chat sessions | ### Summary
Add a tab-based navigation to allow working on multiple chat sessions at the same time.
### Problem to solve
Right now the interface only let you use one chat at a time, which becomes quickly frustrating when you need to juggle between different conversations or tasks. You basically lose context or have ... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-04-23T21:22:07Z | 2026-05-02T04:28:47Z | 2026-04-29T05:16:48Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | alexandre-leng | 72,550,137 | MDQ6VXNlcjcyNTUwMTM3 | User | false |
openclaw/openclaw | 4,318,970,118 | I_kwDOQb6kR88AAAABAW5BBg | 70,785 | https://github.com/openclaw/openclaw/issues/70785 | https://api.github.com/repos/openclaw/openclaw/issues/70785 | [Feature]: Adding automatic model loading and query the provider for the model list | ### Summary
So many other providers, softwares, options automatcially make a list of the available models including open webui
To simplify the stress of the set up process and constant need to make a new list everytime you want to add a new agent, model, or provider this should be a fundimental feature.
### Problem... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-23T21:23:10Z | 2026-04-28T04:47:21Z | 2026-04-25T19:27:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | cdvankammen | 31,197,020 | MDQ6VXNlcjMxMTk3MDIw | User | false |
openclaw/openclaw | 4,318,993,521 | I_kwDOQb6kR88AAAABAW6ccQ | 70,787 | https://github.com/openclaw/openclaw/issues/70787 | https://api.github.com/repos/openclaw/openclaw/issues/70787 | [Bug]: Web UI chat messages disappear immediately - scope upgrade approval loop blocks replies | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Sending a message in the Control UI chat causes the message bubble to disappear immediately with no reply, due to a continuous scope upgrade approval loop that persists even after approving via openclaw devices... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-23T21:28:06Z | 2026-04-29T04:41:52Z | 2026-04-26T07:17:43Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | goktuggokbulut | 52,534,872 | MDQ6VXNlcjUyNTM0ODcy | User | false |
openclaw/openclaw | 4,319,002,604 | I_kwDOQb6kR88AAAABAW6_7A | 70,788 | https://github.com/openclaw/openclaw/issues/70788 | https://api.github.com/repos/openclaw/openclaw/issues/70788 | fix(windows): suppress startup-folder cmd window flash via wscript silent launcher | ## Problem
On Windows, the Startup folder login item (`OpenClaw Gateway.cmd`) launches the gateway using:
\`\`\`batch
start "" /min cmd.exe /d /c C:\Users\<user>\.openclaw\gateway.cmd
\`\`\`
The \`/min\` flag minimizes the window but does not suppress rendering — Windows still composites the \`cmd.exe\` frame before... | open | null | false | 1 | [] | [] | 2026-04-23T21:29:53Z | 2026-05-03T10:19:30Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | gthgomez | 91,163,862 | MDQ6VXNlcjkxMTYzODYy | User | false |
openclaw/openclaw | 4,318,898,363 | I_kwDOQb6kR88AAAABAW0ouw | 70,777 | https://github.com/openclaw/openclaw/issues/70777 | https://api.github.com/repos/openclaw/openclaw/issues/70777 | Main agent intermittently starts with wrong cwd and tries to read ~/.openclaw/agents/main/agent/AGENTS.md instead of workspace AGENTS.md | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
Main agent intermittently starts with cwd /home/elicw instead of the configured workspace /home/elicw/.openclaw/workspace, and then attempts to read /home/elicw/.openclaw/agents/main/agent/AGENTS.md instead of the workspace boo... | closed | completed | true | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-23T21:09:11Z | 2026-04-29T04:35:46Z | 2026-04-26T14:56:27Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | elco5 | 36,783,379 | MDQ6VXNlcjM2NzgzMzc5 | User | false |
openclaw/openclaw | 4,318,872,175 | I_kwDOQb6kR88AAAABAWzCbw | 70,773 | https://github.com/openclaw/openclaw/issues/70773 | https://api.github.com/repos/openclaw/openclaw/issues/70773 | [Bug]: Telegram channel keeps loading stale runtime hash after update/reinstall (Cannot find package 'openclaw' from plugin-runtime-deps) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After updating to OpenClaw `2026.4.22 (00bd2cf)`, Telegram channel startup repeatedly fails with:
`Cannot find package 'openclaw' imported from .../.openclaw/plugin-runtime-deps/openclaw-2026.4.22-4eca5026e977/...`
It appear... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-04-23T21:04:35Z | 2026-04-26T04:28:53Z | 2026-04-24T02:28:05Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | JDTm | 2,218,600 | MDQ6VXNlcjIyMTg2MDA= | User | false |
openclaw/openclaw | 4,319,087,685 | I_kwDOQb6kR88AAAABAXAMRQ | 70,794 | https://github.com/openclaw/openclaw/issues/70794 | https://api.github.com/repos/openclaw/openclaw/issues/70794 | [Bug]: WhatsApp plugin doesn't allow group communication | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Version: 2026.4.15-beta.1
Log line: Listening for personal WhatsApp inbound messages
Config showing groupPolicy: "open" and groups defined
No group messages appearing in logs
### Steps to reproduce
1. Install... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-23T21:48:28Z | 2026-04-28T04:47:21Z | 2026-04-25T19:28:10Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | iso25010 | 109,476,605 | U_kgDOBoZ6_Q | User | false |
openclaw/openclaw | 4,319,129,572 | I_kwDOQb6kR88AAAABAXCv5A | 70,796 | https://github.com/openclaw/openclaw/issues/70796 | https://api.github.com/repos/openclaw/openclaw/issues/70796 | [Feature]: Add a conversations dropdown under Chat in the sidebar to switch between chat threads more easily | ### Summary
Introduce a dropdown list of conversations under the Chat section to quickly access and switch between different chat threads.
### Problem to solve
Right now there is no clear way to navigate between multiple conversations from the sidebar, which makes it hard to reuse past chats or continue previous wor... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-04-23T21:58:07Z | 2026-04-29T04:34:13Z | 2026-04-26T14:56:32Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | alexandre-leng | 72,550,137 | MDQ6VXNlcjcyNTUwMTM3 | User | false |
openclaw/openclaw | 4,197,261,141 | I_kwDOQb6kR876LR9V | 59,931 | https://github.com/openclaw/openclaw/issues/59931 | https://api.github.com/repos/openclaw/openclaw/issues/59931 | Config validation error: tools.web.search.brave no longer valid after update | ## Description
After updating OpenClaw, the gateway fails config validation because `tools.web.search` now has `additionalProperties: false` and the `brave` sub-key is no longer in the schema.
## Steps to Reproduce
1. Have a config with:
```json
"tools": {
"web": {
"search": {
"provider": "brave",
... | closed | completed | true | 3 | [] | [] | 2026-04-02T22:22:08Z | 2026-04-28T04:50:46Z | 2026-04-25T18:37:25Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | bvoorhees | 10,913,615 | MDQ6VXNlcjEwOTEzNjE1 | User | false |
openclaw/openclaw | 4,144,839,591 | I_kwDOQb6kR873DTun | 55,237 | https://github.com/openclaw/openclaw/issues/55237 | https://api.github.com/repos/openclaw/openclaw/issues/55237 | [Bug] openai-codex provider missing image/audio capabilities in media understanding | ## Summary
The `openai-codex` provider (used for OpenAI Pro plan OAuth access) does not register `image` or `audio` capabilities in the media understanding provider registry. This means agents using `openai-codex/gpt-5.4` as their model cannot use the `image` tool or benefit from media understanding preprocessing, des... | closed | completed | true | 1 | [] | [] | 2026-03-26T15:06:58Z | 2026-04-26T04:29:52Z | 2026-04-23T22:00:22Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | stanleeyY | 4,137,404 | MDQ6VXNlcjQxMzc0MDQ= | User | false |
openclaw/openclaw | 4,047,040,024 | I_kwDOQb6kR87xOO4Y | 41,322 | https://github.com/openclaw/openclaw/issues/41322 | https://api.github.com/repos/openclaw/openclaw/issues/41322 | image tool fails on openai-codex/* with {"detail":"Instructions are required"} | ## Environment
- OpenClaw: 2026.3.8
- Provider: openai-codex
- Models tested:
- openai-codex/gpt-5.3-codex
- openai-codex/gpt-5.1-codex-mini
## Summary
Using the `image` tool with OpenAI Codex models fails consistently with:
```json
{"detail":"Instructions are required"}
```
even when `prompt` is provided.
## R... | closed | completed | true | 2 | [] | [] | 2026-03-09T17:54:03Z | 2026-04-26T04:29:52Z | 2026-04-23T22:00:27Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | ion-ctrl | 266,816,281 | U_kgDOD-dLGQ | User | false |
openclaw/openclaw | 4,319,157,550 | I_kwDOQb6kR88AAAABAXEdLg | 70,801 | https://github.com/openclaw/openclaw/issues/70801 | https://api.github.com/repos/openclaw/openclaw/issues/70801 | fix(macos): two-phase stop+start in restartLaunchAgent to eliminate kickstart-before-ensurePidGone race | ## Background
PR #43961 adds `ensurePidGone` to guarantee stale gateway processes are dead before a new instance is started. Greptile's review (4/5 confidence) correctly flags a remaining structural limitation:
In `restartLaunchAgent`, `kickstart -k` launches the **new** process first, then `ensurePidGone` is awaited... | open | null | false | 1 | [] | [] | 2026-04-23T22:04:33Z | 2026-05-03T10:19:31Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | happydog-intj | 212,984 | MDQ6VXNlcjIxMjk4NA== | User | false |
openclaw/openclaw | 4,319,178,059 | I_kwDOQb6kR88AAAABAXFtSw | 70,804 | https://github.com/openclaw/openclaw/issues/70804 | https://api.github.com/repos/openclaw/openclaw/issues/70804 | [Bug]: Slack channel mentions reach OpenClaw but no reply is sent; /openclaw still times out on 2026.4.22 | ## Bug type
Regression / Slack channel integration
## Summary
On macOS with OpenClaw `2026.4.22`, Slack channel mentions are reaching OpenClaw, but OpenClaw never sends a reply back to Slack. In the same setup, `/openclaw` still fails with Slack's `the app did not respond` timeout.
This looks related to existing Slac... | closed | completed | true | 1 | [] | [] | 2026-04-23T22:10:01Z | 2026-04-29T04:36:02Z | 2026-04-26T12:21:20Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | nratai | 277,380,771 | U_kgDOEIh-ow | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.