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 | 4,188,948,075 | I_kwDOQb6kR875rkZr | 59,195 | https://github.com/openclaw/openclaw/issues/59195 | https://api.github.com/repos/openclaw/openclaw/issues/59195 | [Bug]: Long Pop-up Windows Blocking UI | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Approval pop-up windows which contain too many rows of text block the entire UI. No chance to approve. Blocked.
<img width="1911" height="883" alt="Image" src="https://github.com/user-attachments/assets/79b11900-5bb6-4b96-b... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-01T18:17:50Z | 2026-04-28T04:54:01Z | 2026-04-25T17:58:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | leamara-2026 | 270,386,270 | U_kgDOEB3EXg | User | false |
openclaw/openclaw | 4,189,360,948 | I_kwDOQb6kR875tJM0 | 59,220 | https://github.com/openclaw/openclaw/issues/59220 | https://api.github.com/repos/openclaw/openclaw/issues/59220 | web_search broken — DuckDuckGo returns bot detection since 2026-03-30 | **Bug:** web_search broken — DuckDuckGo returns bot detection since 2026-03-30
**Severity:** HIGH
**First logged:** 2026-03-30
**Last logged:** 2026-03-31
**Description:**
All web_search calls fail with "DuckDuckGo returned a bot-detection challenge" starting 2026-03-30T18:44. This blocks x:daily-post, research task... | closed | completed | true | 1 | [] | [] | 2026-04-01T19:24:40Z | 2026-04-28T04:54:00Z | 2026-04-25T17:58:44Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ion-ctrl | 266,816,281 | U_kgDOD-dLGQ | User | false |
openclaw/openclaw | 4,328,769,808 | I_kwDOQb6kR88AAAABAgPJEA | 71,683 | https://github.com/openclaw/openclaw/issues/71683 | https://api.github.com/repos/openclaw/openclaw/issues/71683 | opencode-go/deepseek-v4-flash tool schema failure and Telegram planning text leak after fallback | ## Summary
Using `opencode-go/deepseek-v4-flash` in Telegram fails on tool-enabled requests with a provider schema error, and the fallback path can expose internal planning text in the visible reply.
## OpenClaw Version
- `2026.4.21`
## Environment
- Channel: Telegram direct chat
- Provider/model: `opencode-go/deepse... | closed | completed | true | 2 | [] | [] | 2026-04-25T17:21:56Z | 2026-04-28T04:53:48Z | 2026-04-25T17:58:47Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | MaheshBhushan | 128,616,744 | U_kgDOB6qJKA | User | false |
openclaw/openclaw | 4,183,811,785 | I_kwDOQb6kR875X-bJ | 58,712 | https://github.com/openclaw/openclaw/issues/58712 | https://api.github.com/repos/openclaw/openclaw/issues/58712 | [Bug]: [Follow-up] memos-memory-guide skill shows disabled but works correctly - WSL path issue? | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Additional Report: memos-memory-guide
Following up on #9469, I'm experiencing a similar issue with the `memos-memory-guide` skill.
### Environment
- OS: Windows 11 (Asia/Shanghai timezone)
- OpenClaw version: Latest (20... | closed | not_planned | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-01T03:36:26Z | 2026-04-28T04:53:01Z | 2026-04-25T18:10:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | sixwingskybear | 161,640,283 | U_kgDOCaJvWw | User | false |
openclaw/openclaw | 4,185,571,414 | I_kwDOQb6kR875esBW | 58,911 | https://github.com/openclaw/openclaw/issues/58911 | https://api.github.com/repos/openclaw/openclaw/issues/58911 | [Bug]: Published package is missing plugin-sdk Telegram exports used by third-party plugins | ## Summary
On the currently published OpenClaw package (`2026.3.31` in our environment), the npm package exports do **not** include `./plugin-sdk/telegram` (or `./plugin-sdk/telegram-runtime`).
That breaks third-party plugins that legitimately depend on Telegram runtime helpers via the plugin SDK surface.
Concrete fa... | closed | not_planned | true | 1 | [] | [] | 2026-04-01T09:41:07Z | 2026-04-28T04:53:00Z | 2026-04-25T18:10:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | evgyur | 120,411,712 | U_kgDOBy1WQA | User | false |
openclaw/openclaw | 4,187,721,517 | I_kwDOQb6kR875m48t | 59,097 | https://github.com/openclaw/openclaw/issues/59097 | https://api.github.com/repos/openclaw/openclaw/issues/59097 | Bug: buffered block dispatcher splits messages at ~4000 chars, breaking non-chunked channels | The `dispatchReplyWithBufferedBlockDispatcher` splits LLM replies into chunks at `DEFAULT_CHUNK_LIMIT` (4000 chars) and calls `deliver()` multiple times per turn. Channels that concatenate these chunks (like atticus-channel, which joins with `\n\n`) end up with spurious line breaks mid-sentence at ~4000-char boundaries... | closed | completed | true | 1 | [] | [] | 2026-04-01T15:12:31Z | 2026-04-28T04:53:00Z | 2026-04-25T18:10:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | atticusfawler | 268,220,059 | U_kgDOD_y2mw | User | false |
openclaw/openclaw | 4,189,239,955 | I_kwDOQb6kR875srqT | 59,215 | https://github.com/openclaw/openclaw/issues/59215 | https://api.github.com/repos/openclaw/openclaw/issues/59215 | [Bug]: exec tool cannot authenticate to gateway in loopback mode | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Agent's exec tool fails with "pairing required" error when running local commands, even though the same commands work fine from the terminal and gateway auth is set to "none".
### Steps to reproduce
1. Start OpenClaw 2026... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-01T19:04:59Z | 2026-04-28T04:52:59Z | 2026-04-25T18:10:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | rhimmon1 | 25,462,323 | MDQ6VXNlcjI1NDYyMzIz | User | false |
openclaw/openclaw | 4,328,894,010 | I_kwDOQb6kR88AAAABAgWuOg | 71,697 | https://github.com/openclaw/openclaw/issues/71697 | https://api.github.com/repos/openclaw/openclaw/issues/71697 | Ollama Cloud provider fails in agent/Gateway path with non-alternating conversation roles | ## Summary
Ollama Cloud works directly with the configured `OLLAMA_API_KEY`, but using the same `ollama` provider as an OpenClaw agent/Gateway model fails because the request sent through OpenClaw's agent path is rejected by Ollama Cloud with a non-alternating conversation roles error.
This is related to, but distinc... | closed | completed | true | 2 | [] | [] | 2026-04-25T18:21:08Z | 2026-04-29T04:33:19Z | 2026-04-27T00:03:17Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ismael-81 | 273,866,131 | U_kgDOEFLdkw | User | false |
openclaw/openclaw | 4,328,916,892 | I_kwDOQb6kR88AAAABAgYHnA | 71,705 | https://github.com/openclaw/openclaw/issues/71705 | https://api.github.com/repos/openclaw/openclaw/issues/71705 | [Bug]: Azure gpt-image-2 image generation times out under hardcoded 180s default after #71526 fix | ## Summary
The Azure `gpt-image-2` request-shape issue from #71526 appears fixed by omitting `model` from Azure deployment-scoped image requests, but there is a follow-up problem: OpenClaw's default OpenAI image timeout is too short for slow Azure image generations.
## Background
Issue #71526 identified that Azure ima... | closed | completed | true | 0 | [] | [] | 2026-04-25T18:33:02Z | 2026-04-28T04:45:50Z | 2026-04-25T22:34:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | voytas75 | 13,856,061 | MDQ6VXNlcjEzODU2MDYx | User | false |
openclaw/openclaw | 4,184,145,049 | I_kwDOQb6kR875ZPyZ | 58,755 | https://github.com/openclaw/openclaw/issues/58755 | https://api.github.com/repos/openclaw/openclaw/issues/58755 | [Feature]: Configurable Feishu probe TTL (PROBE_SUCCESS_TTL_MS) | ## Problem
`probeFeishu()` in `extensions/feishu/src/probe.ts` caches bot info for only 10 minutes (`PROBE_SUCCESS_TTL_MS = 600_000`). With multi-account setups (e.g. 3 Feishu accounts), this results in ~12,960 `/open-apis/bot/v3/info` calls per month across all accounts.
Bot info (name, avatar, bot_id) is **essentia... | closed | not_planned | true | 1 | [] | [] | 2026-04-01T05:12:45Z | 2026-04-28T04:50:49Z | 2026-04-25T18:36:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | colakang | 17,260,007 | MDQ6VXNlcjE3MjYwMDA3 | User | false |
openclaw/openclaw | 4,186,706,640 | I_kwDOQb6kR875jBLQ | 59,012 | https://github.com/openclaw/openclaw/issues/59012 | https://api.github.com/repos/openclaw/openclaw/issues/59012 | [macOS] microphone recording commands not available on node | ## 问题描述
macOS 节点缺少麦克风录音相关命令。
## 环境
- Gateway: OpenClaw 2026.3.31
- macOS 节点: NightQ的Mac mini, macOS 26.4.0, app v2026.3.31
## 复现步骤
1. 查看节点权限: `openclaw nodes describe --node "NightQ的Mac mini"`
2. 发现权限中有 `microphone=yes`
3. 但命令列表中没有麦克风录制相关的命令
## 实际行为
节点权限显示:
```
Perms: [accessibility=yes, appleScript=yes, camera=... | closed | not_planned | true | 1 | [] | [] | 2026-04-01T12:54:07Z | 2026-04-28T04:50:49Z | 2026-04-25T18:37:05Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | nightq | 3,429,433 | MDQ6VXNlcjM0Mjk0MzM= | User | false |
openclaw/openclaw | 4,191,166,618 | I_kwDOQb6kR8750CCa | 59,374 | https://github.com/openclaw/openclaw/issues/59374 | https://api.github.com/repos/openclaw/openclaw/issues/59374 | Feature Request: Support multiple Tavily API keys for failover | ## Problem
When the Tavily API key reaches its rate limit or runs out of quota, users need to manually switch to another API key. Currently, OpenClaw only supports a single Tavily API key configuration, which is inconvenient for users who have multiple keys.
## Use Case
In a production environment, users may have:
-... | closed | not_planned | true | 1 | [] | [] | 2026-04-02T03:17:45Z | 2026-04-28T04:50:48Z | 2026-04-25T18:37:12Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | liuguangchao103 | 49,757,969 | MDQ6VXNlcjQ5NzU3OTY5 | User | false |
openclaw/openclaw | 4,194,826,787 | I_kwDOQb6kR876B_oj | 59,701 | https://github.com/openclaw/openclaw/issues/59701 | https://api.github.com/repos/openclaw/openclaw/issues/59701 | OpenClaw 2026.4.1 Resets exec-approvals.json Allowlist - Breaking Automated Workflows | # Bug Report: OpenClaw 2026.4.1 Resets exec-approvals.json Allowlist
**Severity:** High
**Impact:** Breaking change for automated workflows (cron jobs, schedulers)
## Summary
After updating to OpenClaw 2026.4.1, the `exec-approvals.json` file was reset/initialized with strict defaults, removing all previously work... | closed | completed | true | 1 | [] | [] | 2026-04-02T14:02:32Z | 2026-04-28T04:50:47Z | 2026-04-25T18:37:18Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Reneb-cafe | 106,862,786 | U_kgDOBl6Ywg | User | false |
openclaw/openclaw | 4,197,965,921 | I_kwDOQb6kR876N-Bh | 60,016 | https://github.com/openclaw/openclaw/issues/60016 | https://api.github.com/repos/openclaw/openclaw/issues/60016 | [Bug]: Google provider models fall to 200k context fallback due to empty discovery cache | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
All Gemini models resolve to a 200k context window instead of their actual 1,048,576 token limit because the Google extension does not register a provider catalog, leaving models.json empty for Google and causi... | closed | completed | true | 1 | [] | [] | 2026-04-03T02:46:28Z | 2026-04-28T04:50:46Z | 2026-04-25T18:37:31Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Baelson | 7,864,962 | MDQ6VXNlcjc4NjQ5NjI= | User | false |
openclaw/openclaw | 4,328,899,465 | I_kwDOQb6kR88AAAABAgXDiQ | 71,699 | https://github.com/openclaw/openclaw/issues/71699 | https://api.github.com/repos/openclaw/openclaw/issues/71699 | [Bug]: Gateway hard-crashes with 0xC0000409 (STATUS_STACK_BUFFER_OVERRUN) on Windows during Mattermost streaming reply; auto-respawn frequently wedges | ## Bug
The gateway crashes hard on Windows with exit code `3221226505` (`0xC0000409` — `STATUS_STACK_BUFFER_OVERRUN`) during normal operation: incoming Mattermost channel events arriving while the embedded `acpx` runtime is mid-inference. The user-visible symptom is **a Mattermost post that the bot stopped editing mid... | open | null | false | 2 | [] | [] | 2026-04-25T18:23:56Z | 2026-05-02T15:23:22Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | mathiticus3 | 185,200,111 | U_kgDOCwnt7w | User | false |
openclaw/openclaw | 4,328,932,764 | I_kwDOQb6kR88AAAABAgZFnA | 71,709 | https://github.com/openclaw/openclaw/issues/71709 | https://api.github.com/repos/openclaw/openclaw/issues/71709 | [Bug]: slug-generator HTTP 400 misclassified as profile-wide billing failure (5h cooldown), kills all agents on profile | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
A transient HTTP 400 returned to the internal `generateSlugViaLLM` helper is misclassified as a profile-wide billing failure: OpenClaw marks the entire `anthropic:claude-cli` auth profile as `disabled:billing` wit... | open | null | false | 1 | [] | [] | 2026-04-25T18:40:38Z | 2026-05-02T16:01:56Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | nikolaykazakovvs-ux | 233,889,353 | U_kgDODfDeSQ | User | false |
openclaw/openclaw | 4,328,934,502 | I_kwDOQb6kR88AAAABAgZMZg | 71,710 | https://github.com/openclaw/openclaw/issues/71710 | https://api.github.com/repos/openclaw/openclaw/issues/71710 | [Bug]: openclaw-agent ignores SIGTERM under cron, accumulates hung process chains and exhausts host RAM/swap | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
`openclaw agent` processes spawned by cron with `timeout 600 openclaw agent ...` (without `-k <delay>`) accumulate as long-lived hung chains because the agent does not exit on SIGTERM. Within ~2.5 days a multi-fir... | open | null | false | 1 | [] | [] | 2026-04-25T18:41:23Z | 2026-05-02T16:01:58Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | nikolaykazakovvs-ux | 233,889,353 | U_kgDODfDeSQ | User | false |
openclaw/openclaw | 4,328,945,477 | I_kwDOQb6kR88AAAABAgZ3RQ | 71,711 | https://github.com/openclaw/openclaw/issues/71711 | https://api.github.com/repos/openclaw/openclaw/issues/71711 | [Feature] activeUntil / expiresAt for recurring cron jobs | ## Summary
Add a per-job `activeUntil` (or `expiresAt`) field for recurring cron schedules (`--every`, `--cron`). When the configured timestamp passes, the scheduler stops triggering the job and (configurably) removes it from `~/.openclaw/cron/jobs.json`.
## Motivation
OpenClaw cron one-shots (`--at`) auto-delete af... | open | null | false | 1 | [] | [] | 2026-04-25T18:46:49Z | 2026-05-02T13:08:32Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | alonp99 | 8,467,965 | MDQ6VXNlcjg0Njc5NjU= | User | false |
openclaw/openclaw | 4,194,444,475 | I_kwDOQb6kR876AiS7 | 59,677 | https://github.com/openclaw/openclaw/issues/59677 | https://api.github.com/repos/openclaw/openclaw/issues/59677 | [Feature]: In the Feishu environment, the OpenClaw robot is unable to respond to messages sent by other robot accounts, but it responds completely normally to messages from human users. | ### Summary
In the Feishu environment, the OpenClaw robot is unable to respond to messages sent by other robot accounts, but it responds completely normally to messages from human users.
### Problem to solve
In the Feishu (Lark) environment, I have deployed two Feishu robots simultaneously: a standard business rob... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-02T12:59:53Z | 2026-04-28T04:49:48Z | 2026-04-25T18:48:08Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | WangQvQ | 58,406,737 | MDQ6VXNlcjU4NDA2NzM3 | User | false |
openclaw/openclaw | 4,196,285,186 | I_kwDOQb6kR876HjsC | 59,860 | https://github.com/openclaw/openclaw/issues/59860 | https://api.github.com/repos/openclaw/openclaw/issues/59860 | iMessage self-chat echo suppression fails: message-ID short-circuit prevents text-based fallback match | ## Bug: iMessage self-chat echo suppression fails due to message-ID short-circuit in `SentMessageCache.has()`
### Summary
After updating to the latest OpenClaw release, outbound agent replies in iMessage self-chat (Notes to Self / single-person conversation) are echoed back and re-ingested as new inbound messages. Th... | closed | completed | true | 1 | [] | [] | 2026-04-02T18:35:05Z | 2026-04-28T04:49:48Z | 2026-04-25T18:48:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jeffweisbein | 47,037,958 | MDQ6VXNlcjQ3MDM3OTU4 | User | false |
openclaw/openclaw | 4,196,964,366 | I_kwDOQb6kR876KJgO | 59,910 | https://github.com/openclaw/openclaw/issues/59910 | https://api.github.com/repos/openclaw/openclaw/issues/59910 | Inbound images from Slack exceed Anthropic 5MB base64 limit (no auto-resize) | ## Bug
Images sent via Slack are stored raw (Slack does not compress uploads like Telegram does). When the image is between 5MB and 6MB, it passes OpenClaw's `MAX_IMAGE_BYTES` (6MB) check but gets rejected by Anthropic's API:
```
LLM request rejected: messages.107.content.1.image.source.base64: image exceeds 5 MB max... | closed | completed | true | 1 | [] | [] | 2026-04-02T21:01:01Z | 2026-04-28T04:49:47Z | 2026-04-25T18:48:28Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | evetanaka | 257,669,453 | U_kgDOD1u5TQ | User | false |
openclaw/openclaw | 4,328,958,830 | I_kwDOQb6kR88AAAABAgarbg | 71,714 | https://github.com/openclaw/openclaw/issues/71714 | https://api.github.com/repos/openclaw/openclaw/issues/71714 | [Bug] Enter key does not send message in chat input on v4.22+ | ## Bug Description
Since v4.22, pressing the Enter key in the Control UI chat input box does not send the message. Instead, it inserts a newline character.
All send shortcuts are unresponsive. Users must manually click the Send button to send messages.
## Environment
- OpenClaw version: v4.22 and later (including v... | closed | completed | true | 1 | [] | [] | 2026-04-25T18:53:28Z | 2026-04-29T04:35:06Z | 2026-04-26T16:17:11Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | zhong18804784882 | 268,618,080 | U_kgDOEALJYA | User | false |
openclaw/openclaw | 4,328,965,320 | I_kwDOQb6kR88AAAABAgbEyA | 71,715 | https://github.com/openclaw/openclaw/issues/71715 | https://api.github.com/repos/openclaw/openclaw/issues/71715 | WhatsApp: XML tool-tag sanitizer regression in 2026.4.23 + missing exposeErrorText/silentErrorReplies parity | ## Summary
Two related observations from production WhatsApp deployment on `openclaw@2026.4.23`:
1. **Sanitizer regression** — Anthropic-style XML tool-calling tags (`<function_calls>…</function_calls>`, `<invoke>`, `<parameter>`, `<tool_call>`) are bleeding through to user-visible WhatsApp messages on certain agent ... | closed | completed | true | 1 | [] | [] | 2026-04-25T18:56:53Z | 2026-04-29T04:43:27Z | 2026-04-26T10:54:36Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | xkola | 26,909,281 | MDQ6VXNlcjI2OTA5Mjgx | User | false |
openclaw/openclaw | 4,328,968,163 | I_kwDOQb6kR88AAAABAgbP4w | 71,717 | https://github.com/openclaw/openclaw/issues/71717 | https://api.github.com/repos/openclaw/openclaw/issues/71717 | exec tool returns EPERM on Windows, all commands fail | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
在 Windows 11 上，所有通过 exec 工具执行的命令都立即失败，错误信息为 `spawn EPERM`。
### Steps to reproduce
1. 安装 openclaw 并完成配置
2. 通过飞书或 web 界面发送任何命令
3. exec 工具返回错误: `spawn EPERM`
### Expected behavior
NOT_ENOUGH_INFO
### Actual ... | open | null | false | 1 | [
"bug"
] | [] | 2026-04-25T18:58:21Z | 2026-05-02T16:01:59Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | ljcdegj | 129,809,843 | U_kgDOB7y9sw | User | false |
openclaw/openclaw | 4,328,965,448 | I_kwDOQb6kR88AAAABAgbFSA | 71,716 | https://github.com/openclaw/openclaw/issues/71716 | https://api.github.com/repos/openclaw/openclaw/issues/71716 | Subagent sessions not cleaned up after archiveAfterMinutes (orphaned sessions accumulate) | ## Bug Report: Subagent sessions not cleaned up after archiveAfterMinutes
### Issue Summary
Subagent sessions are not being cleaned up after `archiveAfterMinutes: 60` (1 hour) as configured. Sessions persist for days, accumulating in `agents/*/sessions/` directories. The `sessions-archived/` folder remains empty, ind... | closed | not_planned | true | 0 | [] | [] | 2026-04-25T18:56:58Z | 2026-04-28T04:49:41Z | 2026-04-25T18:59:25Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Kondziosz | 118,409,363 | U_kgDOBw7Ikw | User | false |
openclaw/openclaw | 4,328,987,956 | I_kwDOQb6kR88AAAABAgcdNA | 71,719 | https://github.com/openclaw/openclaw/issues/71719 | https://api.github.com/repos/openclaw/openclaw/issues/71719 | [Feature]: TUI, show a short recent-session summary on startup/history load | ### Summary
Show a short recent-session summary in the TUI after connection/history load so users can quickly recover working context when
returning to an existing session.
### Problem to solve
When opening the TUI, it can be hard to immediately recover the latest working context, especially after stepping away for ... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-25T19:08:20Z | 2026-04-29T04:39:09Z | 2026-04-26T09:26:49Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | MertBasar0 | 100,199,793 | U_kgDOBfjtcQ | User | false |
openclaw/openclaw | 3,895,057,471 | I_kwDOQb6kR87oKdw_ | 8,629 | https://github.com/openclaw/openclaw/issues/8629 | https://api.github.com/repos/openclaw/openclaw/issues/8629 | [Showcase] TitanShell - Security First Desktop Client for OpenClaw | ## 👋 Hi OpenClaw Team!
I've built **TitanShell**, a security-first desktop client for OpenClaw.
### Features
- 🛡️ Zero-trust architecture with Tauri 2.0
- 🔐 macOS Keychain + biometric protection
- ⚡ Ephemeral token injection (no credentials on disk)
- 🎨 Tactical terminal UI theme
- 🖥️ Cross-platform: macOS, Wind... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-04T07:14:16Z | 2026-04-28T04:48:35Z | 2026-04-25T19:15:42Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | DaguangZhou | 44,826,597 | MDQ6VXNlcjQ0ODI2NTk3 | User | false |
openclaw/openclaw | 4,064,368,047 | I_kwDOQb6kR87yQVWv | 44,038 | https://github.com/openclaw/openclaw/issues/44038 | https://api.github.com/repos/openclaw/openclaw/issues/44038 | [Feature]: [Community Tool / Proposal] Windows Quick Installer GUI - Lowering the barrier to entry for Windows users | ### Summary
I am proposing (and have developed) the OpenClaw Windows Quick Installer, an open-source desktop GUI application designed to help Windows users quickly install, configure, and manage the OpenClaw CLI. This tool aims to drastically lower the barrier to entry for non-developers and AI enthusiasts who want to... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-03-12T12:36:00Z | 2026-04-28T04:48:34Z | 2026-04-25T19:15:44Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | JustinBIBERRR | 89,630,091 | MDQ6VXNlcjg5NjMwMDkx | User | false |
openclaw/openclaw | 4,135,606,932 | I_kwDOQb6kR872gFqU | 54,493 | https://github.com/openclaw/openclaw/issues/54493 | https://api.github.com/repos/openclaw/openclaw/issues/54493 | Add CatchClaw as a community marketplace resource for OpenClaw | ### Summary
I’ve been building [CatchClaw](https://catchclaw.me), a marketplace for OpenClaw agents.
It focuses on making reusable agents easier to discover, package, install, export, and share. I’m opening this issue mainly to ask whether this kind of community resource is relevant enough to be referenced somewhere ... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-03-25T13:50:35Z | 2026-04-28T04:48:27Z | 2026-04-25T19:16:41Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Lovelcp | 4,645,149 | MDQ6VXNlcjQ2NDUxNDk= | User | false |
openclaw/openclaw | 4,329,014,503 | I_kwDOQb6kR88AAAABAgeE5w | 71,721 | https://github.com/openclaw/openclaw/issues/71721 | https://api.github.com/repos/openclaw/openclaw/issues/71721 | [Bug]: Crash in sandbox: uv_interface_addresses error (networkInterfaces) causes gateway disconnect | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Summary
OpenClaw gateway crashes inside a Nemoclaw/OpenShell sandbox with:
```
SystemError: A system error occurred: uv_interface_addresses returned Unknown system error 1
```
This prevents the TUI from connecting (`ga... | open | null | false | 2 | [
"bug"
] | [] | 2026-04-25T19:22:21Z | 2026-05-02T17:31:56Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | joelvds | 279,323,301 | U_kgDOEKYipQ | User | false |
openclaw/openclaw | 4,328,946,457 | I_kwDOQb6kR88AAAABAgZ7GQ | 71,712 | https://github.com/openclaw/openclaw/issues/71712 | https://api.github.com/repos/openclaw/openclaw/issues/71712 | [RFC] Agent-facing scheduling API with non-forgeable provenance | ## Summary
Document and harden a path for **agents** (not just operators) to create/modify their own recurring cron jobs at runtime, with **non-forgeable provenance** captured by the gateway rather than self-reported by the agent.
Today, `openclaw cron add` is operator-facing and writes `~/.openclaw/cron/jobs.json` d... | open | null | false | 3 | [] | [] | 2026-04-25T18:47:21Z | 2026-05-02T16:55:48Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | alonp99 | 8,467,965 | MDQ6VXNlcjg0Njc5NjU= | User | false |
openclaw/openclaw | 4,076,624,751 | I_kwDOQb6kR87y_Ftv | 46,546 | https://github.com/openclaw/openclaw/issues/46546 | https://api.github.com/repos/openclaw/openclaw/issues/46546 | Android: Talk mode concurrent playback / duplicate response delivery | ## Description
Using Talk mode on Android, we noticed a couple of things with how responses are delivered:
### 1. Concurrent audio playback
When the assistant reply comes back, we sometimes hear multiple audio streams playing at the same time rather than sequentially. This seems to happen when the response triggers m... | closed | completed | true | 1 | [] | [] | 2026-03-14T20:08:02Z | 2026-04-28T04:46:30Z | 2026-04-25T19:44:18Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lde0 | 186,663,029 | U_kgDOCyBAdQ | User | false |
openclaw/openclaw | 4,329,057,176 | I_kwDOQb6kR88AAAABAggrmA | 71,728 | https://github.com/openclaw/openclaw/issues/71728 | https://api.github.com/repos/openclaw/openclaw/issues/71728 | [RFC] Agent truthfulness and confidence signaling for verifiable trust | ## Problem
No mechanism to distinguish known vs inferred vs hallucinated statements in agent output. Users cannot reliably trust agent assertions in autonomous mode.
## Proposed
1. **Confidence tiers**: known / believed / uncertain / unknown — explicit flag on every factual claim
2. **Source attribution**: file+line... | open | null | false | 1 | [] | [] | 2026-04-25T19:44:38Z | 2026-05-02T11:36:31Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | AS76 | 22,101,267 | MDQ6VXNlcjIyMTAxMjY3 | User | false |
openclaw/openclaw | 4,102,950,987 | I_kwDOQb6kR870jhBL | 50,541 | https://github.com/openclaw/openclaw/issues/50541 | https://api.github.com/repos/openclaw/openclaw/issues/50541 | [Bug]: CLI cannot connect to Gateway (missing scope: operator.read) - Downgrade to 2026.3.11 works | ### Bug type
Regression (worked before, now fails)
### Summary
### Issue: CLI Cannot Connect to Gateway – Downgrading from 2026.3.13 to 2026.3.11 Resolves the Problem
**Environment**
- OS: Ubuntu 24.04.4 LTS (Linux 6.8.0-101-generic x86_64)
- Node: 22.22.1
- Problematic version: OpenClaw **2026.3.13**
- Wor... | closed | not_planned | true | 2 | [
"bug",
"regression",
"clownfish"
] | [] | 2026-03-19T16:52:19Z | 2026-04-28T04:32:21Z | 2026-04-25T19:46:17Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | passer593 | 268,878,053 | U_kgDOEAbA5Q | User | false |
openclaw/openclaw | 4,104,746,051 | I_kwDOQb6kR870qXRD | 50,691 | https://github.com/openclaw/openclaw/issues/50691 | https://api.github.com/repos/openclaw/openclaw/issues/50691 | openclaw status reports missing operator.read while gateway is healthy; gateway probe times out on same loopback endpoint | ## Environment
- OpenClaw: `2026.3.13 (61d171a)`
- OS: Windows 10 (x64), Node `24.13.1`
- Gateway mode: `local`
- Bind: `loopback`
- Auth mode: `token`
- Gateway URL: `ws://127.0.0.1:18789`
## Summary
I’m seeing contradictory gateway/auth behavior on the same machine/session:
- `openclaw gateway status` reports servi... | closed | not_planned | true | 2 | [
"clownfish"
] | [] | 2026-03-19T23:18:29Z | 2026-04-28T04:32:21Z | 2026-04-25T19:46:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | StarShipSoundStudios | 264,524,250 | U_kgDOD8RR2g | User | false |
openclaw/openclaw | 4,329,045,391 | I_kwDOQb6kR88AAAABAgf9jw | 71,726 | https://github.com/openclaw/openclaw/issues/71726 | https://api.github.com/repos/openclaw/openclaw/issues/71726 | [Bug]: Isolated cron task summaries routed to Telegram despite notify: silent | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
My agent (Scorchio, running OpenClaw) generated this issue — Garry12gg is posting it on its behalf.
Bug description:
Isolated cron job sessions are sending their terminalSummary to Telegram even when notifyPol... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-25T19:38:20Z | 2026-04-30T04:38:41Z | 2026-04-26T05:27:50Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | garry12gg | 62,000,769 | MDQ6VXNlcjYyMDAwNzY5 | User | false |
openclaw/openclaw | 4,329,032,803 | I_kwDOQb6kR88AAAABAgfMYw | 71,725 | https://github.com/openclaw/openclaw/issues/71725 | https://api.github.com/repos/openclaw/openclaw/issues/71725 | gemini-3.1-flash-lite-preview fails on multi-turn tool use — missing thought_signature support | ## Summary
`gemini-3.1-flash-lite-preview` (and likely other Gemini 3.1 thinking-class models) are broken for any agentic task involving tool calls. The model works fine for single-turn generation but throws a `400 INVALID_ARGUMENT` error on the second turn of any tool-use conversation.
## Root Cause
Google's Gemini... | closed | completed | true | 1 | [] | [] | 2026-04-25T19:31:58Z | 2026-04-28T04:46:28Z | 2026-04-25T19:52:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | packwood | 6,589,673 | MDQ6VXNlcjY1ODk2NzM= | User | false |
openclaw/openclaw | 4,329,052,290 | I_kwDOQb6kR88AAAABAggYgg | 71,727 | https://github.com/openclaw/openclaw/issues/71727 | https://api.github.com/repos/openclaw/openclaw/issues/71727 | [Bug]: Error: Encrypted content item_id did not match the target item id when using GitHub Copilot with GPT-5-mini in OpenClaw | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
github-copilot/gpt-4.1 works fine, but github-copilot/gpt-5-mini doesn't.
### Steps to reproduce
version 2026.4.24，Error: Encrypted content item_id did not match the target item id when using GitHub Copilot w... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-25T19:42:07Z | 2026-04-28T04:46:27Z | 2026-04-25T19:52:43Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Yize9 | 18,671,767 | MDQ6VXNlcjE4NjcxNzY3 | User | false |
openclaw/openclaw | 4,329,088,196 | I_kwDOQb6kR88AAAABAgikxA | 71,739 | https://github.com/openclaw/openclaw/issues/71739 | https://api.github.com/repos/openclaw/openclaw/issues/71739 | perf(cli): agents list --json shouldn't load all extension plugins | ## Problem
`openclaw agents list --json` consistently takes **~7-9s on a fast host and ~11s in a container** on 2026.4.23, while peer commands like `cron list --all --json`, `skills list --json`, and `sessions ... --json` stay sub-second on the same install. Programmatic JSON consumers (dashboards, monitoring scripts,... | closed | completed | true | 0 | [] | [] | 2026-04-25T19:59:38Z | 2026-04-28T04:46:02Z | 2026-04-25T21:07:44Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kaloster | 7,976,579 | MDQ6VXNlcjc5NzY1Nzk= | User | false |
openclaw/openclaw | 4,070,036,319 | I_kwDOQb6kR87yl9Nf | 44,943 | https://github.com/openclaw/openclaw/issues/44943 | https://api.github.com/repos/openclaw/openclaw/issues/44943 | [Bug]: diagnostics.otel: Unrecognized key: "captureContent" | ### Bug type
Regression (worked before, now fails)
### Summary
Hi, “diagnostics.otel: Unrecognized key: "captureContent"” . Why did this error occur? Why is this field not supported, and which version can support? Thanks.
### Steps to reproduce
.
### Expected behavior
/
### Actual behavior
.
### OpenClaw vers... | closed | completed | true | 2 | [
"bug",
"regression"
] | [
"vincentkoc"
] | 2026-03-13T09:45:19Z | 2026-04-28T04:46:24Z | 2026-04-25T20:00:40Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | wsh2836741-sketch | 265,002,886 | U_kgDOD8ufhg | User | false |
openclaw/openclaw | 4,037,022,831 | I_kwDOQb6kR87woBRv | 38,477 | https://github.com/openclaw/openclaw/issues/38477 | https://api.github.com/repos/openclaw/openclaw/issues/38477 | [Feature]: OpenTelemetry Traces: Missing Parent-Child Span Relationships | ### Summary
OpenClaw's OpenTelemetry integration exports independent traces for each diagnostic event instead of creating hierarchical parent-child span relationships. This prevents proper distributed tracing visualization in tools like AWS X-Ray, Jaeger, and other APM platforms.
## Current Behavior
Each operation (... | closed | completed | true | 1 | [
"enhancement"
] | [
"vincentkoc"
] | 2026-03-07T02:17:50Z | 2026-04-28T04:46:23Z | 2026-04-25T20:01:07Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | huizi-2013 | 94,334,425 | U_kgDOBZ9t2Q | User | false |
openclaw/openclaw | 4,222,425,821 | I_kwDOQb6kR877rRrd | 62,964 | https://github.com/openclaw/openclaw/issues/62964 | https://api.github.com/repos/openclaw/openclaw/issues/62964 | Feature Request: Add child spans for detailed OTEL traces | ## Feature Request: Add child spans for detailed OTEL traces
### Problem
Currently, OpenClaw's OTEL instrumentation only provides coarse-grained spans. For example, `openclaw.message.processed` is a single span that encompasses the entire message processing time (~228 seconds), but there is no breakdown of where that ... | closed | completed | true | 1 | [] | [
"vincentkoc"
] | 2026-04-08T05:55:05Z | 2026-04-28T04:46:22Z | 2026-04-25T20:01:12Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | shaohq | 40,241,246 | MDQ6VXNlcjQwMjQxMjQ2 | User | false |
openclaw/openclaw | 3,892,980,950 | I_kwDOQb6kR87oCizW | 8,249 | https://github.com/openclaw/openclaw/issues/8249 | https://api.github.com/repos/openclaw/openclaw/issues/8249 | Feature: Enhanced OpenTelemetry trace context for tool call correlation | # OpenClaw Feature Request: Enhanced OpenTelemetry Trace Context
## Summary
Add proper trace context propagation so tool call spans are children of model usage spans, enabling hierarchical trace visualization in Jaeger/etc.
## Current Behavior
- `openclaw.model.usage` spans are created independently
- `openclaw.tool.... | closed | completed | true | 1 | [
"enhancement"
] | [
"vincentkoc"
] | 2026-02-03T20:05:35Z | 2026-04-28T04:46:21Z | 2026-04-25T20:01:17Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | evancohen-openclaw | 258,490,701 | U_kgDOD2hBTQ | User | false |
openclaw/openclaw | 4,069,360,065 | I_kwDOQb6kR87yjYHB | 44,790 | https://github.com/openclaw/openclaw/issues/44790 | https://api.github.com/repos/openclaw/openclaw/issues/44790 | [Bug]: ACP sessions leave orphaned `claude` processes, swap exhaustion | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
ACP (Agent Communication Protocol) sessions spawned via `acpx` / `claude-agent-a
cp` leave orphaned `claude` CLI processes running after session completion. Thes
e processes accumulate over time, consuming all available swap memory and degr... | closed | completed | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-03-13T07:06:57Z | 2026-04-28T04:46:21Z | 2026-04-25T20:01:41Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | danielvos1998 | 120,246,690 | U_kgDOByrRog | User | false |
openclaw/openclaw | 4,329,095,399 | I_kwDOQb6kR88AAAABAgjA5w | 71,741 | https://github.com/openclaw/openclaw/issues/71741 | https://api.github.com/repos/openclaw/openclaw/issues/71741 | Ollama web-search provider: bare `TypeError: fetch failed` despite valid key + reachable host | # Ollama web-search provider fails with `TypeError: fetch failed` even when ollama.com is reachable and key is valid
**OpenClaw version:** 2026.4.23 (a979721)
**Node:** v22.22.2
**OS:** Linux (Azure VM, x64)
**Provider:** `ollama` (Ollama Cloud Web Search)
## Summary
`openclaw capability web search --provider ollama... | closed | completed | true | 1 | [] | [] | 2026-04-25T20:02:56Z | 2026-04-28T04:45:56Z | 2026-04-25T22:06:24Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | madhvidua | 16,516,740 | MDQ6VXNlcjE2NTE2NzQw | User | false |
openclaw/openclaw | 4,063,315,844 | I_kwDOQb6kR87yMUeE | 43,899 | https://github.com/openclaw/openclaw/issues/43899 | https://api.github.com/repos/openclaw/openclaw/issues/43899 | [Bug]: Telegram forum-topic ACP turns can pass a Telegram permalink to ACP instead of the actual message body | ## Summary
Telegram forum-topic messages sent to a topic-bound ACP session can be serialized into the ACP prompt as a **Telegram permalink** (for example `https://t.me/c/<chat>/<topic>/<message>`) instead of the actual inbound message body.
When this happens, the ACP session appears alive and correctly bound, but the... | closed | completed | true | 2 | [] | [] | 2026-03-12T09:27:18Z | 2026-04-28T04:46:20Z | 2026-04-25T20:03:40Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | thetechpadawan | 260,382,279 | U_kgDOD4UeRw | User | false |
openclaw/openclaw | 4,069,022,882 | I_kwDOQb6kR87yiFyi | 44,688 | https://github.com/openclaw/openclaw/issues/44688 | https://api.github.com/repos/openclaw/openclaw/issues/44688 | [Feature]: Talk Mode — configurable speech recognition locale (non-English languages) | ## Summary
Talk Mode speech recognition should support configurable locale beyond `en-US`, enabling non-English speakers to use voice input.
## Problem to solve
Talk Mode on iOS (and likely macOS) uses `SFSpeechRecognizer` with what appears to be a hardcoded or device-default `en-US` locale. When a non-English speak... | closed | completed | true | 1 | [] | [] | 2026-03-13T05:31:43Z | 2026-04-28T04:46:19Z | 2026-04-25T20:06:13Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | morozsm | 4,393,731 | MDQ6VXNlcjQzOTM3MzE= | User | false |
openclaw/openclaw | 4,329,103,021 | I_kwDOQb6kR88AAAABAgjerQ | 71,743 | https://github.com/openclaw/openclaw/issues/71743 | https://api.github.com/repos/openclaw/openclaw/issues/71743 | [Bug]: Several commands tries to install plugins devs. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to 026.4.24 (cbcfdf6), several commands tries to in node packages.
I found it with "agents bind" and "plugins list".
"openclaw doctor" also wants to install node packages each time.
`
~@Prometheus .openclaw ... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-25T20:06:29Z | 2026-04-28T04:45:57Z | 2026-04-25T22:06:13Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fmatray | 8,267,716 | MDQ6VXNlcjgyNjc3MTY= | User | false |
openclaw/openclaw | 4,329,116,597 | I_kwDOQb6kR88AAAABAgkTtQ | 71,744 | https://github.com/openclaw/openclaw/issues/71744 | https://api.github.com/repos/openclaw/openclaw/issues/71744 | [Bug]: Watchdog fallback chain silently drops primary-failure root cause when fallback also fails — only terminal error surfaces in runtime_events | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When the watchdog fallback chain rotates a primary model to a fallback (e.g. Anthropic → local Ollama) and the fallback also fails (e.g. RAM exhaustion at the fallback host), the **root cause of the original prima... | closed | completed | true | 3 | [] | [] | 2026-04-25T20:13:32Z | 2026-04-30T04:31:58Z | 2026-04-28T04:08:52Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | nikolaykazakovvs-ux | 233,889,353 | U_kgDODfDeSQ | User | false |
openclaw/openclaw | 4,234,724,919 | I_kwDOQb6kR878aMY3 | 63,914 | https://github.com/openclaw/openclaw/issues/63914 | https://api.github.com/repos/openclaw/openclaw/issues/63914 | sessions_spawn: LLM-exposed `runtime` parameter creates spawn_failed footgun for in-config agents | ## Summary
The `sessions_spawn` tool exposes a `runtime` enum (`subagent` | `acp`) directly to the LLM. When an LLM agent (in our case Claude Haiku 4.5 acting as a triage router) escalates to an in-config specialist agent, the model can hallucinate `runtime: "acp"` even though that mode is for spawning **external** AC... | closed | completed | true | 1 | [] | [] | 2026-04-09T20:25:07Z | 2026-04-28T04:46:17Z | 2026-04-25T20:15:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tmote | 21,976,980 | MDQ6VXNlcjIxOTc2OTgw | User | false |
openclaw/openclaw | 4,329,134,759 | I_kwDOQb6kR88AAAABAglapw | 71,749 | https://github.com/openclaw/openclaw/issues/71749 | https://api.github.com/repos/openclaw/openclaw/issues/71749 | [Bug]: Windows Telegram provider still crashes with ERR_UNSUPPORTED_ESM_URL_SCHEME in 2026.4.24 unless native Jiti is disabled | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
On native Windows, OpenClaw 2026.4.24 starts successfully, but the Telegram provider crash-loops on startup with:
Only URLs with a scheme in: file, data, and node are supported by the default ESM loader. On Windows, absolut... | closed | completed | true | 4 | [
"bug"
] | [] | 2026-04-25T20:23:09Z | 2026-04-28T04:32:42Z | 2026-04-25T22:06:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | smeyer9 | 17,261,975 | MDQ6VXNlcjE3MjYxOTc1 | User | false |
openclaw/openclaw | 4,329,067,622 | I_kwDOQb6kR88AAAABAghUZg | 71,730 | https://github.com/openclaw/openclaw/issues/71730 | https://api.github.com/repos/openclaw/openclaw/issues/71730 | [Bug]: update 4.24 - Failed to install bundled plugin runtime deps: Error: npm error code EACCES - TONS of missing npm packages | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Bundled plugin runtime deps are missing. │
│ - @anthropic-ai/sdk@0.90.0 (used by amazon-bedrock-mantle) │
│ - @anthropic-ai/vertex-sdk@^0.16.0 (use... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-25T19:50:14Z | 2026-04-28T04:46:16Z | 2026-04-25T20:23:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | darealgege | 30,112,470 | MDQ6VXNlcjMwMTEyNDcw | User | false |
openclaw/openclaw | 4,227,335,443 | I_kwDOQb6kR877-AUT | 63,348 | https://github.com/openclaw/openclaw/issues/63348 | https://api.github.com/repos/openclaw/openclaw/issues/63348 | [Bug]: ACPX model forwarding to Claude Code triggers prompt cache invalidation on every turn | ## Bug type
Behavior bug
## Beta release blocker
No
## Summary
When ACPX drives a Claude Code session via ACP, the Anthropic prompt cache is invalidated on the majority of turns, forcing full cache recreation (~90K+ tokens) instead of cache reads. The issue does not occur when using Claude Code CLI directly on the... | closed | duplicate | true | 2 | [] | [] | 2026-04-08T20:10:33Z | 2026-04-28T04:46:15Z | 2026-04-25T20:24:44Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kidroca | 12,156,624 | MDQ6VXNlcjEyMTU2NjI0 | User | false |
openclaw/openclaw | 4,062,521,350 | I_kwDOQb6kR87yJSgG | 43,756 | https://github.com/openclaw/openclaw/issues/43756 | https://api.github.com/repos/openclaw/openclaw/issues/43756 | Feature Request: Slack threadBindings with spawnAcpSessions support (parity with Discord/Telegram) | ## Feature Request
### Summary
Add `threadBindings` support to the Slack channel plugin, specifically `spawnAcpSessions`, to enable thread-bound ACP agent sessions in Slack — matching the existing Discord and Telegram implementations.
### Current State
- **Discord**: Full `threadBindings` support with `spawnAcpSessio... | closed | duplicate | true | 4 | [] | [] | 2026-03-12T06:31:39Z | 2026-04-28T04:46:14Z | 2026-04-25T20:24:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | sigvardt | 6,000,238 | MDQ6VXNlcjYwMDAyMzg= | User | false |
openclaw/openclaw | 4,329,084,109 | I_kwDOQb6kR88AAAABAgiUzQ | 71,734 | https://github.com/openclaw/openclaw/issues/71734 | https://api.github.com/repos/openclaw/openclaw/issues/71734 | [RFC] Trusted tool policy stage and plugin tool metadata | ## Summary
Add a trusted pre-plugin tool policy stage and plugin-owned tool metadata so bundled/trusted plugins can enforce host-level policy and so plugin tools can appear with first-class catalog/display/safety metadata.
This is proposed SDK surface, not currently implemented API.
## Why this matters
Plan Mode mu... | closed | not_planned | true | 3 | [] | [] | 2026-04-25T19:57:50Z | 2026-04-30T15:15:36Z | 2026-04-26T12:06:22Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | 100yenadmin | 239,388,517 | U_kgDODkTHZQ | User | false |
openclaw/openclaw | 4,329,084,027 | I_kwDOQb6kR88AAAABAgiUew | 71,733 | https://github.com/openclaw/openclaw/issues/71733 | https://api.github.com/repos/openclaw/openclaw/issues/71733 | [RFC] Durable next-turn injections and agent turn preparation hooks | ## Summary
Add a host-owned durable queue for plugin next-turn injections plus a typed `agent_turn_prepare` hook so plugins can schedule future prompt additions with exactly-once semantics.
This is proposed SDK surface, not currently implemented API.
## Why this matters
Plan Mode approval acceptance, revision, ques... | closed | not_planned | true | 3 | [] | [] | 2026-04-25T19:57:47Z | 2026-04-30T15:15:33Z | 2026-04-26T14:58:38Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | 100yenadmin | 239,388,517 | U_kgDODkTHZQ | User | false |
openclaw/openclaw | 4,329,083,956 | I_kwDOQb6kR88AAAABAgiUNA | 71,732 | https://github.com/openclaw/openclaw/issues/71732 | https://api.github.com/repos/openclaw/openclaw/issues/71732 | [RFC] Plugin session extensions and patch actions | ## Summary
Add a namespaced plugin session extension API so plugins can persist typed per-session state, expose an explicit public projection to gateway/UI clients, and mutate that state through gateway-authorized patch actions.
This is proposed SDK surface, not currently implemented API.
## Why this matters
Plan M... | closed | not_planned | true | 3 | [] | [] | 2026-04-25T19:57:45Z | 2026-04-30T15:15:30Z | 2026-04-26T14:56:53Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | 100yenadmin | 239,388,517 | U_kgDODkTHZQ | User | false |
openclaw/openclaw | 4,329,084,363 | I_kwDOQb6kR88AAAABAgiVyw | 71,737 | https://github.com/openclaw/openclaw/issues/71737 | https://api.github.com/repos/openclaw/openclaw/issues/71737 | [RFC] Agent events, run context, scheduler lifecycle, and heartbeat contributions | ## Summary
Expose safe plugin subscriptions to agent events, namespaced run-scoped plugin context, session-scoped scheduler helpers, and heartbeat prompt contributions.
This is proposed SDK surface, not currently implemented API.
## Why this matters
Plan Mode parity requires plan snapshot persistence, subagent trac... | closed | not_planned | true | 3 | [] | [] | 2026-04-25T19:57:58Z | 2026-04-30T15:15:43Z | 2026-04-26T16:17:15Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | 100yenadmin | 239,388,517 | U_kgDODkTHZQ | User | false |
openclaw/openclaw | 4,329,084,277 | I_kwDOQb6kR88AAAABAgiVdQ | 71,736 | https://github.com/openclaw/openclaw/issues/71736 | https://api.github.com/repos/openclaw/openclaw/issues/71736 | [RFC] Control UI plugin contribution slots | ## Summary
Add data-driven Control UI contribution slots for plugin chat modes, approval cards, event classifiers, input guards, and status surfaces.
This is proposed SDK surface, not currently implemented API.
## Why this matters
Plan Mode should not patch `mode-switcher`, `app-tool-stream`, `app-render`, chat vie... | open | null | false | 8 | [] | [] | 2026-04-25T19:57:55Z | 2026-05-02T16:55:49Z | null | CONTRIBUTOR | null | 20260502T180029Z | 2026-05-02T18:00:29Z | 100yenadmin | 239,388,517 | U_kgDODkTHZQ | User | false |
openclaw/openclaw | 4,329,084,206 | I_kwDOQb6kR88AAAABAgiVLg | 71,735 | https://github.com/openclaw/openclaw/issues/71735 | https://api.github.com/repos/openclaw/openclaw/issues/71735 | [RFC] Scoped plugin commands, trusted command ownership, and continuation | ## Summary
Extend plugin commands with declarative gateway scopes, trusted command-root ownership, and a `continueAgent` result for commands that should resume an agent run after handling.
This is proposed SDK surface, not currently implemented API.
## Why this matters
Plan Mode `/plan accept`, `/plan revise`, and ... | closed | not_planned | true | 3 | [] | [] | 2026-04-25T19:57:53Z | 2026-04-30T15:15:39Z | 2026-04-26T10:22:46Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | 100yenadmin | 239,388,517 | U_kgDODkTHZQ | User | false |
openclaw/openclaw | 4,329,154,464 | I_kwDOQb6kR88AAAABAgmnoA | 71,752 | https://github.com/openclaw/openclaw/issues/71752 | https://api.github.com/repos/openclaw/openclaw/issues/71752 | update: global install verify false-positive on dist/extensions/*/.openclaw-install-stage artifacts | ## Summary
`openclaw update` can fail at `global install verify` with `unexpected packaged dist file .../.openclaw-install-stage/...` even when the package update itself succeeds.
## Observed Behavior
- Update output shows `Update Result: ERROR`
- Failed step: `global install verify`
- Example errors:
- `unexpected ... | closed | completed | true | 1 | [] | [] | 2026-04-25T20:33:42Z | 2026-04-30T04:38:16Z | 2026-04-26T06:28:26Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | waynegault | 39,455,578 | MDQ6VXNlcjM5NDU1NTc4 | User | false |
openclaw/openclaw | 4,329,134,603 | I_kwDOQb6kR88AAAABAglaCw | 71,748 | https://github.com/openclaw/openclaw/issues/71748 | https://api.github.com/repos/openclaw/openclaw/issues/71748 | [Feature] Agent Circadian Rhythm — Morning Brief and Evening Wrap | ## Problem
Personal AI assistants lack a daily activation rhythm. They respond to events but don't wake up with context like humans do. A professional who wakes at 5am needs a brief BEFORE handling emails, not after.
## Proposed
**Morning Brief** (configurable time): Reads recent memory, open items, upcoming events ... | closed | completed | true | 1 | [] | [] | 2026-04-25T20:23:04Z | 2026-04-28T04:46:14Z | 2026-04-25T20:36:40Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | AS76 | 22,101,267 | MDQ6VXNlcjIyMTAxMjY3 | User | false |
openclaw/openclaw | 4,329,162,302 | I_kwDOQb6kR88AAAABAgnGPg | 71,757 | https://github.com/openclaw/openclaw/issues/71757 | https://api.github.com/repos/openclaw/openclaw/issues/71757 | Matrix channel: doctor --fix fails with 'must NOT have additional properties' before it can migrate config | ## Bug
`openclaw doctor` and `openclaw gateway status` report:
```
Error: Config validation failed: channels.matrix: invalid config: must NOT have additional properties
```
The Matrix config was created via `openclaw channels add` or manual setup and contains the legacy `dm.policy` nested format:
```json
{
"chann... | closed | completed | true | 1 | [] | [] | 2026-04-25T20:37:53Z | 2026-04-28T04:43:37Z | 2026-04-26T01:34:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | aweiker | 38,160 | MDQ6VXNlcjM4MTYw | User | false |
openclaw/openclaw | 4,329,162,262 | I_kwDOQb6kR88AAAABAgnGFg | 71,756 | https://github.com/openclaw/openclaw/issues/71756 | https://api.github.com/repos/openclaw/openclaw/issues/71756 | [Bug]: Session lifecycle never finalizes — status: "running" persists after session close, breaking maintenance eviction and dashboard observability | **Bug type:** Behavior bug
**Summary**
When a session ends normally, the Gateway never updates the `status` field in `sessions.json`. The field stays `"running"` regardless of how the session ended — whether it ran for 2 minutes via a cron job or 8 hours interactively. This has two concrete consequences:
**Consequen... | closed | not_planned | true | 1 | [] | [] | 2026-04-25T20:37:52Z | 2026-04-29T04:39:08Z | 2026-04-26T09:26:53Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | SpecialAgentGlitch | 260,494,040 | U_kgDOD4bS2A | User | false |
openclaw/openclaw | 4,329,060,108 | I_kwDOQb6kR88AAAABAgg3DA | 71,729 | https://github.com/openclaw/openclaw/issues/71729 | https://api.github.com/repos/openclaw/openclaw/issues/71729 | Control UI CSP blocks blob: avatar URLs generated by avatar fetch path | ## Summary
Control UI still serves a Content Security Policy that does not allow `blob:` images:
```text
img-src 'self' data:
```
This can block avatar rendering in frontend paths that fetch `/avatar/<agentId>` and then render the result via `URL.createObjectURL(await response.blob())`.
This is related to #70700, b... | closed | completed | true | 3 | [] | [] | 2026-04-25T19:46:11Z | 2026-04-28T04:33:09Z | 2026-04-25T20:49:13Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ismael-81 | 273,866,131 | U_kgDOEFLdkw | User | false |
openclaw/openclaw | 4,329,172,937 | I_kwDOQb6kR88AAAABAgnvyQ | 71,759 | https://github.com/openclaw/openclaw/issues/71759 | https://api.github.com/repos/openclaw/openclaw/issues/71759 | Bonjour/mDNS conflict can crash gateway with unhandled CIAO cancellation rejection | ### Summary
When two OpenClaw gateways are running on the same host/network namespace, the Bonjour/mDNS advertiser can hit a name/hostname conflict and crash the gateway with an unhandled promise rejection from `@homebridge/ciao`.
This looks like a recoverable local-discovery conflict, but it currently takes down the... | closed | completed | true | 2 | [] | [] | 2026-04-25T20:43:17Z | 2026-04-28T04:35:29Z | 2026-04-25T20:53:58Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | johnturek | 53,944 | MDQ6VXNlcjUzOTQ0 | User | false |
openclaw/openclaw | 4,329,144,816 | I_kwDOQb6kR88AAAABAgmB8A | 71,751 | https://github.com/openclaw/openclaw/issues/71751 | https://api.github.com/repos/openclaw/openclaw/issues/71751 | [Bug]: Docker container crashes on 4.24 due to [openclaw] Unhandled promise rejection: CIAO PROBING CANCELLED | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
2026-04-25T20:14:47.695+00:00 [openclaw] Unhandled promise rejection: CIAO PROBING CANCELLED
2026-04-25T20:14:47.700+00:00 [openclaw] wrote stability bundle: /home/node/.openclaw/logs/stability/openclaw-stability-2026-04-25T20-... | closed | completed | true | 3 | [
"bug",
"bug:crash"
] | [] | 2026-04-25T20:28:42Z | 2026-04-28T04:46:11Z | 2026-04-25T20:53:58Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | k0Iry | 6,511,925 | MDQ6VXNlcjY1MTE5MjU= | User | false |
openclaw/openclaw | 4,329,177,517 | I_kwDOQb6kR88AAAABAgoBrQ | 71,761 | https://github.com/openclaw/openclaw/issues/71761 | https://api.github.com/repos/openclaw/openclaw/issues/71761 | [Bug]: All channel messages are injected twice into agent context after 2026.4.24 update (duplicate turns, double token cost) | ### Describe the bug
After updating from `2026.4.23` to `2026.4.24`, **every inbound message** — from **WebChat (Control UI)** and **NapCat (QQ plugin)** — is injected **twice** into the agent transcript, causing the agent to reply twice and consuming **2× tokens** per turn.
This is **not** a NapCat-specific issue. W... | closed | completed | true | 6 | [] | [] | 2026-04-25T20:45:20Z | 2026-04-28T04:30:40Z | 2026-04-26T01:10:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | WhiteGiverMa | 152,406,589 | U_kgDOCRWKPQ | User | false |
openclaw/openclaw | 4,180,902,827 | I_kwDOQb6kR875M4Or | 58,488 | https://github.com/openclaw/openclaw/issues/58488 | https://api.github.com/repos/openclaw/openclaw/issues/58488 | acpx: model forwarding to ACP agents uses Claude-specific _meta convention; breaks non-Claude harnesses (e.g. kilocode) | ## Problem
When acpx spawns an ACP harness agent and wants to pass a model, it uses a Claude Code-specific metadata convention:
```js
_meta: { claudeCode: { options: { model: "openrouter/deepseek/deepseek-v3.2" } } }
```
This is passed in the `newSession` ACP call. Any non-Claude harness (e.g. `kilocode`, `opencode`... | closed | completed | true | 3 | [] | [] | 2026-03-31T17:12:16Z | 2026-04-28T04:44:12Z | 2026-04-26T01:03:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | CarsonGreenland | 3,248,403 | MDQ6VXNlcjMyNDg0MDM= | User | false |
openclaw/openclaw | 4,329,202,027 | I_kwDOQb6kR88AAAABAgphaw | 71,766 | https://github.com/openclaw/openclaw/issues/71766 | https://api.github.com/repos/openclaw/openclaw/issues/71766 | [Bug]: CLI probe commands (openclaw gateway status) create read-only device pairing record that blocks admin operations | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
Yes
### Summary
Running CLI probe commands like `openclaw gateway status` or `openclaw status` creates a read-only (`operator.read`) device pairing record in `devices/paired.json`. This read-only record subsequently blocks an... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-25T20:58:06Z | 2026-04-28T04:45:58Z | 2026-04-25T22:04:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | SunboZ | 54,214,234 | MDQ6VXNlcjU0MjE0MjM0 | User | false |
openclaw/openclaw | 4,329,205,473 | I_kwDOQb6kR88AAAABAgpu4Q | 71,767 | https://github.com/openclaw/openclaw/issues/71767 | https://api.github.com/repos/openclaw/openclaw/issues/71767 | [Bug]: Invalid skill slug: spclaudehome/skill-vetter | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When I use "openclaw skills install spclaudehome/skill-vetter" to install skill-vetter, it outputs "Invalid skill slug: spclaudehome/skill-vetter"
### Steps to reproduce
openclaw skills install spclaudehome/s... | closed | not_planned | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-25T20:59:59Z | 2026-04-29T04:32:16Z | 2026-04-26T09:26:56Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | xujintao | 5,418,373 | MDQ6VXNlcjU0MTgzNzM= | User | false |
openclaw/openclaw | 4,198,494,817 | I_kwDOQb6kR876P_Jh | 60,092 | https://github.com/openclaw/openclaw/issues/60092 | https://api.github.com/repos/openclaw/openclaw/issues/60092 | [Bug]: DEFAULT IT TO 0.0.0.0 for the love of all that is holy . | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Default binding to 0.0.0.0 instead of localhost reduces the pain process of trying to make this accessible on the network .
### Steps to reproduce
config.json
### Expected behavior
be able to just access on the network. ... | closed | not_planned | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-03T05:38:44Z | 2026-04-28T04:46:08Z | 2026-04-25T21:00:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | cdvankammen | 31,197,020 | MDQ6VXNlcjMxMTk3MDIw | User | false |
openclaw/openclaw | 3,883,159,717 | I_kwDOQb6kR87ndFCl | 6,654 | https://github.com/openclaw/openclaw/issues/6654 | https://api.github.com/repos/openclaw/openclaw/issues/6654 | [Feature]: Use zerobrew instead of homebrew | ## Summary
[zerobrew](https://github.com/lucasgelfond/zerobrew) seems much faster (5x-20x) than homebrew. Granted it's only used to install skills, but would it be a good update?
## Proposed solution
1) Propose zerobrew installation instead of homebrew during setup.
2) In skill installations, try to use zerobrew fi... | closed | completed | true | 1 | [
"enhancement",
"r: third-party-extension"
] | [] | 2026-02-01T22:45:19Z | 2026-04-28T04:46:07Z | 2026-04-25T21:05:07Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jmoraispk | 19,170,303 | MDQ6VXNlcjE5MTcwMzAz | User | false |
openclaw/openclaw | 4,329,224,686 | I_kwDOQb6kR88AAAABAgq57g | 71,770 | https://github.com/openclaw/openclaw/issues/71770 | https://api.github.com/repos/openclaw/openclaw/issues/71770 | Feature Request: Automatic Token-Based Model Failover | As discussed with the user (Bhushan), OpenClaw would benefit from a feature that automatically switches between configured LLM providers (e.g., Gemini and OpenAI) based on real-time token consumption thresholds. \n\n**Problem:** Users currently have to manually monitor and switch models to optimize for cost or API limi... | open | null | false | 1 | [] | [] | 2026-04-25T21:09:58Z | 2026-05-02T16:55:50Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | kountilya | 44,292,019 | MDQ6VXNlcjQ0MjkyMDE5 | User | false |
openclaw/openclaw | 4,329,255,119 | I_kwDOQb6kR88AAAABAgswzw | 71,775 | https://github.com/openclaw/openclaw/issues/71775 | https://api.github.com/repos/openclaw/openclaw/issues/71775 | [Bug]: Amnesia | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Last night, while I was performing a task via WeChat, I went to sleep around 11 p.m. When I woke up at 5 a.m. this morning and resumed the conversation from yesterday, he had completely forgotten our previous dialogue. Durin... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-04-25T21:21:51Z | 2026-05-03T04:35:15Z | 2026-04-30T19:40:51Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | caikun2018 | 39,227,100 | MDQ6VXNlcjM5MjI3MTAw | User | false |
openclaw/openclaw | 4,329,257,653 | I_kwDOQb6kR88AAAABAgs6tQ | 71,776 | https://github.com/openclaw/openclaw/issues/71776 | https://api.github.com/repos/openclaw/openclaw/issues/71776 | gateway.cmd does not inherit system PATH when run under macOS LaunchAgent | ## Summary
When OpenClaw is launched via macOS LaunchAgent (e.g., for headless cron scheduling on a Mac), `gateway.cmd` runs with a minimal PATH that doesn't include user-installed binaries. Tool invocations that depend on `node`, `python3`, `git`, etc. either fail silently or pick up wrong versions (often falling bac... | closed | completed | true | 1 | [
"app: macos",
"gateway",
"maintainer",
"bug:behavior"
] | [] | 2026-04-25T21:22:45Z | 2026-04-30T04:32:54Z | 2026-04-27T22:47:50Z | CONTRIBUTOR | null | 20260430T060018Z | 2026-04-30T06:00:18Z | fabianwilliams | 92,543,063 | U_kgDOBYQYVw | User | false |
openclaw/openclaw | 4,329,258,601 | I_kwDOQb6kR88AAAABAgs-aQ | 71,777 | https://github.com/openclaw/openclaw/issues/71777 | https://api.github.com/repos/openclaw/openclaw/issues/71777 | `cron run` manual trigger doesn't faithfully reproduce the scheduled execution environment | ## Summary
Running `cron run <job>` manually (for testing/debugging) behaves differently from the scheduled run. Observed differences include working directory, environment variables, and session context loading. This makes cron bugs hard to reproduce on a fresh shell — you can't trust that "it works when I run it man... | open | null | false | 1 | [
"agents",
"maintainer",
"bug:behavior"
] | [] | 2026-04-25T21:23:04Z | 2026-04-30T20:55:07Z | null | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | fabianwilliams | 92,543,063 | U_kgDOBYQYVw | User | false |
openclaw/openclaw | 4,329,313,019 | I_kwDOQb6kR88AAAABAgwS-w | 71,783 | https://github.com/openclaw/openclaw/issues/71783 | https://api.github.com/repos/openclaw/openclaw/issues/71783 | QMD collection name mismatch causes memory search failures on boot | ## Bug Summary
### Symptom
On boot, OpenClaw fails to add the `life` memory collection and all QMD memory searches fail with `Collection not found: life`. Log output:
```
qmd collection add /Users/clark/life --name life --glob **/*.md failed (code 1):
A collection already exists for this path and pattern:
Name: lif... | open | null | false | 1 | [] | [] | 2026-04-25T21:44:11Z | 2026-05-02T21:50:20Z | null | NONE | null | 20260503T000641Z | 2026-05-03T00:06:41Z | andrewtclark | 20,686,658 | MDQ6VXNlcjIwNjg2NjU4 | User | false |
openclaw/openclaw | 4,329,313,012 | I_kwDOQb6kR88AAAABAgwS9A | 71,782 | https://github.com/openclaw/openclaw/issues/71782 | https://api.github.com/repos/openclaw/openclaw/issues/71782 | MEMORY.md bootstrap truncation silently drops context with no warning | ## Bug Summary
### Symptom
On agent session start, the workspace bootstrap file is truncated in injected context:
```
workspace bootstrap file MEMORY.md is 14296 chars (limit 12000); truncating in injected context
```
### Impact
The agent loses the tail of its MEMORY.md on every session, silently dropping memory ent... | closed | completed | true | 1 | [] | [] | 2026-04-25T21:44:11Z | 2026-04-29T04:39:08Z | 2026-04-26T09:27:06Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | andrewtclark | 20,686,658 | MDQ6VXNlcjIwNjg2NjU4 | User | false |
openclaw/openclaw | 4,329,329,049 | I_kwDOQb6kR88AAAABAgxRmQ | 71,784 | https://github.com/openclaw/openclaw/issues/71784 | https://api.github.com/repos/openclaw/openclaw/issues/71784 | Bug: memory search live embedding fails ~20–40% with `fetch failed | other side closed` (provider-agnostic; upstream healthy) | # Bug: Memory search transient `fetch failed | other side closed` / `Client network socket disconnected before secure TLS connection was established` for live embedding queries (provider-agnostic)
## Summary
Live memory search queries fail intermittently (~20–40% of calls) with one of two transient TLS/socket errors ... | open | null | false | 3 | [] | [] | 2026-04-25T21:50:50Z | 2026-05-02T16:41:20Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | kevinheinrichs | 113,271,657 | U_kgDOBsBjaQ | User | false |
openclaw/openclaw | 4,329,341,216 | I_kwDOQb6kR88AAAABAgyBIA | 71,785 | https://github.com/openclaw/openclaw/issues/71785 | https://api.github.com/repos/openclaw/openclaw/issues/71785 | Control UI webchat: duplicate assistant messages after update to 2026.4.24 | ## Description
After updating to OpenClaw **2026.4.24**, the webchat Control UI renders **duplicate assistant messages** on every reply. Each assistant response appears twice in the chat history.
## Environment
- **OpenClaw version:** 2026.4.24 (cbcfdf6)
- **OS:** Windows 11
- **Channel:** webchat (Control UI)
- **S... | closed | completed | true | 3 | [] | [] | 2026-04-25T21:57:04Z | 2026-04-29T04:42:28Z | 2026-04-26T07:06:10Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | duong24mse23163-svg | 271,048,958 | U_kgDOECfg_g | User | false |
openclaw/openclaw | 4,209,934,492 | I_kwDOQb6kR8767oCc | 61,713 | https://github.com/openclaw/openclaw/issues/61713 | https://api.github.com/repos/openclaw/openclaw/issues/61713 | [Bug]: macOS: agent exec / gateway exec cannot reach SSH host, while terminal and direct runExecProcess work | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to 2026.4.2 (d74a122) on macOS, OpenClaw agent/gateway exec can no longer reach an SSH NAS target ("No route to host"), while the same host succeeds from Terminal, direct Node spawn, and direct bundled runExe... | open | null | false | 2 | [
"bug",
"regression"
] | [] | 2026-04-06T06:10:07Z | 2026-05-02T09:30:08Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | Tangdou0322 | 32,410,256 | MDQ6VXNlcjMyNDEwMjU2 | User | false |
openclaw/openclaw | 4,329,366,295 | I_kwDOQb6kR88AAAABAgzjFw | 71,789 | https://github.com/openclaw/openclaw/issues/71789 | https://api.github.com/repos/openclaw/openclaw/issues/71789 | [Performance]: Windows CLI startup takes 120+ seconds - plugin discovery cache ineffective across processes | **Problem:**
CLI commands (openclaw status, etc.) take 118-124 seconds to start on Windows.
**Root Cause:**
The plugin discovery cache is in-memory only (const discoveryCache = new Map()). Each CLI invocation is a new Node.js process, so the cache is always empty on startup. The DEFAULT_DISCOVERY_CACHE_MS = 1000 (1 ... | closed | completed | true | 1 | [] | [] | 2026-04-25T22:10:54Z | 2026-04-29T04:39:07Z | 2026-04-26T09:27:09Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | liuying1991 | 106,902,534 | U_kgDOBl80Bg | User | false |
openclaw/openclaw | 4,329,371,669 | I_kwDOQb6kR88AAAABAgz4FQ | 71,790 | https://github.com/openclaw/openclaw/issues/71790 | https://api.github.com/repos/openclaw/openclaw/issues/71790 | [Bug]: healthcheck reports Tailscale "off" when Tailscale daemon is running | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The healthcheck skill (and/or `openclaw security audit --deep`) reports
"Tailscale: vypnutý/off" even when Tailscale is fully active and connected
to a private tailnet.
### Steps to reproduce
1. Mac mini ... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-25T22:14:00Z | 2026-04-28T04:44:14Z | 2026-04-26T00:58:11Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | pesvobodak | 46,964,858 | MDQ6VXNlcjQ2OTY0ODU4 | User | false |
openclaw/openclaw | 4,329,384,153 | I_kwDOQb6kR88AAAABAg0o2Q | 71,794 | https://github.com/openclaw/openclaw/issues/71794 | https://api.github.com/repos/openclaw/openclaw/issues/71794 | [Bug]: 4.23 silently breaks agent_end hooks for non-bundled plugins (openclaw-mem0 auto-capture) | ## Bug Description
After upgrading to OpenClaw 2026.4.23, the `openclaw-mem0` plugin's `agent_end` hook is silently blocked with:
```
[gateway] typed hook "agent_end" blocked because non-bundled plugins must set plugins.entries.openclaw-mem0.hooks.allowConversationAccess=true
```
This means **auto-capture stops work... | closed | completed | true | 1 | [] | [] | 2026-04-25T22:21:17Z | 2026-04-29T04:36:20Z | 2026-04-26T12:06:27Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Veda-openclaw | 267,342,076 | U_kgDOD-9Q_A | User | false |
openclaw/openclaw | 4,329,384,231 | I_kwDOQb6kR88AAAABAg0pJw | 71,795 | https://github.com/openclaw/openclaw/issues/71795 | https://api.github.com/repos/openclaw/openclaw/issues/71795 | Mattermost: All inbound messages get System: prefix, Discord doesn't | ## Bug Description
Every inbound message from Mattermost is wrapped with a `System:` prefix before being submitted to the model. Discord (same agent, same session, same model) does not have this issue.
**Mattermost (broken) — first line of every prompt.submitted:**
```
System: [2026-04-26 05:26:13 GMT+8] Mattermost m... | closed | completed | true | 1 | [] | [] | 2026-04-25T22:21:19Z | 2026-05-01T04:44:26Z | 2026-04-28T08:41:49Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | JeyoWang | 268,744,013 | U_kgDOEAS1TQ | User | false |
openclaw/openclaw | 4,329,391,998 | I_kwDOQb6kR88AAAABAg1Hfg | 71,798 | https://github.com/openclaw/openclaw/issues/71798 | https://api.github.com/repos/openclaw/openclaw/issues/71798 | [Bug]: Matrix cron announce lowercases room IDs and scheduled delivery fails | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Matrix cron announce delivery stores the Matrix room ID lowercased when creating a scheduled job, so the agent generates the scheduled final message but delivery fails because Matrix room IDs are case-sensitive... | closed | completed | true | 1 | [] | [] | 2026-04-25T22:25:52Z | 2026-04-28T04:43:57Z | 2026-04-26T01:22:06Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | nickdecooman | 2,564,899 | MDQ6VXNlcjI1NjQ4OTk= | User | false |
openclaw/openclaw | 4,329,398,110 | I_kwDOQb6kR88AAAABAg1fXg | 71,800 | https://github.com/openclaw/openclaw/issues/71800 | https://api.github.com/repos/openclaw/openclaw/issues/71800 | logging.maxFileBytes config value not applied at gateway start (likely logger init race) | ## Summary
Setting \`logging.maxFileBytes\` in \`~/.openclaw/openclaw.json\` to a value above the default 500 MB is not picked up by the gateway logger. The default \`DEFAULT_MAX_LOG_FILE_BYTES = 500 * 1024 * 1024\` continues to apply across multiple gateway restarts (\`launchctl bootout\`+\`bootstrap\`), even though ... | closed | completed | true | 1 | [] | [] | 2026-04-25T22:29:27Z | 2026-04-29T04:37:38Z | 2026-04-26T10:22:51Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | 1OAKDesign | 143,361,737 | U_kgDOCIuGyQ | User | false |
openclaw/openclaw | 4,329,220,096 | I_kwDOQb6kR88AAAABAgqoAA | 71,769 | https://github.com/openclaw/openclaw/issues/71769 | https://api.github.com/repos/openclaw/openclaw/issues/71769 | [Bug]: gateway crash loop on 2026.4.24 (bonjour plugin) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Bug report (filed upstream)
### Summary
On 2026.4.24, the gateway enters an unrecoverable crash loop on systems
where the bonjour service stays in \`announcing\` state. The bonjour plugin
defines a handler sp... | closed | duplicate | true | 2 | [
"bug",
"duplicate",
"dedupe:child",
"regression"
] | [] | 2026-04-25T21:07:31Z | 2026-04-28T04:45:54Z | 2026-04-25T22:30:48Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jplavoiemtl | 36,643,005 | MDQ6VXNlcjM2NjQzMDA1 | User | false |
openclaw/openclaw | 4,329,233,099 | I_kwDOQb6kR88AAAABAgrayw | 71,771 | https://github.com/openclaw/openclaw/issues/71771 | https://api.github.com/repos/openclaw/openclaw/issues/71771 | [v2026.4.24] Gateway crash loop: CIAO ANNOUNCEMENT CANCELLED in [plugins] bonjour | ## Summary
After upgrading from `2026.4.23` to `2026.4.24` (commit `cbcfdf6`) via `npm i -g openclaw` (manual-update), the gateway entered a tight crash loop with an unhandled promise rejection from the bonjour mDNS advertiser. Crashes occurred roughly every ~16 seconds. Downgrading to `2026.4.23` (commit `a979721`) im... | closed | duplicate | true | 2 | [
"duplicate",
"dedupe:child"
] | [] | 2026-04-25T21:13:35Z | 2026-04-28T04:45:53Z | 2026-04-25T22:30:52Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | spartoviMD | 179,311,308 | U_kgDOCrASzA | User | false |
openclaw/openclaw | 4,329,301,767 | I_kwDOQb6kR88AAAABAgvnBw | 71,781 | https://github.com/openclaw/openclaw/issues/71781 | https://api.github.com/repos/openclaw/openclaw/issues/71781 | Bonjour plugin crashes gateway in restart loop on macOS (CIAO ANNOUNCEMENT CANCELLED) | ## Bug Summary
**Regression introduced ~2026.4.3–2026.4.24**
### Symptom
`openclaw gateway` crashes repeatedly with `Unhandled promise rejection: CIAO ANNOUNCEMENT CANCELLED` every ~34 seconds. The gateway reaches ready state but never stabilizes.
### Root Cause (two parts)
1. **Stuck mDNS announcement:** The CIAO ... | closed | duplicate | true | 1 | [
"duplicate",
"dedupe:child"
] | [] | 2026-04-25T21:40:34Z | 2026-04-28T04:45:53Z | 2026-04-25T22:30:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | andrewtclark | 20,686,658 | MDQ6VXNlcjIwNjg2NjU4 | User | false |
openclaw/openclaw | 4,329,397,434 | I_kwDOQb6kR88AAAABAg1cug | 71,799 | https://github.com/openclaw/openclaw/issues/71799 | https://api.github.com/repos/openclaw/openclaw/issues/71799 | 2026.4.24 gateway crash loop: Unhandled promise rejection 'CIAO PROBING CANCELLED' from bonjour | ## Summary
After upgrading from `2026.4.23` to `2026.4.24` on macOS, the gateway enters a crash-loop on startup. Each gateway process exits within ~25 seconds with an unhandled promise rejection from the bonjour mDNS plugin during its announcing/probing phase. `openclaw doctor` cannot complete because the gateway is n... | closed | duplicate | true | 1 | [
"duplicate",
"dedupe:child"
] | [] | 2026-04-25T22:29:02Z | 2026-04-28T04:45:52Z | 2026-04-25T22:31:01Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | 1OAKDesign | 143,361,737 | U_kgDOCIuGyQ | User | false |
openclaw/openclaw | 4,329,406,854 | I_kwDOQb6kR88AAAABAg2Bhg | 71,802 | https://github.com/openclaw/openclaw/issues/71802 | https://api.github.com/repos/openclaw/openclaw/issues/71802 | [Bug]: 2026.4.24 regression — ENOTEMPTY on bundled plugin-sdk causes feishu (and potentially all bundled channels) to fail loading | ## Bug Summary
After upgrading to OpenClaw 2026.4.24, the bundled runtime dependency repair logic uses non-recursive `fs.rmdir()` to remove the `plugin-sdk` directory, which contains 300+ files. This causes `ENOTEMPTY` error on every startup, preventing feishu (and potentially all bundled channel plugins) from loading... | closed | completed | true | 1 | [] | [] | 2026-04-25T22:34:30Z | 2026-04-28T04:45:23Z | 2026-04-26T00:18:28Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | colin-chang | 24,368,158 | MDQ6VXNlcjI0MzY4MTU4 | User | false |
openclaw/openclaw | 4,329,409,900 | I_kwDOQb6kR88AAAABAg2NbA | 71,803 | https://github.com/openclaw/openclaw/issues/71803 | https://api.github.com/repos/openclaw/openclaw/issues/71803 | CLI watchdog kills sessions that are correctly idle while waiting on a Monitor task | ## CLI watchdog kills sessions that are correctly idle while waiting on a Monitor task
### Summary
The `agent/cli-backend` watchdog terminates the Claude CLI process after `noOutputTimeoutMs` (default 180s) of no stdout, even when the agent is deliberately idle inside a `Monitor` tool call waiting for a long-running s... | open | null | false | 1 | [] | [] | 2026-04-25T22:36:14Z | 2026-05-02T16:55:52Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | beaujohnson1 | 192,929,506 | U_kgDOC3_e4g | User | false |
openclaw/openclaw | 4,329,411,519 | I_kwDOQb6kR88AAAABAg2Tvw | 71,804 | https://github.com/openclaw/openclaw/issues/71804 | https://api.github.com/repos/openclaw/openclaw/issues/71804 | [Bug]: 2026.4.24 regression — Discord exec approval no longer sends interactive prompt via DM; approval callbacks loop with repeated errors after user approves | ## Bug Summary
Two related regressions in OpenClaw 2026.4.24 affecting Discord exec approval flow:
1. **Approval prompts no longer appear in Discord DM** — Before upgrading, exec approval requests were sent as interactive Discord messages (with Allow/Deny buttons) via DM. After upgrading to 2026.4.24, approval prompt... | closed | duplicate | true | 1 | [] | [] | 2026-04-25T22:37:11Z | 2026-04-28T04:45:20Z | 2026-04-26T00:20:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | colin-chang | 24,368,158 | MDQ6VXNlcjI0MzY4MTU4 | User | false |
openclaw/openclaw | 4,329,412,952 | I_kwDOQb6kR88AAAABAg2ZWA | 71,805 | https://github.com/openclaw/openclaw/issues/71805 | https://api.github.com/repos/openclaw/openclaw/issues/71805 | 2026.4.24: ENOTEMPTY race on plugin-sdk during parallel plugin/channel bootstrap | ## Summary
On `openclaw@2026.4.24` (commit `cbcfdf6`), the gateway fails to fully boot because multiple bundled plugin/channel loaders race on the same shared `plugin-sdk` directory under `~/.openclaw/plugin-runtime-deps/openclaw-2026.4.24-*/dist/extensions/node_modules/openclaw/plugin-sdk`. One loader wins; the rest c... | closed | completed | true | 1 | [] | [] | 2026-04-25T22:38:03Z | 2026-04-28T04:45:22Z | 2026-04-26T00:18:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jsc2304 | 181,968,212 | U_kgDOCtidVA | User | false |
openclaw/openclaw | 4,329,383,401 | I_kwDOQb6kR88AAAABAg0l6Q | 71,793 | https://github.com/openclaw/openclaw/issues/71793 | https://api.github.com/repos/openclaw/openclaw/issues/71793 | claude-cli live session: hardcoded 256KB stdout buffer causes spurious "Claude CLI stdout buffer exceeded limit." failures | ## Summary
Since `2026.4.22` (warm stdio sessions for `claude-cli`, #69679), embedded agents using `claude-cli` provider intermittently fail with:
```
Embedded agent failed before reply: Claude CLI stdout buffer exceeded limit.
```
The error is non-recoverable for the current session (user has to `/new`) and surface... | closed | completed | true | 1 | [] | [] | 2026-04-25T22:20:50Z | 2026-04-28T04:42:48Z | 2026-04-26T02:36:07Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | chacher86 | 54,685,632 | MDQ6VXNlcjU0Njg1NjMy | User | false |
openclaw/openclaw | 4,189,037,839 | I_kwDOQb6kR875r6UP | 59,201 | https://github.com/openclaw/openclaw/issues/59201 | https://api.github.com/repos/openclaw/openclaw/issues/59201 | sendAnnounce queue path passes deliver:true without channel for sessions with no origin | ## Bug
When the gateway has multiple channels configured (e.g., telegram + bluebubbles), the `sendAnnounce` function in the announce queue path (`src/agents/subagent-announce-delivery.ts`) unconditionally sets `deliver: !requesterIsSubagent`.
For Paperclip adapter sessions (`agent:eng:paperclip:issue:*`), which have ... | closed | completed | true | 1 | [
"bug",
"agents"
] | [] | 2026-04-01T18:31:32Z | 2026-04-28T04:45:49Z | 2026-04-25T22:49:24Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | larrylhollan | 265,963,893 | U_kgDOD9pJdQ | User | false |
openclaw/openclaw | 4,329,434,216 | I_kwDOQb6kR88AAAABAg3saA | 71,810 | https://github.com/openclaw/openclaw/issues/71810 | https://api.github.com/repos/openclaw/openclaw/issues/71810 | [Bug]: OpenClaw 2026.4.24 — Telegram Channel Startup Failures | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After updating to 2026.4.24, the Telegram bundled channel fails to load consistently on startup due to cascading ENOTEMPTY errors in the shared node_modules directory (/home/<user>/node_modules/). This causes the gateway to ... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-25T22:51:02Z | 2026-04-28T04:45:21Z | 2026-04-26T00:18:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | maxdes63 | 166,340,227 | U_kgDOCeomgw | User | false |
openclaw/openclaw | 4,329,439,165 | I_kwDOQb6kR88AAAABAg3_vQ | 71,811 | https://github.com/openclaw/openclaw/issues/71811 | https://api.github.com/repos/openclaw/openclaw/issues/71811 | ACP harness sessions refuse subagent announce: `<<<BEGIN_OPENCLAW_INTERNAL_CONTEXT>>>` envelope is sent into harnesses that don't have OpenClaw's system prompt | # Bug: OpenClaw injects `<<<BEGIN_OPENCLAW_INTERNAL_CONTEXT>>>` envelopes into ACP harness sessions
**Filed:** 2026-04-25
**Severity:** Medium — breaks subagent/cron announce delivery to ACP-routed orchestrator sessions; harness refuses (correctly) and the user-facing reply ends up being a "this looks like prompt inje... | closed | completed | true | 1 | [] | [] | 2026-04-25T22:53:59Z | 2026-04-28T04:42:32Z | 2026-04-26T02:43:28Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jordiagi | 257,037,191 | U_kgDOD1IThw | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.