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 | 4,354,290,210 | I_kwDOQb6kR88AAAABA4kyIg | 74,676 | https://github.com/openclaw/openclaw/issues/74676 | https://api.github.com/repos/openclaw/openclaw/issues/74676 | [Bug]: api.resolvePath returns undefined for extension plugins during service start | ## Description
`api.resolvePath('.')` returns `undefined` when called from within a plugin's registered service `start()` function for extension-based (non-stock) plugins installed under `~/.openclaw/extensions/`. This causes any extension plugin that uses `api.resolvePath` for file I/O during service startup to fail.... | closed | completed | true | 3 | [] | [] | 2026-04-29T23:33:56Z | 2026-05-02T04:26:24Z | 2026-04-30T03:14:36Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | jimdawdy-hub | 262,052,366 | U_kgDOD56aDg | User | false |
openclaw/openclaw | 4,354,307,175 | I_kwDOQb6kR88AAAABA4l0Zw | 74,679 | https://github.com/openclaw/openclaw/issues/74679 | https://api.github.com/repos/openclaw/openclaw/issues/74679 | Bug: x-openclaw-session-key header bypasses agent routing on /v1/chat/completions | ## Summary
When `x-openclaw-session-key` is sent as an HTTP request header on the `/v1/chat/completions` endpoint, agent routing via the `model: "openclaw/<agentId>"` field is silently ignored. The request is handled by the default (main) agent regardless of the model field. Removing the header and using `sessionKey` ... | open | null | false | 1 | [] | [] | 2026-04-29T23:39:27Z | 2026-05-04T02:38:35Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | R2D4agent | 272,577,753 | U_kgDOED802Q | User | false |
openclaw/openclaw | 4,354,335,208 | I_kwDOQb6kR88AAAABA4nh6A | 74,681 | https://github.com/openclaw/openclaw/issues/74681 | https://api.github.com/repos/openclaw/openclaw/issues/74681 | Non-deterministic timeout enforcement in isolated agentTurn runner (timeoutSeconds unreliable) | ## Bug Report: Non-deterministic timeout enforcement in isolated agent runner
**OpenClaw version:** 2026.4.26
**Date observed:** 2026-04-30
**Environment:** macOS Darwin 25.4.0 (arm64), node v22.22.1
### Summary
`timeoutSeconds` on cron `agentTurn` jobs is not reliably enforced in isolated session runs. The timeout ... | open | null | false | 3 | [] | [] | 2026-04-29T23:47:14Z | 2026-05-03T04:44:16Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | list3r | 1,835,336 | MDQ6VXNlcjE4MzUzMzY= | User | false |
openclaw/openclaw | 4,354,340,037 | I_kwDOQb6kR88AAAABA4n0xQ | 74,682 | https://github.com/openclaw/openclaw/issues/74682 | https://api.github.com/repos/openclaw/openclaw/issues/74682 | [Bug]: 2026.4.26 — auth resolver fails to attach Authorization header to OpenAI requests; all openai/* and codex/* models 401 with "Missing bearer" | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
On OpenClaw 2026.4.26 with agentRuntime.id=codex, every outbound OpenAI request to https://api.openai.com/v1/responses
returns 401 "Missing bearer or basic authentication in header" across both auth modes (api_key via ope... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-29T23:48:40Z | 2026-05-02T04:26:52Z | 2026-04-29T23:54:08Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | anonomouslylearning444 | 280,545,887 | U_kgDOELjKXw | User | false |
openclaw/openclaw | 4,354,343,671 | I_kwDOQb6kR88AAAABA4oC9w | 74,683 | https://github.com/openclaw/openclaw/issues/74683 | https://api.github.com/repos/openclaw/openclaw/issues/74683 | [Bug]: MSTeams "channel unresolved" error | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
I have my MS Teams team and channel IDs configured as the docs say here: https://docs.openclaw.ai/channels/msteams#team-and-channel-ids-common-gotcha
Yet when the gateway starts, I see the log message:
```
[m... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-29T23:49:52Z | 2026-05-02T04:26:39Z | 2026-04-30T02:01:40Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | dseravalli | 1,758,219 | MDQ6VXNlcjE3NTgyMTk= | User | false |
openclaw/openclaw | 4,354,363,336 | I_kwDOQb6kR88AAAABA4pPyA | 74,684 | https://github.com/openclaw/openclaw/issues/74684 | https://api.github.com/repos/openclaw/openclaw/issues/74684 | [Bug]: sessions_spawn does not expose child process PID for running ACP sessions | ## Summary
`sessions_spawn(runtime="acp", ...)` returns session identifiers such as `childSessionKey` / `sessionKey`, but does not expose the OS process PID(s) for the spawned ACP adapter / child agent process. When the child appears stuck, is still running, or has empty projected history, operators have no first-clas... | open | null | false | 1 | [] | [] | 2026-04-29T23:55:34Z | 2026-05-03T09:49:53Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | El-Ingenier0 | 84,151,108 | MDQ6VXNlcjg0MTUxMTA4 | User | false |
openclaw/openclaw | 4,354,394,456 | I_kwDOQb6kR88AAAABA4rJWA | 74,688 | https://github.com/openclaw/openclaw/issues/74688 | https://api.github.com/repos/openclaw/openclaw/issues/74688 | v2026.4.27: `plugin-runtime-deps` tarball is missing the `chokidar` dependency, breaking all CLI subcommands at first `import` | # v2026.4.27: `plugin-runtime-deps` tarball is missing the `chokidar` dependency, breaking all CLI subcommands at first `import`
## Summary
After a fresh install of `openclaw@2026.4.27`, the gateway boots and runs cleanly, but **every CLI subcommand that hits the plugin manager** (e.g. `openclaw memory promote`) fail... | closed | completed | true | 1 | [] | [] | 2026-04-30T00:03:53Z | 2026-05-02T04:26:51Z | 2026-04-30T00:08:50Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | haumanto | 72,009,799 | MDQ6VXNlcjcyMDA5Nzk5 | User | false |
openclaw/openclaw | 4,354,487,554 | I_kwDOQb6kR88AAAABA4w1Ag | 74,694 | https://github.com/openclaw/openclaw/issues/74694 | https://api.github.com/repos/openclaw/openclaw/issues/74694 | [Bug]: Telegram group chats are not ingested/responded to by OpenClaw bot, while direct messages work normally, with evidence of group-to-supergroup migration handling | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Telegram group chats are not ingested/responded to by OpenClaw bot, while direct messages work normally, with evidence of group-to-supergroup migration handling
### Steps to reproduce
1. Configure Telegram bot in OpenClaw
... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-04-30T00:24:50Z | 2026-05-02T04:26:50Z | 2026-04-30T00:29:38Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | kabes1123 | 280,476,456 | U_kgDOELe7KA | User | false |
openclaw/openclaw | 4,145,021,248 | I_kwDOQb6kR873EAFA | 55,249 | https://github.com/openclaw/openclaw/issues/55249 | https://api.github.com/repos/openclaw/openclaw/issues/55249 | Feature: Session labels / nicknames for easier identification | ## Problem
When listing sessions via `openclaw sessions`, every session is identified by auto-generated keys like:
```
agent:main:main
agent:main:msteams:channel:19:44ff67b4064941d0b45ccba274c3e403@thread.tacv2
agent:main:webchat:gateway-client
```
These are opaque and hard to distinguish at a glance, especially whe... | open | null | false | 2 | [] | [] | 2026-03-26T15:27:58Z | 2026-05-02T01:14:43Z | null | NONE | null | 20260502T030029Z | 2026-05-02T03:00:29Z | tazmon95 | 3,894,763 | MDQ6VXNlcjM4OTQ3NjM= | User | false |
openclaw/openclaw | 4,354,626,506 | I_kwDOQb6kR88AAAABA45Tyg | 74,707 | https://github.com/openclaw/openclaw/issues/74707 | https://api.github.com/repos/openclaw/openclaw/issues/74707 | Gateway RPC: add SDK-facing task ledger APIs | ## Goal
Expose background/detached work through SDK-facing task ledger RPCs.
The SDK has task nouns, and OpenClaw already has background activity concepts. External apps need a stable way to show subagents, ACP jobs, cron-like jobs, CLI tasks, and async work without reading internal session logs.
## Proposed methods... | open | null | false | 1 | [] | [] | 2026-04-30T00:59:36Z | 2026-04-30T01:03:53Z | null | MEMBER | null | 20260430T030024Z | 2026-04-30T03:00:24Z | BunsDev | 68,980,965 | MDQ6VXNlcjY4OTgwOTY1 | User | false |
openclaw/openclaw | 4,354,626,603 | I_kwDOQb6kR88AAAABA45UKw | 74,708 | https://github.com/openclaw/openclaw/issues/74708 | https://api.github.com/repos/openclaw/openclaw/issues/74708 | Gateway RPC: add SDK-facing environment discovery APIs | ## Goal
Start SDK environment support with read-only Gateway RPCs for discovery/status.
The SDK design includes local, Gateway, node, managed, and ephemeral environments. Before create/delete/provisioning, app clients should be able to discover what environments exist and what capabilities they expose.
## Proposed f... | open | null | false | 2 | [
"maintainer"
] | [] | 2026-04-30T00:59:38Z | 2026-05-02T05:27:04Z | null | MEMBER | null | 20260502T060142Z | 2026-05-02T06:01:42Z | BunsDev | 68,980,965 | MDQ6VXNlcjY4OTgwOTY1 | User | false |
openclaw/openclaw | 4,354,626,710 | I_kwDOQb6kR88AAAABA45Ulg | 74,709 | https://github.com/openclaw/openclaw/issues/74709 | https://api.github.com/repos/openclaw/openclaw/issues/74709 | Docs: clarify app SDK vs plugin SDK boundary | ## Goal
Make the app SDK vs plugin SDK distinction unmistakable in docs and examples.
The new `@openclaw/sdk` package is for external applications and automations talking to Gateway. The existing plugin SDK is for trusted code running inside OpenClaw to register tools, channels, providers, hooks, and runtimes.
## Ac... | open | null | false | 1 | [
"maintainer"
] | [] | 2026-04-30T00:59:41Z | 2026-04-30T01:04:04Z | null | MEMBER | null | 20260430T030024Z | 2026-04-30T03:00:24Z | BunsDev | 68,980,965 | MDQ6VXNlcjY4OTgwOTY1 | User | false |
openclaw/openclaw | 4,354,626,272 | I_kwDOQb6kR88AAAABA45S4A | 74,706 | https://github.com/openclaw/openclaw/issues/74706 | https://api.github.com/repos/openclaw/openclaw/issues/74706 | Gateway RPC: add SDK-facing artifacts APIs | ## Goal
Add SDK-facing artifact RPCs so external apps can discover and render rich outputs.
OpenMeow and other app clients will need stable access to generated files, diffs, screenshots, logs, media, trajectories, and PR links without parsing runtime internals.
## Proposed methods
- `artifacts.list`
- `artifacts.ge... | closed | completed | true | 1 | [] | [] | 2026-04-30T00:59:33Z | 2026-05-03T04:34:34Z | 2026-05-01T00:35:50Z | MEMBER | null | 20260503T060030Z | 2026-05-03T06:00:30Z | BunsDev | 68,980,965 | MDQ6VXNlcjY4OTgwOTY1 | User | false |
openclaw/openclaw | 4,144,716,296 | I_kwDOQb6kR873C1oI | 55,224 | https://github.com/openclaw/openclaw/issues/55224 | https://api.github.com/repos/openclaw/openclaw/issues/55224 | [Bug] Tool parameters randomly dropped in long sessions (exec/write/sessions_spawn) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
In long sessions (several hundred conversation turns), the exec, write, and sessions_spawn tools randomly drop required parameters. The tool receives an empty object {} instead of the actual parameters, causing... | closed | not_planned | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-03-26T14:53:12Z | 2026-05-02T04:26:47Z | 2026-04-30T01:07:28Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | tonyfeng2018 | 37,775,747 | MDQ6VXNlcjM3Nzc1NzQ3 | User | false |
openclaw/openclaw | 4,147,175,501 | I_kwDOQb6kR873MOBN | 55,363 | https://github.com/openclaw/openclaw/issues/55363 | https://api.github.com/repos/openclaw/openclaw/issues/55363 | [Bug]: WhatsApp message send drops --force-document before gateway payload | ## Summary
On WhatsApp, `openclaw message send --force-document` is accepted by the CLI but does not make it into the outbound payload. The message is still sent through the normal audio/media path instead of being forced into a document payload.
This appears to be a general `message send` / gateway transport gap for... | closed | not_planned | true | 1 | [] | [] | 2026-03-26T20:36:23Z | 2026-05-02T04:26:46Z | 2026-04-30T01:07:35Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | mySebbe | 14,988,278 | MDQ6VXNlcjE0OTg4Mjc4 | User | false |
openclaw/openclaw | 4,162,986,401 | I_kwDOQb6kR874IiGh | 56,837 | https://github.com/openclaw/openclaw/issues/56837 | https://api.github.com/repos/openclaw/openclaw/issues/56837 | Feature Request: Add proxy support for model providers (SOCKS5/HTTP) | null | closed | completed | true | 1 | [] | [] | 2026-03-29T06:30:45Z | 2026-05-02T04:26:46Z | 2026-04-30T01:07:39Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | sysit | 9,957,218 | MDQ6VXNlcjk5NTcyMTg= | User | false |
openclaw/openclaw | 4,187,147,161 | I_kwDOQb6kR875ksuZ | 59,050 | https://github.com/openclaw/openclaw/issues/59050 | https://api.github.com/repos/openclaw/openclaw/issues/59050 | Configurable Workspace File Load Order in System Prompt Assembly | ### Summary
Allow users to configure the injection order of workspace bootstrap files (AGENTS.md, SOUL.md, TOOLS.md, IDENTITY.md, USER.md, HEARTBEAT.md) into the system prompt via `config/default.json` or per-agent config.
### Problem to solve
OpenClaw hardcodes AGENTS.md as the first workspace file injected into th... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-01T13:56:28Z | 2026-05-02T04:26:45Z | 2026-04-30T01:07:45Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | WSW-creator | 261,031,819 | U_kgDOD48Hiw | User | false |
openclaw/openclaw | 4,354,677,134 | I_kwDOQb6kR88AAAABA48Zjg | 74,721 | https://github.com/openclaw/openclaw/issues/74721 | https://api.github.com/repos/openclaw/openclaw/issues/74721 | `/new` command fails with empty messages error on QQ Bot channel | ## Bug Description
Sending `/new` command in QQ Bot (C2C private chat) results in an error: `"⚠️ Something went wrong while processing your request. Please try again, or use /new to start a fresh session."` — which is a circular reference since `/new` itself caused the error.
## Root Cause
From gateway logs, the err... | closed | completed | true | 1 | [] | [] | 2026-04-30T01:14:53Z | 2026-05-02T04:26:42Z | 2026-04-30T01:20:22Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | ZHTodd223 | 279,258,292 | U_kgDOEKUktA | User | false |
openclaw/openclaw | 4,354,679,237 | I_kwDOQb6kR88AAAABA48hxQ | 74,722 | https://github.com/openclaw/openclaw/issues/74722 | https://api.github.com/repos/openclaw/openclaw/issues/74722 | [Feature]: Language-aware TTS routing for inbound voice replies | ### Summary
Add language-aware TTS provider/model/voice routing so auto-TTS replies can use the detected input language instead of a single fixed voice.
### Problem to solve
OpenClaw currently supports `messages.tts.auto: "inbound"`, which correctly controls when audio is returned: text input gets text output, voice... | open | null | false | 2 | [] | [] | 2026-04-30T01:15:42Z | 2026-05-03T04:44:18Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | reisball | 73,487,521 | MDQ6VXNlcjczNDg3NTIx | User | false |
openclaw/openclaw | 4,354,708,353 | I_kwDOQb6kR88AAAABA4-TgQ | 74,726 | https://github.com/openclaw/openclaw/issues/74726 | https://api.github.com/repos/openclaw/openclaw/issues/74726 | MCP tools not accessible to main agent despite bundle-mcp configuration | MCP servers configured under mcp.servers are not accessible to the main OpenClaw agent. The built-in image tool always dispatches to the model provider, never to the MCP server.
Environment: OpenClaw v2026.4.26, tools.profile: coding, mcp.servers configured and running.
The MCP server (minimax-coding-plan-mcp) IS hea... | closed | not_planned | true | 2 | [] | [] | 2026-04-30T01:26:12Z | 2026-05-03T04:38:23Z | 2026-04-30T05:19:26Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | luuq2026 | 268,951,258 | U_kgDOEAfe2g | User | false |
openclaw/openclaw | 4,354,720,231 | I_kwDOQb6kR88AAAABA4_B5w | 74,728 | https://github.com/openclaw/openclaw/issues/74728 | https://api.github.com/repos/openclaw/openclaw/issues/74728 | [Bug]: 2026.4.27 still consumes ~7-20% idle CPU with event loop blocking up to 35s (mitigated but not fixed from 4.26) | ## Bug type
Regression (worked before, now fails)
## Beta release blocker
No
## Summary
**2026.4.27 still consumes ~7-20% idle CPU with periodic event loop blocking up to 35s (mitigated but not fixed from 4.26)**
v2026.4.26's 100% CPU spin bug (#73647, now duplicate of #73532) was partially mitigated in v2026.4.2... | closed | not_planned | true | 1 | [] | [] | 2026-04-30T01:30:01Z | 2026-05-02T04:26:42Z | 2026-04-30T01:34:55Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | leo-afk-sudo | 267,242,069 | U_kgDOD-3KVQ | User | false |
openclaw/openclaw | 4,354,721,214 | I_kwDOQb6kR88AAAABA4_Fvg | 74,729 | https://github.com/openclaw/openclaw/issues/74729 | https://api.github.com/repos/openclaw/openclaw/issues/74729 | [Feishu] message read action 不支持翻页拉取历史消息列表 | ## 问题描述
飞书 channel 的 `message read` action 每次只返回单条消息，无法翻页获取历史消息列表。
## 复现步骤
1. 在飞书群聊中
2. 调用 message read 接口，传入 chatId + limit + before/around 参数
3. 无论传什么参数，始终只返回 messageId 对应的单条消息
## 期望行为
支持分页拉取消息列表，类似飞书 API 的 GET /im/v1/messages 接口：
- 返回多条消息
- 支持 page_token 翻页
- 支持 before/after 时间筛选
## 环境信息
- OpenClaw 版本: 2026.4.23
... | open | null | false | 1 | [] | [] | 2026-04-30T01:30:22Z | 2026-05-03T05:00:23Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | seiya00329 | 277,226,939 | U_kgDOEIYluw | User | false |
openclaw/openclaw | 4,354,694,513 | I_kwDOQb6kR88AAAABA49dcQ | 74,723 | https://github.com/openclaw/openclaw/issues/74723 | https://api.github.com/repos/openclaw/openclaw/issues/74723 | Browser plugin's src/** not emitted to dist; getPwAiModule() returns null in production builds | ## Summary
`@openclaw/browser-plugin` is missing `openclaw.bundle.stageRuntimeDependencies: true` in its `package.json`. As a result, the plugin's `src/**` tree is never compiled to `dist/extensions/browser/src/**`, the runtime's dynamic `import('./pw-ai.js')` fails with `ERR_MODULE_NOT_FOUND`, and every Playwright-ba... | closed | completed | true | 3 | [] | [] | 2026-04-30T01:21:36Z | 2026-05-02T04:26:41Z | 2026-04-30T01:40:09Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | ShadowWalker2014 | 7,563,362 | MDQ6VXNlcjc1NjMzNjI= | User | false |
openclaw/openclaw | 4,354,739,196 | I_kwDOQb6kR88AAAABA5AL_A | 74,732 | https://github.com/openclaw/openclaw/issues/74732 | https://api.github.com/repos/openclaw/openclaw/issues/74732 | docs+feat: Document oMLX (Apple Silicon MLX) as memorySearch embedding provider | ### Summary
Document (and ideally natively support) oMLX — an Apple Silicon-native MLX inference server with an OpenAI-compatible API — as an embedding provider for memorySearch. Currently undocumented; users migrating from Ollama have no clear path.
### Problem to solve
oMLX (https://github.com/jundot/omlx) is an M... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-30T01:36:30Z | 2026-05-03T20:25:14Z | null | NONE | null | 20260503T210034Z | 2026-05-03T21:00:34Z | mppyes-ai | 264,929,184 | U_kgDOD8p_oA | User | false |
openclaw/openclaw | 4,354,478,427 | I_kwDOQb6kR88AAAABA4wRWw | 74,693 | https://github.com/openclaw/openclaw/issues/74693 | https://api.github.com/repos/openclaw/openclaw/issues/74693 | [Bug]: openclaw status crashes loading channel setup fallback runtime loader | ## Summary
`openclaw status` crashes on `2026.4.27` when read-only channel resolution enters the channel setup runtime fallback path for an external channel plugin.
The failure appears to be a packaged/bundled path issue in `dist/read-only-*.js`: it tries to resolve `../../plugins/loader.js` relative to a bundled fil... | closed | completed | true | 4 | [] | [] | 2026-04-30T00:22:56Z | 2026-05-03T04:38:09Z | 2026-04-30T07:08:27Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | giangthb | 45,944,331 | MDQ6VXNlcjQ1OTQ0MzMx | User | false |
openclaw/openclaw | 4,354,715,683 | I_kwDOQb6kR88AAAABA4-wIw | 74,727 | https://github.com/openclaw/openclaw/issues/74727 | https://api.github.com/repos/openclaw/openclaw/issues/74727 | lossless-claw: current-conversation recall can miss persisted turns after session-key drift | ## Summary
`lossless-claw` recall tools can return false negatives when the default/current conversation resolver points at an older active conversation for the same logical user/channel flow, while newer turns are being persisted under a different active conversation/session key.
In this state, `lcm_grep(...)` with ... | closed | not_planned | true | 1 | [] | [] | 2026-04-30T01:28:32Z | 2026-05-02T04:26:39Z | 2026-04-30T02:01:46Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | Adam-Researchh | 261,827,313 | U_kgDOD5sq8Q | User | false |
openclaw/openclaw | 4,354,811,360 | I_kwDOQb6kR88AAAABA5El4A | 74,743 | https://github.com/openclaw/openclaw/issues/74743 | https://api.github.com/repos/openclaw/openclaw/issues/74743 | [Bug]: Custom session cron jobs do not persist in destination session. | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When you create a cron job with --session session:custom-id ([see docs](https://docs.openclaw.ai/automation/cron-jobs)), responses to the cron prompt appear in the destination channel but not in the session log... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-30T02:00:01Z | 2026-05-03T06:50:55Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | DreamNaber | 123,891,618 | U_kgDOB2Jvog | User | false |
openclaw/openclaw | 4,354,851,470 | I_kwDOQb6kR88AAAABA5HCjg | 74,745 | https://github.com/openclaw/openclaw/issues/74745 | https://api.github.com/repos/openclaw/openclaw/issues/74745 | Outbound message rendering: internal metadata label leaking into chat replies | ## Bug Description
Internal session metadata labels are leaking into outbound Telegram messages after the 2026.4.27 update.
## What's happening
Two issues observed in the same session:
1. The literal text `[assistant copied inbound metadata omitted]` is appearing as part of the delivered Telegram message text — it ... | closed | completed | true | 1 | [] | [] | 2026-04-30T02:11:21Z | 2026-05-02T04:26:20Z | 2026-04-30T03:58:21Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | adamwdear | 258,995,361 | U_kgDOD2_0oQ | User | false |
openclaw/openclaw | 4,354,626,015 | I_kwDOQb6kR88AAAABA45R3w | 74,705 | https://github.com/openclaw/openclaw/issues/74705 | https://api.github.com/repos/openclaw/openclaw/issues/74705 | Gateway RPC: add SDK-facing tools.invoke | ## Goal
Add a Gateway RPC method for SDK-facing generic tool invocation.
There is an HTTP `/tools/invoke` route, but `@openclaw/sdk` currently has `oc.tools.invoke()` scaffolded as unsupported. The SDK should have a clean Gateway RPC method with the same policy/approval semantics as other tool invocation paths.
## P... | closed | completed | true | 3 | [] | [] | 2026-04-30T00:59:30Z | 2026-05-04T04:37:12Z | 2026-05-01T08:16:55Z | MEMBER | null | 20260504T060031Z | 2026-05-04T06:00:31Z | BunsDev | 68,980,965 | MDQ6VXNlcjY4OTgwOTY1 | User | false |
openclaw/openclaw | 4,354,865,785 | I_kwDOQb6kR88AAAABA5H6eQ | 74,749 | https://github.com/openclaw/openclaw/issues/74749 | https://api.github.com/repos/openclaw/openclaw/issues/74749 | gemini-cli backend hardcodes unsupported --skip-trust flag — every call fails on gemini-cli ≥ 0.38.x | ## Summary
The Google Gemini CLI backend at `dist/extensions/google/cli-backend.js` passes `--skip-trust` to `gemini`, but current `gemini-cli` (tested 0.38.1 and 0.38.2) doesn't recognize that flag. Yargs rejects with `Unknown arguments: skip-trust, skipTrust` and gemini exits non-zero, so every model attempt fails a... | open | null | false | 1 | [] | [] | 2026-04-30T02:16:16Z | 2026-05-03T08:26:22Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | cemendes | 16,995,556 | MDQ6VXNlcjE2OTk1NTU2 | User | false |
openclaw/openclaw | 4,140,364,506 | I_kwDOQb6kR872yPLa | 54,874 | https://github.com/openclaw/openclaw/issues/54874 | https://api.github.com/repos/openclaw/openclaw/issues/54874 | [Bug] Slow typing in webchat input with keystroke delay | ### Bug type
Regression (worked before, now fails)
### Summary
Summary Typing in the webchat input box is extremely slow - each keystroke causes a noticeable delay of 500ms-1s before the character appears. This makes typing feel unresponsive.
### Steps to reproduce
1. Open Safari or Chrome on Mac Mini<br>2. Navig... | open | null | false | 1 | [
"bug",
"regression"
] | [] | 2026-03-26T04:07:03Z | 2026-05-01T20:10:44Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | xiaotianmdjuk-droid | 265,267,997 | U_kgDOD8-rHQ | User | false |
openclaw/openclaw | 4,144,617,570 | I_kwDOQb6kR873Cdhi | 55,216 | https://github.com/openclaw/openclaw/issues/55216 | https://api.github.com/repos/openclaw/openclaw/issues/55216 | [Bug] Telegram media group: silent image loss when one photo fails to download | ## Summary
When sending multiple images as a Telegram media group (album), if one photo fails to download (network error, API timeout, file too large, etc.), it is **silently dropped** with no notification to the user. The AI receives fewer images than were sent, with no indication of what happened.
**Version:** `202... | open | null | false | 1 | [] | [] | 2026-03-26T14:42:47Z | 2026-05-01T20:10:47Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | Round-Polar-Bear | 84,654,411 | MDQ6VXNlcjg0NjU0NDEx | User | false |
openclaw/openclaw | 4,144,808,223 | I_kwDOQb6kR873DMEf | 55,235 | https://github.com/openclaw/openclaw/issues/55235 | https://api.github.com/repos/openclaw/openclaw/issues/55235 | [Feature]: Generate and update openclaw.json JSON schema on bootstrap/update | ### Summary
Add seamless support for JSON schema for editors
### Problem to solve
After the merge of #54523, it was added the ability to generate a JSON schema was added, but it requires the user to do extra actions to connect it.
### Proposed solution
Generate the json schema on update/bootstrap of OpenCLAW loca... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-26T15:03:36Z | 2026-05-01T20:10:49Z | null | CONTRIBUTOR | null | 20260501T210037Z | 2026-05-01T21:00:37Z | kvokka | 15,954,013 | MDQ6VXNlcjE1OTU0MDEz | User | false |
openclaw/openclaw | 4,144,875,328 | I_kwDOQb6kR873DcdA | 55,239 | https://github.com/openclaw/openclaw/issues/55239 | https://api.github.com/repos/openclaw/openclaw/issues/55239 | 请求支持 Telegram Bot 命令菜单中文化 | ## 描述\nTelegram Bot 注册的命令（如 /tts、/help 等）在 Telegram 聊天界面显示的命令描述是英文的，对于中文用户来说不够友好。\n\n## 期望行为\n希望能够在 OpenClaw 配置中支持为 Telegram Bot 命令设置不同语言的描述，例如添加中文命令说明：\n- /tts → 文字转语音\n- /help → 帮助\n- 等等\n\n## 使用场景\n用户看到的是英文命令说明，看不懂怎么使用。希望能够配置成中文命令说明。\n\n## 补充\n当前语言：简体中文\n操作系统：macOS\nOpenClaw 版本：2026.3.24 | open | null | false | 1 | [] | [] | 2026-03-26T15:10:56Z | 2026-05-01T20:10:50Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | wx2cyj | 123,572,594 | U_kgDOB12Rcg | User | false |
openclaw/openclaw | 4,144,911,553 | I_kwDOQb6kR873DlTB | 55,242 | https://github.com/openclaw/openclaw/issues/55242 | https://api.github.com/repos/openclaw/openclaw/issues/55242 | [Bug] Feishu card message header leaks identity.emoji field containing non-emoji text | ## Bug Description
When sending long messages via Feishu (using card format), the Feishu card header incorrectly displays the full contents of the `identity.emoji` field, including any descriptive text, instead of just emoji characters.
## Steps to Reproduce
1. Configure an agent identity with `emoji` field containi... | open | null | false | 1 | [] | [] | 2026-03-26T15:15:04Z | 2026-05-01T20:10:52Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | leedongyu1128 | 57,846,547 | MDQ6VXNlcjU3ODQ2NTQ3 | User | false |
openclaw/openclaw | 4,145,904,041 | I_kwDOQb6kR873HXmp | 55,297 | https://github.com/openclaw/openclaw/issues/55297 | https://api.github.com/repos/openclaw/openclaw/issues/55297 | Feature request: auto-update for skills and plugins | ## Problem
Skills installed via `clawhub install` and plugins installed via `openclaw plugins install` have no auto-update mechanism. Once installed, they sit at whatever version was current at install time unless the user manually runs `clawhub update` or reinstalls.
This is a friction point for plugin/skill authors... | open | null | false | 2 | [] | [] | 2026-03-26T17:22:28Z | 2026-05-01T20:10:53Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | mitchelldavis44 | 6,354,781 | MDQ6VXNlcjYzNTQ3ODE= | User | false |
openclaw/openclaw | 4,147,942,579 | I_kwDOQb6kR873PJSz | 55,412 | https://github.com/openclaw/openclaw/issues/55412 | https://api.github.com/repos/openclaw/openclaw/issues/55412 | GatewayDrainingError should auto-retry, not surface to user | ## Problem
When the gateway restarts (e.g., after `config.patch`), any in-flight agent run that triggers a new command during the drain window gets `GatewayDrainingError`. This falls through to the generic error handler in `agent-runner.runtime` and surfaces to the user as:
```
⚠️ Agent failed before reply: Gateway i... | open | null | false | 1 | [] | [] | 2026-03-26T23:17:18Z | 2026-05-04T22:44:33Z | null | NONE | null | 20260505T000034Z | 2026-05-05T00:00:34Z | assimetria-ai | 263,905,545 | U_kgDOD7rhCQ | User | false |
openclaw/openclaw | 4,148,407,195 | I_kwDOQb6kR873Q6ub | 55,450 | https://github.com/openclaw/openclaw/issues/55450 | https://api.github.com/repos/openclaw/openclaw/issues/55450 | [API] Expose underlying model name in /v1/chat/completions response | ## Problem
Currently the OpenClaw OpenAI-compatible API returns `"model": "openclaw"` in all chat completion responses, regardless of the actual underlying model being used (e.g. `anthropic/claude-sonnet-4-6`, `ollama/mistral`, etc.).
## Request
Expose the actual model identifier in the response — either:
- As the `... | open | null | false | 1 | [] | [] | 2026-03-27T01:21:24Z | 2026-05-01T20:10:55Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | k9spy | 7,834,738 | MDQ6VXNlcjc4MzQ3Mzg= | User | false |
openclaw/openclaw | 4,148,848,673 | I_kwDOQb6kR873Smgh | 55,517 | https://github.com/openclaw/openclaw/issues/55517 | https://api.github.com/repos/openclaw/openclaw/issues/55517 | [Bug]: plugin runtime subagent.run from before_dispatch does not trigger subagent_delivery_target or final external delivery | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
In OpenClaw 2026.3.24, when a plugin `before_dispatch` hook starts a child run via `api.runtime.subagent.run()` and returns an immediate handled reply, the child result is not later auto-delivered back to the s... | open | null | false | 2 | [] | [] | 2026-03-27T03:08:06Z | 2026-05-01T20:10:56Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | oky120002 | 7,135,584 | MDQ6VXNlcjcxMzU1ODQ= | User | false |
openclaw/openclaw | 4,149,274,367 | I_kwDOQb6kR873UOb_ | 55,564 | https://github.com/openclaw/openclaw/issues/55564 | https://api.github.com/repos/openclaw/openclaw/issues/55564 | Gateway replays stale queued messages after restart | ## Description
After a gateway restart (e.g. VPS reboot, `openclaw gateway restart`, or service crash), the gateway delivers all queued/pending outbound messages at once. This creates a flood of stale messages that are no longer relevant.
## Steps to Reproduce
1. Have an active conversation with the agent via Telegr... | open | null | false | 1 | [] | [] | 2026-03-27T04:42:16Z | 2026-05-01T20:10:58Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | JamesTsetsekas | 8,147,662 | MDQ6VXNlcjgxNDc2NjI= | User | false |
openclaw/openclaw | 4,151,408,124 | I_kwDOQb6kR873cXX8 | 55,734 | https://github.com/openclaw/openclaw/issues/55734 | https://api.github.com/repos/openclaw/openclaw/issues/55734 | [Bug]: OpenClaw navigates URLs into Chrome internal UI components (chrome://omnibox-popup.top-chrome/, chrome://tab-search.top-chrome/) instead of actual web tabs when using CDP mode | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When using OpenClaw to control a browser via CDP (Chrome DevTools Protocol), OpenClaw may mistakenly select Chrome's internal UI components as the target for navigation. This causes requested URLs to open insid... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-03-27T10:32:57Z | 2026-05-01T20:10:59Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | Demine4 | 31,314,996 | MDQ6VXNlcjMxMzE0OTk2 | User | false |
openclaw/openclaw | 4,152,597,004 | I_kwDOQb6kR873g5oM | 55,801 | https://github.com/openclaw/openclaw/issues/55801 | https://api.github.com/repos/openclaw/openclaw/issues/55801 | Tool policy audit logging: missing 'which rule matched' in logs | ## Problem
When a tool is allowed or denied by policy, there is no log entry indicating **which rule** was matched. This makes it impossible to audit whether security policies are working as intended.
## Current Behavior
Tool calls are logged with start/end events at DEBUG level:
```
embedded run tool start: runId=... | open | null | false | 1 | [] | [] | 2026-03-27T13:20:20Z | 2026-05-01T20:11:02Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | laterequation | 138,646,637 | U_kgDOCEOUbQ | User | false |
openclaw/openclaw | 4,156,361,679 | I_kwDOQb6kR873vQvP | 56,068 | https://github.com/openclaw/openclaw/issues/56068 | https://api.github.com/repos/openclaw/openclaw/issues/56068 | UI seamColor config setting is ignored by frontend | ## Problem
The `ui.seamColor` setting in `openclaw.json` has no visible effect on the web UI. Every accent color (buttons, sidebar highlights, selection states, focus rings) stays hardcoded red regardless of what value is configured.
## Evidence
The backend correctly:
- Validates `ui.seamColor` in the Zod/TypeBox sc... | open | null | false | 1 | [] | [] | 2026-03-27T22:56:53Z | 2026-05-01T20:11:08Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | dennis-lynch | 59,386,026 | MDQ6VXNlcjU5Mzg2MDI2 | User | false |
openclaw/openclaw | 4,160,177,856 | I_kwDOQb6kR87390bA | 56,437 | https://github.com/openclaw/openclaw/issues/56437 | https://api.github.com/repos/openclaw/openclaw/issues/56437 | [Feature Request] Native fallbackRunbook injection in agents config | ## Problem
When an agent falls back to a secondary model (e.g. OpenRouter/Qwen, Nemotron), the fallback model receives the raw cron payload message but **no system prompt, no workspace bootstrap files, and no identity context**. The result is not graceful degradation — it is a cold start with amnesia.
## What we obse... | open | null | false | 3 | [] | [] | 2026-03-28T14:27:03Z | 2026-05-01T20:11:10Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | darrenfurr | 133,894 | MDQ6VXNlcjEzMzg5NA== | User | false |
openclaw/openclaw | 4,162,140,379 | I_kwDOQb6kR874FTjb | 56,694 | https://github.com/openclaw/openclaw/issues/56694 | https://api.github.com/repos/openclaw/openclaw/issues/56694 | read tool 'without path' warning is a false positive when model uses 'file' or 'filePath' parameter names | ## Description
The embedded agent runtime emits a warning `read tool called without path` whenever the LLM calls the `read` tool using `file` or `filePath` as the parameter name instead of `path` or `file_path`. This is a **false positive** — the tool call works correctly, but the validation check only covers 2 of the... | open | null | false | 1 | [] | [] | 2026-03-29T01:21:20Z | 2026-05-01T19:07:28Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | wangmingliang-ms | 69,189,193 | MDQ6VXNlcjY5MTg5MTkz | User | false |
openclaw/openclaw | 4,174,856,750 | I_kwDOQb6kR87410Iu | 58,012 | https://github.com/openclaw/openclaw/issues/58012 | https://api.github.com/repos/openclaw/openclaw/issues/58012 | [Bug]: strict9 tool-call-id regression — Mistral via proxy providers fails again after provider-capabilities refactors | ## Bug Description
The strict9 tool-call-id fix for Mistral models routed through proxy providers (OpenRouter, Together, etc.) has **regressed**. Tool calls fail with `invalid_function_call` (code 3280) because the 9-char alphanumeric sanitization is no longer applied.
## Root Cause
The original fix (commit `4d06910... | open | null | false | 3 | [] | [] | 2026-03-31T00:30:55Z | 2026-05-01T20:11:21Z | null | CONTRIBUTOR | null | 20260501T210037Z | 2026-05-01T21:00:37Z | MoerAI | 26,067,127 | MDQ6VXNlcjI2MDY3MTI3 | User | false |
openclaw/openclaw | 4,182,434,426 | I_kwDOQb6kR875SuJ6 | 58,565 | https://github.com/openclaw/openclaw/issues/58565 | https://api.github.com/repos/openclaw/openclaw/issues/58565 | Auth profile failover should differentiate 401 (dead key) from 429/529 (transient) | ## Problem
When an auth profile's API key is invalid or expired (HTTP 401), OpenClaw puts the profile into cooldown with the same timing as rate limits (429) or overloaded responses (529). This means:
1. Every request first tries the dead key
2. Waits for the 401 failure
3. Enters cooldown (1 minute observed)
4. Fall... | open | null | false | 1 | [] | [] | 2026-03-31T21:17:54Z | 2026-05-01T20:11:23Z | null | CONTRIBUTOR | null | 20260501T210037Z | 2026-05-01T21:00:37Z | amittell | 1,388,680 | MDQ6VXNlcjEzODg2ODA= | User | false |
openclaw/openclaw | 4,184,817,687 | I_kwDOQb6kR875b0AX | 58,826 | https://github.com/openclaw/openclaw/issues/58826 | https://api.github.com/repos/openclaw/openclaw/issues/58826 | [Feature]: Add built-in token budget and request limit controls | ### Summary
Introduce built-in controls for token budgets and request limits at request/session/daily levels to prevent uncontrolled usage.
### Problem to solve
OpenClaw currently lacks built-in mechanisms for enforcing token usage limits or request caps.
While external tools such as diagnostics-otel can observe to... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-01T07:31:53Z | 2026-04-30T23:22:19Z | null | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Agrfeblve | 204,475,155 | U_kgDODDALEw | User | false |
openclaw/openclaw | 4,202,808,230 | I_kwDOQb6kR876gcOm | 60,607 | https://github.com/openclaw/openclaw/issues/60607 | https://api.github.com/repos/openclaw/openclaw/issues/60607 | [Bug] Agent completes task but sends no reply to user (dispatch complete with queuedFinal=false, replies=0) | ## Bug Description
When an agent completes a task but produces no replyable content (e.g., model returns empty response, or NO_REPLY token, or all renderable payloads are empty), the OpenClaw gateway logs `dispatch complete (queuedFinal=false, replies=0)` and the user receives **no notification at all**. This wastes t... | open | null | false | 1 | [] | [] | 2026-04-04T00:50:05Z | 2026-05-03T06:07:11Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | 3964tian | 200,762,110 | U_kgDOC_di_g | User | false |
openclaw/openclaw | 4,203,985,196 | I_kwDOQb6kR876k7ks | 60,737 | https://github.com/openclaw/openclaw/issues/60737 | https://api.github.com/repos/openclaw/openclaw/issues/60737 | [Feature]: Per-DM/per-chat default ACP binding for auto-spawning ACP sessions on new topics | [Feature]: Per-DM/per-chat default ACP binding for auto-spawning ACP sessions on new topics
## Problem
When using Telegram DM forum topics with ACP agents (Claude Code, Codex, etc.), every new topic requires manually running `/acp spawn claude` or asking the main agent to spawn an ACP session. There is no way to conf... | open | null | false | 1 | [] | [] | 2026-04-04T07:18:03Z | 2026-05-03T11:49:07Z | null | CONTRIBUTOR | null | 20260503T120038Z | 2026-05-03T12:00:38Z | Lukavyi | 1,013,690 | MDQ6VXNlcjEwMTM2OTA= | User | false |
openclaw/openclaw | 4,218,635,201 | I_kwDOQb6kR877c0PB | 62,557 | https://github.com/openclaw/openclaw/issues/62557 | https://api.github.com/repos/openclaw/openclaw/issues/62557 | `openclaw update` does not restart system-wide systemd services | ## Problem
When OpenClaw is installed npm-global and run via a **system-wide** systemd unit (`/etc/systemd/system/openclaw.service`, e.g. as a dedicated `openclaw` system user inside an LXC/VM), `openclaw update` correctly upgrades the binary but then prints:
```
Restarting service...
Gateway service disabled.
Start ... | closed | completed | true | 2 | [] | [] | 2026-04-07T15:07:22Z | 2026-05-03T04:34:10Z | 2026-05-01T02:47:17Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | robinchoice | 168,186,104 | U_kgDOCgZQ-A | User | false |
openclaw/openclaw | 4,237,305,563 | I_kwDOQb6kR878kCbb | 64,157 | https://github.com/openclaw/openclaw/issues/64157 | https://api.github.com/repos/openclaw/openclaw/issues/64157 | [Bug]: Multi-Surface Message Routing Failure — Replies lost when switching from Feishu to Web Control UI mid-conversation | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When a conversation starts on Feishu (mobile/desktop app) and the user sends a follow-up message via the OpenClaw Web Control UI, subsequent assistant replies are no longer delivered/pushed to Feishu. The Feish... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-10T07:12:41Z | 2026-05-03T12:31:06Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | aaahzb | 7,029,088 | MDQ6VXNlcjcwMjkwODg= | User | false |
openclaw/openclaw | 4,354,905,008 | I_kwDOQb6kR88AAAABA5KTsA | 74,752 | https://github.com/openclaw/openclaw/issues/74752 | https://api.github.com/repos/openclaw/openclaw/issues/74752 | TTS tool: returns (spoken) but doesn't send voice message to Telegram | ## Bug Description
**TTS tool returns "(spoken)" but doesn't send voice message to Telegram - text is sent instead**
### Environment
- OpenClaw version: 2026.4.27 (cbc2ba0)
- Node.js: v24.14.1
- Platform: Linux (Ubuntu)
- Channel: Telegram
- TTS Config:
```json
{
"auto": "always",
"provider": "microsoft... | closed | completed | true | 2 | [] | [] | 2026-04-30T02:29:46Z | 2026-05-04T04:34:33Z | 2026-05-02T02:16:43Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | Loveworld3033 | 120,953,918 | U_kgDOBzWcPg | User | false |
openclaw/openclaw | 4,354,566,155 | I_kwDOQb6kR88AAAABA41oCw | 74,698 | https://github.com/openclaw/openclaw/issues/74698 | https://api.github.com/repos/openclaw/openclaw/issues/74698 | [Bug]: Telegram group /status silently no-replies on 2026.4.27 while normal messages work | ## Bug
Telegram group slash commands such as `/status` no longer produce a response in group sessions on OpenClaw 2026.4.27, while normal group messages still route to the agent successfully.
## Environment
- OpenClaw: 2026.4.27
- OS: macOS 15.7.3 (x64)
- Node: 25.5.0
- Channel: Telegram
- Gateway: local loopback, r... | open | null | false | 1 | [] | [] | 2026-04-30T00:43:34Z | 2026-05-03T07:45:40Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | tess020126-cmyk | 259,169,068 | U_kgDOD3KbLA | User | false |
openclaw/openclaw | 4,354,621,148 | I_kwDOQb6kR88AAAABA44-3A | 74,702 | https://github.com/openclaw/openclaw/issues/74702 | https://api.github.com/repos/openclaw/openclaw/issues/74702 | openclaw-lark plugin fails to load after OpenClaw 2026.4.27: Cannot find module openclaw/plugin-sdk/zalouser | ## Bug Description
After updating OpenClaw to 2026.4.27, the `@larksuite/openclaw-lark` plugin (tried versions 2026.4.7 and 2026.4.8) fails to load with the following error:
```
Error: Cannot find module /openclaw/dist/plugin-sdk/root-alias.cjs/zalouser
- /home/goron/.openclaw/extensions/openclaw-lark/src/messaging... | closed | completed | true | 2 | [] | [] | 2026-04-30T00:58:15Z | 2026-05-02T04:26:26Z | 2026-04-30T03:14:31Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | Goron01 | 196,734,406 | U_kgDOC7ntxg | User | false |
openclaw/openclaw | 4,354,908,173 | I_kwDOQb6kR88AAAABA5KgDQ | 74,754 | https://github.com/openclaw/openclaw/issues/74754 | https://api.github.com/repos/openclaw/openclaw/issues/74754 | Auto-update detects new version but never installs it | ## Steps to reproduce
1. Set `update.auto.enabled = true`, `update.auto.stableDelayHours = 1`, `update.checkOnStart = true` in `openclaw.json`
2. Start gateway. Wait for `update-check.json` to detect a newer version.
3. Wait 2+ hours (well past the 1-hour `stableDelayHours`).
4. Run `openclaw --version` — still on old... | open | null | false | 1 | [] | [] | 2026-04-30T02:30:45Z | 2026-05-03T09:13:40Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | esigntax | 271,068,566 | U_kgDOECgtlg | User | false |
openclaw/openclaw | 4,354,600,083 | I_kwDOQb6kR88AAAABA43skw | 74,701 | https://github.com/openclaw/openclaw/issues/74701 | https://api.github.com/repos/openclaw/openclaw/issues/74701 | amazon-bedrock: `xhigh` thinking level missing for Claude Opus 4.7 (and adaptive for sonnet-4.6) | ## Summary
The `amazon-bedrock` provider extension hard-codes `resolveThinkingProfile` and **never advertises `xhigh`**, even for Claude Opus 4.7 (where the downstream stream runtime already maps xhigh → xhigh). As a result, `/think` autocomplete on channels where the session is bound to a Bedrock Opus 4.7 model (e.g.... | closed | completed | true | 3 | [] | [] | 2026-04-30T00:52:53Z | 2026-05-02T04:26:35Z | 2026-04-30T02:37:49Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | prasad-yashdeep | 45,161,513 | MDQ6VXNlcjQ1MTYxNTEz | User | false |
openclaw/openclaw | 4,354,922,108 | I_kwDOQb6kR88AAAABA5LWfA | 74,759 | https://github.com/openclaw/openclaw/issues/74759 | https://api.github.com/repos/openclaw/openclaw/issues/74759 | Secret broker plugin: first-class `secret.get`/`secret.has` over Bitwarden / Keychain / 1Password / Vault | ## Summary
Add a generic "secret broker" plugin so agents can resolve named secrets at call time without copy-pasting them into chat or env files.
## Why
OpenClaw runs as a 24/7 personal agent. Side projects (Slack bots, GitHub PR runners, fast-flights scraper, etc.) all need credentials, and the user shouldn't have... | open | null | false | 3 | [] | [] | 2026-04-30T02:34:46Z | 2026-05-03T07:45:41Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Tiger0710 | 26,865,662 | MDQ6VXNlcjI2ODY1NjYy | User | false |
openclaw/openclaw | 4,354,919,253 | I_kwDOQb6kR88AAAABA5LLVQ | 74,757 | https://github.com/openclaw/openclaw/issues/74757 | https://api.github.com/repos/openclaw/openclaw/issues/74757 | Slack tooling: `slack` tool is invisible to Claude Code, and admin actions (channel/app create, invite, member lookup) are missing | ## Summary
Two related gaps in the Slack extension surfaced while wiring up notifications for a side project (flight-watch). Filing them together because they have one root cause (Slack tooling that the agent should be able to drive end-to-end is partly missing, partly invisible).
### 1. `slack` tool is not discovera... | open | null | false | 3 | [] | [] | 2026-04-30T02:34:00Z | 2026-05-03T08:26:23Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Tiger0710 | 26,865,662 | MDQ6VXNlcjI2ODY1NjYy | User | false |
openclaw/openclaw | 4,354,790,687 | I_kwDOQb6kR88AAAABA5DVHw | 74,741 | https://github.com/openclaw/openclaw/issues/74741 | https://api.github.com/repos/openclaw/openclaw/issues/74741 | [Bug]: Signal SSE consumer times out after 10s while /api/v1/events works via curl | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
OpenClaw’s Signal channel times out and does not record inbound activity even though the configured signal-cli SSE endpoint is reachable and emits received Signal messages via curl.
### Steps to reproduce
1. ... | closed | completed | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-04-30T01:53:49Z | 2026-05-03T04:37:05Z | 2026-04-30T14:15:40Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | k7n4n5t3w4rt | 358,035 | MDQ6VXNlcjM1ODAzNQ== | User | false |
openclaw/openclaw | 4,148,059,219 | I_kwDOQb6kR873PlxT | 55,424 | https://github.com/openclaw/openclaw/issues/55424 | https://api.github.com/repos/openclaw/openclaw/issues/55424 | [Bug]: Write failed error shown but file actually saved successfully (100% success rate) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Write failed / Edit failed 错误提示频繁出现，但实际文件 100% 写入成功。这是 OpenClaw 工具层响应超时问题，不是文件系统问题。
### Steps to reproduce
1. 使用 OpenClaw 写入文件（write 或 edit 工具）
2. 观察飞书界面底部显示"Write failed"错误提示
3. 验证文件实际状态（Test-Path + Get-Item）
4. 结果显示文件已成功... | open | null | false | 1 | [
"bug",
"regression"
] | [] | 2026-03-26T23:47:27Z | 2026-05-01T21:07:32Z | null | NONE | null | 20260502T000037Z | 2026-05-02T00:00:37Z | hnfjf4289-sketch | 267,009,132 | U_kgDOD-o8bA | User | false |
openclaw/openclaw | 4,354,625,879 | I_kwDOQb6kR88AAAABA45RVw | 74,704 | https://github.com/openclaw/openclaw/issues/74704 | https://api.github.com/repos/openclaw/openclaw/issues/74704 | SDK: stabilize app-client happy path for agents, sessions, runs | ## Goal
Stabilize the current `@openclaw/sdk` happy path for real external app clients.
OpenMeow is a useful dogfood client because it exercises the core app-SDK flow without requiring future managed/cloud features.
## Happy path to validate
1. Connect to Gateway.
2. Discover agents/models.
3. Create or resume a se... | open | null | false | 8 | [
"maintainer"
] | [
"BunsDev",
"Copilot"
] | 2026-04-30T00:59:27Z | 2026-05-02T13:41:21Z | null | MEMBER | null | 20260502T150030Z | 2026-05-02T15:00:30Z | BunsDev | 68,980,965 | MDQ6VXNlcjY4OTgwOTY1 | User | false |
openclaw/openclaw | 4,354,924,222 | I_kwDOQb6kR88AAAABA5Levg | 74,760 | https://github.com/openclaw/openclaw/issues/74760 | https://api.github.com/repos/openclaw/openclaw/issues/74760 | [Bug]: The same problem: mentioning the bot in a group chat does not get a reply, but it works in P2P. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
gateway log：
10:28:21 [moltbot-popo] [POPO] Dispatching to agent (session=agent:main:moltbot-popo:group:8314302)
10:28:21 [moltbot-popo] [POPO] prepareStreaming: suppressor registered for 8314302
10:28:21 [moltbot-popo] [POP... | closed | not_planned | true | 2 | [
"bug"
] | [] | 2026-04-30T02:35:19Z | 2026-05-03T04:35:57Z | 2026-04-30T16:42:52Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | zym951223 | 51,111,325 | MDQ6VXNlcjUxMTExMzI1 | User | false |
openclaw/openclaw | 4,354,921,080 | I_kwDOQb6kR88AAAABA5LSeA | 74,758 | https://github.com/openclaw/openclaw/issues/74758 | https://api.github.com/repos/openclaw/openclaw/issues/74758 | macOS host integration: `launchd` plugin for installing/managing LaunchAgents from agent code | ## Summary
Add a small host-integration plugin (or extension) that lets agents install/load/unload macOS launchd LaunchAgents on the host running OpenClaw.
## Why
Long-running side projects on this Mac mini (e.g. a flight-price watcher, a heartbeat job) want to drop a `~/Library/LaunchAgents/com.user.foo.plist` and ... | closed | not_planned | true | 1 | [] | [] | 2026-04-30T02:34:29Z | 2026-05-02T04:26:32Z | 2026-04-30T02:45:29Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | Tiger0710 | 26,865,662 | MDQ6VXNlcjI2ODY1NjYy | User | false |
openclaw/openclaw | 4,151,409,817 | I_kwDOQb6kR873cXyZ | 55,735 | https://github.com/openclaw/openclaw/issues/55735 | https://api.github.com/repos/openclaw/openclaw/issues/55735 | [Suggestion] The agent has no version-aware self-knowledge — it misdiagnoses intentional new behaviors as bugs after every upgrade | ## Summary
OpenClaw ships as an AI agent that users naturally turn to when something feels "wrong" after an upgrade. But the agent has **no awareness of its own version history or design changes**. This means it consistently misdiagnoses intentional new behaviors as bugs — and users end up chasing phantom problems.
... | open | null | false | 1 | [] | [] | 2026-03-27T10:33:11Z | 2026-05-01T20:11:01Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | derekcscn | 269,405,863 | U_kgDOEA7Opw | User | false |
openclaw/openclaw | 4,354,958,341 | I_kwDOQb6kR88AAAABA5NkBQ | 74,763 | https://github.com/openclaw/openclaw/issues/74763 | https://api.github.com/repos/openclaw/openclaw/issues/74763 | sessions_spawn(model=...) override silently dropped — written to runtime fields instead of override fields | ## Summary
Subagent runs spawned via `sessions_spawn(model=...)` (or the gateway `subagent.spawn` method's `model` param) silently ignore the requested model and run on the agent's resolved primary instead. The spawn response returns `modelApplied: true`, which is misleading.
## Root Cause
In `dist/subagent-spawn--Q... | closed | completed | true | 1 | [] | [] | 2026-04-30T02:44:53Z | 2026-05-02T04:26:30Z | 2026-04-30T02:49:57Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | dsgjarvis | 262,263,496 | U_kgDOD6HSyA | User | false |
openclaw/openclaw | 4,354,441,000 | I_kwDOQb6kR88AAAABA4t_KA | 74,692 | https://github.com/openclaw/openclaw/issues/74692 | https://api.github.com/repos/openclaw/openclaw/issues/74692 | [Bug]: 2026.4.27 builtin memory can fail because plugin-runtime-deps snapshot misses sqlite-vec | ## Summary
After upgrading to `OpenClaw 2026.4.27`, builtin memory/vector search can fail because the plugin runtime dependency snapshot under `~/.openclaw/plugin-runtime-deps/openclaw-2026.4.27-*/node_modules` does not contain `sqlite-vec`, even though the main global OpenClaw install does.
This breaks memory startup... | closed | completed | true | 5 | [] | [] | 2026-04-30T00:15:28Z | 2026-05-02T04:26:29Z | 2026-04-30T02:52:06Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | mozi1924 | 70,095,417 | MDQ6VXNlcjcwMDk1NDE3 | User | false |
openclaw/openclaw | 4,354,976,854 | I_kwDOQb6kR88AAAABA5OsVg | 74,766 | https://github.com/openclaw/openclaw/issues/74766 | https://api.github.com/repos/openclaw/openclaw/issues/74766 | [Bug]: WhatsApp group agents cannot create polls because same-chat runtime hides message tool | ## Summary
In WhatsApp group auto-reply sessions, the agent no longer receives the `message` tool even when the agent config explicitly allows it. Because native polls are exposed via `message` action `poll`, the agent cannot create WhatsApp polls and falls back to a text-only "reply with numbers" pseudo-poll.
This a... | closed | completed | true | 1 | [] | [] | 2026-04-30T02:50:05Z | 2026-05-02T04:26:28Z | 2026-04-30T02:56:20Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | dyoung522 | 422,900 | MDQ6VXNlcjQyMjkwMA== | User | false |
openclaw/openclaw | 4,354,984,206 | I_kwDOQb6kR88AAAABA5PJDg | 74,767 | https://github.com/openclaw/openclaw/issues/74767 | https://api.github.com/repos/openclaw/openclaw/issues/74767 | [Bug]: Feishu streaming card messages not searchable — interactive card content not indexed by Feishu search | ## Summary
Messages delivered via Feishu interactive card streaming mode are **not searchable** in Feishu's built-in conversation search ("搜索会话内容"). The same content delivered as a plain text message (streaming disabled) is searchable.
## Steps to Reproduce
1. Enable Feishu streaming card mode (`channels.feishu.stre... | open | null | false | 1 | [] | [] | 2026-04-30T02:52:13Z | 2026-05-03T11:50:25Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | TwinsLee | 1,460,694 | MDQ6VXNlcjE0NjA2OTQ= | User | false |
openclaw/openclaw | 4,354,975,643 | I_kwDOQb6kR88AAAABA5Onmw | 74,765 | https://github.com/openclaw/openclaw/issues/74765 | https://api.github.com/repos/openclaw/openclaw/issues/74765 | [Bug]: v2026.4.27 regression — channels.discord.accounts.*.token: unresolved SecretRef "env:default:..." during openclaw gateway restart | **Title:**
[Bug]: v2026.4.27 regression — `channels.discord.accounts.*.token: unresolved SecretRef "env:default:..."` during `openclaw gateway restart`
**Description:**
### Steps to Reproduce
1. Upgrade to **v2026.4.27** (commit `cbc2ba0`).
2. Run `openclaw gateway restart` (or trigger a systemd service restart aft... | closed | completed | true | 3 | [] | [] | 2026-04-30T02:49:44Z | 2026-05-02T04:26:19Z | 2026-04-30T04:10:40Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | PrinceOfEgypt | 2,649,235 | MDQ6VXNlcjI2NDkyMzU= | User | false |
openclaw/openclaw | 4,355,047,965 | I_kwDOQb6kR88AAAABA5TCHQ | 74,774 | https://github.com/openclaw/openclaw/issues/74774 | https://api.github.com/repos/openclaw/openclaw/issues/74774 | Gateway leaks MCP child process trees on session end (mcp-server-filesystem, mcp-server-github) — ~70 MB RSS each, accumulates 50+/day | ## Summary
The gateway spawns `mcp-server-{filesystem,github}` (and similar npm-distributed MCP servers) as child process trees per agent session. **When an agent session ends, the child process tree is never reaped.** Each leaked tree is `npm exec → sh -c → node mcp-server-*` and holds ~70 MB RSS. Over hours of norma... | closed | completed | true | 1 | [] | [] | 2026-04-30T03:10:15Z | 2026-05-02T04:26:23Z | 2026-04-30T03:15:31Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | mlopez14 | 5,874,667 | MDQ6VXNlcjU4NzQ2Njc= | User | false |
openclaw/openclaw | 4,355,063,600 | I_kwDOQb6kR88AAAABA5T_MA | 74,777 | https://github.com/openclaw/openclaw/issues/74777 | https://api.github.com/repos/openclaw/openclaw/issues/74777 | Local memorySearch runtime deps are not retained/resolved in plugin-runtime-deps | ## Summary
Local `memorySearch` fails because OpenClaw's managed `plugin-runtime-deps` environment does not retain or resolve packages required by the local memory runtime.
Observed missing packages:
- `node-llama-cpp`
- `sqlite-vec`
- `chokidar`
The main failure is:
```text
Local embeddings unavailable.
Reason: o... | closed | completed | true | 2 | [] | [] | 2026-04-30T03:14:51Z | 2026-05-03T04:37:03Z | 2026-04-30T14:23:21Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | LLagoon3 | 115,124,830 | U_kgDOBtyqXg | User | false |
openclaw/openclaw | 4,355,075,599 | I_kwDOQb6kR88AAAABA5UuDw | 74,782 | https://github.com/openclaw/openclaw/issues/74782 | https://api.github.com/repos/openclaw/openclaw/issues/74782 | Feature: `openclaw logs --follow` should auto-reconnect instead of exiting on transient gateway disconnect | ## Problem
When running `openclaw logs --follow`, any transient gateway WebSocket disconnect (e.g., gateway restart, temporary connection drop) causes the CLI to immediately `process.exit(1)`, terminating the log tail session.
This is particularly annoying during gateway restarts or when brief connectivity hiccups oc... | closed | completed | false | 1 | [] | [] | 2026-04-30T03:19:03Z | 2026-05-03T09:18:51Z | 2026-05-03T09:18:51Z | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | subaochen | 1,390,258 | MDQ6VXNlcjEzOTAyNTg= | User | false |
openclaw/openclaw | 4,355,101,928 | I_kwDOQb6kR88AAAABA5WU6A | 74,788 | https://github.com/openclaw/openclaw/issues/74788 | https://api.github.com/repos/openclaw/openclaw/issues/74788 | [Bug]: OpenRouter plugin missing supportsXHighThinking — xhigh rejected for DeepSeek V4 models | ### Summary
The OpenRouter provider plugin does not implement `supportsXHighThinking`, so the `/think xhigh` directive is rejected for `openrouter/deepseek/deepseek-v4-pro` and `openrouter/deepseek/deepseek-v4-flash`. The underlying API fully supports it — `reasoning.effort: "xhigh"` maps to DeepSeek max reasoning on ... | closed | completed | false | 1 | [] | [
"vincentkoc"
] | 2026-04-30T03:25:25Z | 2026-05-03T17:53:19Z | 2026-05-03T17:53:12Z | NONE | null | 20260503T180029Z | 2026-05-03T18:00:29Z | razhangwei | 4,484,566 | MDQ6VXNlcjQ0ODQ1NjY= | User | false |
openclaw/openclaw | 4,355,124,909 | I_kwDOQb6kR88AAAABA5XurQ | 74,798 | https://github.com/openclaw/openclaw/issues/74798 | https://api.github.com/repos/openclaw/openclaw/issues/74798 | MCP进程泄漏：gateway session结束时未清理vibe-trading-mcp子进程 | ## 问题描述
每次通过飞书等channel调用Vibe-Trading MCP工具时，openclaw-gateway会spawn一个 \`vibe-trading-mcp\` 子进程。但当session结束时，这些子进程未被回收，持续累积直至变成僵尸进程。
**现象**：宿主机上 \`ps -ef|grep mcp\` 可以看到大量残留进程，父进程均为 \`openclaw-gateway\`，创建时间从某天开始每天新增一个。
## 根本原因
问题出在 OpenClaw Gateway 的 MCP 生命周期管理：
1. **泄漏链路**：gateway 路由的 session 通过 \`getOrCreateSessi... | closed | completed | true | 2 | [] | [] | 2026-04-30T03:31:08Z | 2026-05-03T04:36:15Z | 2026-04-30T15:31:48Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | sawchen | 27,667,422 | MDQ6VXNlcjI3NjY3NDIy | User | false |
openclaw/openclaw | 4,355,122,984 | I_kwDOQb6kR88AAAABA5XnKA | 74,796 | https://github.com/openclaw/openclaw/issues/74796 | https://api.github.com/repos/openclaw/openclaw/issues/74796 | [bug]: openai-codex xhigh thinking level is downgraded in gateway because provider thinking profile is not visible at startup | ### Summary
In OpenClaw `2026.4.27`, `openai-codex/gpt-5.5` appears to support `xhigh` at the provider level, but gateway sessions reject or downgrade `xhigh` to `high`. The shared thinking resolver behaves as if the `openai-codex` provider profile is not present in the active gateway plugin registry.
This affects no... | closed | completed | true | 4 | [] | [] | 2026-04-30T03:30:34Z | 2026-05-04T04:36:06Z | 2026-05-01T14:51:42Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | maxschachere | 7,988,602 | MDQ6VXNlcjc5ODg2MDI= | User | false |
openclaw/openclaw | 4,355,136,227 | I_kwDOQb6kR88AAAABA5Ya4w | 74,800 | https://github.com/openclaw/openclaw/issues/74800 | https://api.github.com/repos/openclaw/openclaw/issues/74800 | Telegram plugin commands can fail with 'Something went wrong' when command handler result is undefined in v2026.4.27 | ## Summary
After upgrading to `v2026.4.27`, Telegram custom/plugin commands such as `/bf` and `/asset` can surface:
> Something went wrong while processing your request. Please try again.
In my case these are local routing/plugin commands (non-AI scripts). The underlying script can still run, but Telegram returns the... | closed | completed | false | 1 | [] | [
"vincentkoc"
] | 2026-04-30T03:34:20Z | 2026-05-03T17:59:31Z | 2026-05-03T17:59:25Z | NONE | null | 20260503T180029Z | 2026-05-03T18:00:29Z | githubdulong | 59,136,159 | MDQ6VXNlcjU5MTM2MTU5 | User | false |
openclaw/openclaw | 4,355,140,746 | I_kwDOQb6kR88AAAABA5Ysig | 74,801 | https://github.com/openclaw/openclaw/issues/74801 | https://api.github.com/repos/openclaw/openclaw/issues/74801 | Log rotation cron job checks wrong file — journald is the real storage pressure | ## Bug
The built-in log rotation cron job (checks `~/.openclaw/logs/openclaw.log` for 50 MB) is watching the wrong file.
When OpenClaw runs as a systemd user service, `openclaw.log` stays near-empty because the file transport is never initialized. All actual log output goes to **journald**, which has no size cap and ... | open | null | false | 1 | [] | [] | 2026-04-30T03:35:30Z | 2026-05-03T07:05:43Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | jeffsteinbok-openclaw | 264,727,320 | U_kgDOD8drGA | User | false |
openclaw/openclaw | 4,355,141,873 | I_kwDOQb6kR88AAAABA5Yw8Q | 74,803 | https://github.com/openclaw/openclaw/issues/74803 | https://api.github.com/repos/openclaw/openclaw/issues/74803 | Isolated cron agentTurn jobs without lightContext/toolsAllow timeout on cold start before doing any work | ## Bug
Isolated `agentTurn` cron jobs that don't set `lightContext: true` and don't restrict `toolsAllow` spend their entire timeout budget on session bootstrap (loading full agent context, memory, workspace files) and timeout before executing any actual work.
## Steps to reproduce
1. Create a cron job with `session... | open | null | false | 1 | [] | [] | 2026-04-30T03:35:45Z | 2026-05-03T08:26:25Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | jeffsteinbok-openclaw | 264,727,320 | U_kgDOD8drGA | User | false |
openclaw/openclaw | 4,355,160,365 | I_kwDOQb6kR88AAAABA5Z5LQ | 74,806 | https://github.com/openclaw/openclaw/issues/74806 | https://api.github.com/repos/openclaw/openclaw/issues/74806 | Bug: /new command in Feishu reports 'messages must not be empty' error | \n## Bug Description\n\nWhen sending `/new` in a Feishu private chat, OpenClaw reports the error:\n\n> ⚠️ Something went wrong while processing your request. Please try again, or use /new to start a fresh session.\n\nThe server-side error logged is:\n\n```\nLLM request rejected: invalid params, messages must not be emp... | closed | completed | true | 1 | [] | [] | 2026-04-30T03:41:33Z | 2026-05-02T04:26:22Z | 2026-04-30T03:46:55Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | zuoxiaojun | 24,306,381 | MDQ6VXNlcjI0MzA2Mzgx | User | false |
openclaw/openclaw | 4,355,165,640 | I_kwDOQb6kR88AAAABA5aNyA | 74,807 | https://github.com/openclaw/openclaw/issues/74807 | https://api.github.com/repos/openclaw/openclaw/issues/74807 | [Bug] Cron task summary ignores agent self-correction - picks tool error over final assistant reply | ## Problem
When a Cron `agentTurn` job runs in an isolated session, the final summary/announcement picks the **last tool error message** as the task result instead of the agent's final self-corrected assistant reply.
### Observed behaviour
1. Agent runs a task that involves file editing via tool calls.
2. First `edi... | open | null | false | 1 | [] | [] | 2026-04-30T03:43:20Z | 2026-05-03T13:22:16Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | u5008907541-jpg | 238,644,415 | U_kgDODjlsvw | User | false |
openclaw/openclaw | 4,355,073,324 | I_kwDOQb6kR88AAAABA5UlLA | 74,780 | https://github.com/openclaw/openclaw/issues/74780 | https://api.github.com/repos/openclaw/openclaw/issues/74780 | tools/invoke returns "Tool not available: message" in 2026.4.26 even with healthy telegram channel | ## OpenClaw version
2026.4.26 (be8c246) on macOS 25.4.0 (Darwin / Apple Silicon), Node 25.9.0.
## What I expected
`POST /tools/invoke` with `{"tool":"message","args":{"action":"send","channel":"telegram","target":"<chat-id>","message":"…"}}` should deliver a Telegram message via the configured bot, the same way agent ... | closed | completed | true | 2 | [] | [] | 2026-04-30T03:18:09Z | 2026-05-02T04:26:22Z | 2026-04-30T03:47:16Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | rjdjohnston | 293,686 | MDQ6VXNlcjI5MzY4Ng== | User | false |
openclaw/openclaw | 4,355,183,709 | I_kwDOQb6kR88AAAABA5bUXQ | 74,810 | https://github.com/openclaw/openclaw/issues/74810 | https://api.github.com/repos/openclaw/openclaw/issues/74810 | LLM-only modelRun fails when inherited tool allowlist is non-empty | ### Bug type
Tool policy / model-run bug
### Version
OpenClaw 2026.4.27 (cbc2ba0)
### Summary
LLM-only runs that intentionally disable tools can fail before reaching the provider when the runtime inherits a global tool allowlist. The error says no callable tools remain, even though tools are intentionally disabled fo... | closed | completed | false | 3 | [] | [] | 2026-04-30T03:48:19Z | 2026-05-03T13:49:11Z | 2026-05-03T13:49:11Z | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | abnershang | 75,654,486 | MDQ6VXNlcjc1NjU0NDg2 | User | false |
openclaw/openclaw | 4,355,182,995 | I_kwDOQb6kR88AAAABA5bRkw | 74,809 | https://github.com/openclaw/openclaw/issues/74809 | https://api.github.com/repos/openclaw/openclaw/issues/74809 | Top-level network proxy breaks Discord TLS when using local HTTP proxy | ### Bug type
Network proxy / transport bug
### Version
OpenClaw 2026.4.27 (cbc2ba0)
### Summary
Enabling the documented top-level network proxy with a local HTTP forward proxy caused Discord gateway TLS failures. The same local proxy works with raw `curl` and Undici `ProxyAgent`, so the failure appears specific to Op... | closed | completed | false | 2 | [] | [
"jesse-merhi"
] | 2026-04-30T03:48:07Z | 2026-05-03T08:43:25Z | 2026-05-03T08:43:25Z | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | abnershang | 75,654,486 | MDQ6VXNlcjc1NjU0NDg2 | User | false |
openclaw/openclaw | 4,355,226,727 | I_kwDOQb6kR88AAAABA5d8Zw | 74,817 | https://github.com/openclaw/openclaw/issues/74817 | https://api.github.com/repos/openclaw/openclaw/issues/74817 | Feishu: @所有人 not delivered to bot + requireMention:false causes multi-bot echo loop | # Feishu (飞书) Bot: `@所有人` Not Received + Multi-Bot Echo Loop
## Issue Summary
Two related problems with Feishu group chat bot configuration in OpenClaw:
---
## Problem 1: `@所有人` ( `@everyone` ) Not Delivered to Bot
**Severity:** Medium
**Description:**
When a user mentions `@所有人` in a Feishu group, the bot does n... | closed | not_planned | true | 2 | [] | [] | 2026-04-30T04:02:04Z | 2026-05-03T04:34:40Z | 2026-04-30T23:29:42Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | shao-zhijie | 265,209,338 | U_kgDOD87F-g | User | false |
openclaw/openclaw | 4,355,254,319 | I_kwDOQb6kR88AAAABA5foLw | 74,822 | https://github.com/openclaw/openclaw/issues/74822 | https://api.github.com/repos/openclaw/openclaw/issues/74822 | Gateway: recurring '⚠️ Something went wrong' error requires manual restart (Telegram) | ## Summary
Recurring gateway error in a Telegram session that makes the assistant unresponsive. Requires manual `openclaw gateway restart` each time to recover.
## Error message shown to user
> ⚠️ Something went wrong while processing your request. Please try again, or use /new to start a fresh session.
## Reproduc... | open | null | false | 2 | [] | [] | 2026-04-30T04:10:53Z | 2026-05-04T12:44:11Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | braddelong | 27,788,138 | MDQ6VXNlcjI3Nzg4MTM4 | User | false |
openclaw/openclaw | 4,355,268,627 | I_kwDOQb6kR88AAAABA5ggEw | 74,832 | https://github.com/openclaw/openclaw/issues/74832 | https://api.github.com/repos/openclaw/openclaw/issues/74832 | openclaw status fails on Telegram file SecretRef while resolving allowlist metadata | ## Summary
On `2026.4.27`, `openclaw status` can hard-fail when the Telegram bot token is configured via a file-backed SecretRef, even though the gateway is live and `openclaw secrets audit` reports the SecretRef as resolved.
Observed error:
```text
channels.telegram.botToken: unresolved SecretRef "file:filemain:/te... | open | null | false | 1 | [] | [] | 2026-04-30T04:15:49Z | 2026-05-03T08:40:19Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | coolmanns | 117,685,070 | U_kgDOBwO7Tg | User | false |
openclaw/openclaw | 4,355,256,984 | I_kwDOQb6kR88AAAABA5fymA | 74,823 | https://github.com/openclaw/openclaw/issues/74823 | https://api.github.com/repos/openclaw/openclaw/issues/74823 | [Bug]: sqlite-vec extension fails to load silently despite vec0.so being present and functional | ## Description
The `memory-core` plugin's vector store reports `sqlite-vec unavailable` on every memory chunk write, even though the `sqlite-vec` npm package (v0.1.9) is installed and the native `vec0.so` extension loads successfully when called manually inside the container.
This causes two problems:
1. **Silent per... | closed | completed | true | 1 | [] | [] | 2026-04-30T04:11:50Z | 2026-05-02T04:26:17Z | 2026-04-30T04:17:29Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | ranklancer | 43,620,530 | MDQ6VXNlcjQzNjIwNTMw | User | false |
openclaw/openclaw | 4,355,284,092 | I_kwDOQb6kR88AAAABA5hcfA | 74,837 | https://github.com/openclaw/openclaw/issues/74837 | https://api.github.com/repos/openclaw/openclaw/issues/74837 | Bug: sessions_spawn model parameter silently ignored — sub-agents always use orchestrator default | ## Bug Report
### Summary
The `model` parameter on `sessions_spawn()` is silently ignored. The response returns `modelApplied: true`, but the sub-agent runtime always uses the orchestrator's default model instead of the requested model.
### Environment
- **OpenClaw version:** 2026.4.27 (cbc2ba0)
- **Platform:** macOS... | open | null | false | 2 | [] | [] | 2026-04-30T04:20:44Z | 2026-05-03T04:44:20Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | danen-carlson | 4,970,765 | MDQ6VXNlcjQ5NzA3NjU= | User | false |
openclaw/openclaw | 4,355,273,090 | I_kwDOQb6kR88AAAABA5gxgg | 74,835 | https://github.com/openclaw/openclaw/issues/74835 | https://api.github.com/repos/openclaw/openclaw/issues/74835 | Add global provider request proxy/default SSRF policy for model providers | ### Feature request
Add a global/default provider request policy for model providers, so operators can configure an explicit proxy and fake-IP/private-network allowance once instead of repeating it under every `models.providers.<id>.request` block.
### Version / context
OpenClaw 2026.4.27 (cbc2ba0)
### Problem
The do... | open | null | false | 1 | [] | [] | 2026-04-30T04:17:28Z | 2026-05-03T01:23:17Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | abnershang | 75,654,486 | MDQ6VXNlcjc1NjU0NDg2 | User | false |
openclaw/openclaw | 4,146,959,616 | I_kwDOQb6kR873LZUA | 55,349 | https://github.com/openclaw/openclaw/issues/55349 | https://api.github.com/repos/openclaw/openclaw/issues/55349 | Control UI: Threaded reply / quote-reply for chat messages | ## Feature Request
### Summary
Add the ability to reply to a specific message in the Control UI chat, similar to how iMessage and WhatsApp handle threaded replies.
### Current Behavior
The chat is purely linear — there is no way to reference or reply to a specific earlier message.
### Desired Behavior
- Long-press o... | closed | not_planned | true | 1 | [] | [] | 2026-03-26T19:57:00Z | 2026-05-04T04:35:30Z | 2026-05-01T21:08:01Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | armando226 | 51,886,777 | MDQ6VXNlcjUxODg2Nzc3 | User | false |
openclaw/openclaw | 4,147,285,962 | I_kwDOQb6kR873Mo_K | 55,366 | https://github.com/openclaw/openclaw/issues/55366 | https://api.github.com/repos/openclaw/openclaw/issues/55366 | [Bug]: Hard context reset destroys in-progress task state with no recovery path | ## Summary
When an agent hits the context limit and compaction cannot recover enough space, OpenClaw fires a hard session reset with this message:
> ⚠️ Context limit exceeded. I've reset our conversation to start fresh - please try again.
The reset **wipes all in-progress task state**, leaving the agent with no mem... | open | null | false | 1 | [] | [] | 2026-03-26T20:58:00Z | 2026-05-01T21:07:30Z | null | NONE | null | 20260502T000037Z | 2026-05-02T00:00:37Z | shiftymix | 193,249,208 | U_kgDOC4S_uA | User | false |
openclaw/openclaw | 4,148,814,114 | I_kwDOQb6kR873SeEi | 55,512 | https://github.com/openclaw/openclaw/issues/55512 | https://api.github.com/repos/openclaw/openclaw/issues/55512 | Markdown table columns misaligned with CJK characters on Telegram and Discord | ## Description
Markdown tables containing CJK (Chinese/Japanese/Korean) characters are rendered with misaligned columns on both Telegram and Discord channels.
## Problem
- **Telegram**: Tables are wrapped in `<pre>` blocks, but column padding uses `String.length` instead of display width. CJK characters occupy **2 t... | open | null | false | 2 | [] | [] | 2026-03-27T03:00:10Z | 2026-05-01T21:07:33Z | null | NONE | null | 20260502T000037Z | 2026-05-02T00:00:37Z | zjy282 | 8,154,648 | MDQ6VXNlcjgxNTQ2NDg= | User | false |
openclaw/openclaw | 4,150,800,588 | I_kwDOQb6kR873aDDM | 55,686 | https://github.com/openclaw/openclaw/issues/55686 | https://api.github.com/repos/openclaw/openclaw/issues/55686 | [Feature Request] Make command options clickable buttons in Feishu (e.g., /reasoning) | ## Description
Currently, when users invoke certain commands like `/reasoning` in Feishu, the options are displayed as plain text (e.g., \"Options: on, off, stream\"), requiring users to manually type their response.
## Expected Behavior
Options should be rendered as interactive Feishu card buttons that users can cl... | open | null | false | 1 | [] | [] | 2026-03-27T09:01:15Z | 2026-05-01T21:07:35Z | null | NONE | null | 20260502T000037Z | 2026-05-02T00:00:37Z | xb22133 | 66,028,269 | MDQ6VXNlcjY2MDI4MjY5 | User | false |
openclaw/openclaw | 4,150,882,168 | I_kwDOQb6kR873aW94 | 55,694 | https://github.com/openclaw/openclaw/issues/55694 | https://api.github.com/repos/openclaw/openclaw/issues/55694 | Agent陷入工具调用失败死循环，导致重复发送消息刷屏 | ## 问题描述
当工具调用失败时，Agent 会陷入无限重试循环，每次重试前都发送消息，导致用户收到大量重复消息刷屏。
## 复现场景
在飞书对话中，Agent 尝试使用 `exec` 工具执行命令，但参数错误（缺少必需的 `command` 参数）。随后 Agent 连续重试 20+ 次，每次重试前都发送类似消息：
> "让我保存新 token 并提交 issue"
结果用户在话题中收到 6+ 条几乎相同的消息。
## 日志片段
```
08:38:57 - exec 调用失败: Validation failed - command must have required property
08:39:03 - 再次... | open | null | false | 4 | [] | [] | 2026-03-27T09:13:57Z | 2026-05-01T21:07:36Z | null | NONE | null | 20260502T000037Z | 2026-05-02T00:00:37Z | 51Google | 17,562,570 | MDQ6VXNlcjE3NTYyNTcw | User | false |
openclaw/openclaw | 4,151,075,919 | I_kwDOQb6kR873bGRP | 55,705 | https://github.com/openclaw/openclaw/issues/55705 | https://api.github.com/repos/openclaw/openclaw/issues/55705 | [Bug]: 控制面板 Subagent 数量显示异常 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
**版本：** OpenClaw 2026.3.24（最新）
**系统：** macOS
**复现步骤：**
1. 打开控制面板（127.0.0.1:18789）
2. 查看 Subagent 列表，显示 14 个活跃 subagent
3. 但实际通过 API/CLI 查询 subagent 列表为空
**预期行为：** 控制面板显示的数量应与实际一致
**附加信息：** 控制面板 UI 显示的版本号为 v2024.3.24（旧版本号格式... | open | null | false | 1 | [
"bug",
"regression"
] | [] | 2026-03-27T09:43:19Z | 2026-05-01T21:07:37Z | null | NONE | null | 20260502T000037Z | 2026-05-02T00:00:37Z | KoraYip-07 | 270,113,321 | U_kgDOEBmaKQ | User | false |
openclaw/openclaw | 4,153,203,848 | I_kwDOQb6kR873jNyI | 55,836 | https://github.com/openclaw/openclaw/issues/55836 | https://api.github.com/repos/openclaw/openclaw/issues/55836 | [Feature]: TUI: vi/vim keybinding mode for input editor | ## Feature Request
### Summary
Add a vi/vim keybinding mode to the TUI input editor, similar to `set -o vi` in bash/readline.
### Motivation
Users who prefer modal text editing (vi/vim style) currently have no way to enable it in the TUI input box. Shell-level `set -o vi` affects the shell prompt but not the TUI's ow... | open | null | false | 1 | [] | [] | 2026-03-27T14:28:44Z | 2026-05-01T21:07:39Z | null | NONE | null | 20260502T000037Z | 2026-05-02T00:00:37Z | tomwhipple | 635,272 | MDQ6VXNlcjYzNTI3Mg== | User | false |
openclaw/openclaw | 4,154,091,958 | I_kwDOQb6kR873mmm2 | 55,897 | https://github.com/openclaw/openclaw/issues/55897 | https://api.github.com/repos/openclaw/openclaw/issues/55897 | api_key auth: MRU sticky ordering + silent 402 failover | ## Problem
When using multiple API keys for the same provider (e.g., 3 keys with separate billing quotas), two issues arise:
### 1. Key selection after /new (session reset)
`orderProfilesByMode` sorts by `lastUsed` ascending (LRU), so after `/new`, the least-recently-used key is selected first. If that key has no quo... | open | null | false | 1 | [] | [] | 2026-03-27T16:17:32Z | 2026-05-01T20:11:04Z | null | NONE | null | 20260501T210037Z | 2026-05-01T21:00:37Z | qmpnqs8nmq-create | 267,704,485 | U_kgDOD_TYpQ | User | false |
openclaw/openclaw | 4,157,579,502 | I_kwDOQb6kR873z6Du | 56,174 | https://github.com/openclaw/openclaw/issues/56174 | https://api.github.com/repos/openclaw/openclaw/issues/56174 | [Feature]: Availablity On Termux | ### Summary
Add Openclaw to Termux. termux is something like a linux terminal in your android.
### Problem to solve
Termux: it is something like a linux terminal in your android. So i have been using this for a long time. And i have tried a lot of opensources. But for openclaw. I tried many times. But it is failin... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-28T04:49:33Z | 2026-05-03T04:38:12Z | 2026-04-30T06:43:18Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | amanmohdtp | 234,303,912 | U_kgDODfcxqA | User | false |
openclaw/openclaw | 4,209,914,917 | I_kwDOQb6kR8767jQl | 61,709 | https://github.com/openclaw/openclaw/issues/61709 | https://api.github.com/repos/openclaw/openclaw/issues/61709 | [Feature]: openclaw auth migrate — atomic auth profile switch across all agent stores | ## Problem
When switching a provider's auth mode (e.g., Anthropic OAuth token → API key), each agent has its own `auth-profiles.json` under `~/.openclaw/agents/<id>/agent/auth-profiles.json`. There is no CLI command to update all agent stores atomically. The result: partial migrations where some agents are updated and... | open | null | false | 1 | [] | [] | 2026-04-06T06:03:46Z | 2026-05-03T13:21:04Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | ScientificProgrammer | 4,348,294 | MDQ6VXNlcjQzNDgyOTQ= | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.