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 | 4,319,115,872 | I_kwDOQb6kR88AAAABAXB6YA | 70,795 | https://github.com/openclaw/openclaw/issues/70795 | https://api.github.com/repos/openclaw/openclaw/issues/70795 | `read` tool remains hard-bound to workspace root despite broader FS policy / runtime patches | `read` tool remains hard-bound to workspace root despite broader FS policy / runtime patches
## Summary
The `read` tool in this session still rejects paths outside `~/.openclaw/workspace` with errors like:
- `Path escapes sandbox root (~/.openclaw/workspace): /opt/homebrew/...`
- `Path escapes sandbox root (~/.openc... | open | null | false | 2 | [] | [] | 2026-04-23T21:55:08Z | 2026-05-03T02:15:10Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | cowtipper2000-cmyk | 264,339,154 | U_kgDOD8F-0g | User | false |
openclaw/openclaw | 4,319,143,577 | I_kwDOQb6kR88AAAABAXDmmQ | 70,797 | https://github.com/openclaw/openclaw/issues/70797 | https://api.github.com/repos/openclaw/openclaw/issues/70797 | Tool-call payload can leak into user-visible reply during background exec/process flows | ## Summary
In some sessions, internal tool orchestration payloads can be rendered into the user-visible reply instead of being handled entirely inside the runtime.
Example of leaked content seen by the user:
```text
Aku install sshpass dulu sekarang.
{"tool_calls":[{"function":{"arguments":"{\"action\":\"poll\",\"se... | open | null | false | 3 | [] | [] | 2026-04-23T22:00:59Z | 2026-05-03T10:41:27Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | erul2 | 90,251,663 | MDQ6VXNlcjkwMjUxNjYz | User | false |
openclaw/openclaw | 4,319,231,374 | I_kwDOQb6kR88AAAABAXI9jg | 70,808 | https://github.com/openclaw/openclaw/issues/70808 | https://api.github.com/repos/openclaw/openclaw/issues/70808 | [Bug] Gateway never disposes stdio MCP runtimes on session end — child processes accumulate | ## Description
Gateway-managed sessions (Telegram, webchat, cron) spawn stdio MCP child processes via `pi-bundle-mcp-runtime` (`createSessionMcpRuntime`), but never call `disposeSessionMcpRuntime()` when sessions end. Processes accumulate indefinitely until the gateway restarts.
## Reproduction
1. Configure a stdio ... | closed | completed | true | 2 | [] | [] | 2026-04-23T22:24:07Z | 2026-04-30T04:44:39Z | 2026-04-25T07:16:04Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | dertbv | 3,845,627 | MDQ6VXNlcjM4NDU2Mjc= | User | false |
openclaw/openclaw | 4,319,237,368 | I_kwDOQb6kR88AAAABAXJU-A | 70,810 | https://github.com/openclaw/openclaw/issues/70810 | https://api.github.com/repos/openclaw/openclaw/issues/70810 | [Bug]: Discord slash commands stuck "Loading" in Thread channels — GuildThreadChannel.get parentId fails on partial rawData | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Discord slash commands (e.g. /status) in Guild Thread channels fail with "Cannot access rawData on partial Channel" error, causing infinite "Loading..." state.
### Steps to reproduce
1. Set `channels.discord.threadBindings... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-04-23T22:25:27Z | 2026-04-26T04:29:22Z | 2026-04-24T01:19:43Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | lzyling | 23,039,860 | MDQ6VXNlcjIzMDM5ODYw | User | false |
openclaw/openclaw | 4,319,314,998 | I_kwDOQb6kR88AAAABAXOENg | 70,818 | https://github.com/openclaw/openclaw/issues/70818 | https://api.github.com/repos/openclaw/openclaw/issues/70818 | [Bug]: Discord provider fails with "Cannot find package 'openclaw'" in profile-specific plugin-runtime-deps after global npm upgrade | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
After upgrading OpenClaw globally via sudo npm install -g openclaw@latest, the Discord channel fails to connect on named gateway profiles (e.g. --profile mneme, --profile peitho) with:
[discord] [default] chann... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-23T22:45:08Z | 2026-04-26T04:29:21Z | 2026-04-24T01:19:46Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | craig-c63 | 15,031,680 | MDQ6VXNlcjE1MDMxNjgw | User | false |
openclaw/openclaw | 4,319,313,936 | I_kwDOQb6kR88AAAABAXOAEA | 70,817 | https://github.com/openclaw/openclaw/issues/70817 | https://api.github.com/repos/openclaw/openclaw/issues/70817 | CLI `openclaw memory status --deep` fails with "Unknown memory embedding provider: local" in v2026.4.22 | ## Summary
In v2026.4.22, the CLI command `openclaw memory status --deep` fails immediately with:
```
Error: Unknown memory embedding provider: local
at getAdapter (file:///.../dist/manager-xvPcitvj.js:143:22)
at createEmbeddingProvider (file:///.../dist/manager-xvPcitvj.js:195:25)
at MemoryIndexManager.l... | closed | completed | true | 1 | [] | [] | 2026-04-23T22:44:54Z | 2026-04-26T04:29:12Z | 2026-04-24T02:10:16Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | clawdiaboat | 262,558,217 | U_kgDOD6ZSCQ | User | false |
openclaw/openclaw | 4,319,346,855 | I_kwDOQb6kR88AAAABAXQApw | 70,823 | https://github.com/openclaw/openclaw/issues/70823 | https://api.github.com/repos/openclaw/openclaw/issues/70823 | RFC: machine-readable memory-slot ownership — `plugins.slots.<slot>.owner` | ## Summary
Today, a plugin that claims an OpenClaw slot — e.g. `plugins.slots.memory = "plur-claw"` — can only be *detected* as the owner of that slot by string-comparing the slot value against its own plugin ID. That comparison works for happy-path installs, but it collapses as soon as two plugins want the same slot:... | open | null | false | 1 | [] | [] | 2026-04-23T22:54:11Z | 2026-05-03T12:31:30Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | plur9 | 1,336,797 | MDQ6VXNlcjEzMzY3OTc= | User | false |
openclaw/openclaw | 4,319,372,950 | I_kwDOQb6kR88AAAABAXRmlg | 70,825 | https://github.com/openclaw/openclaw/issues/70825 | https://api.github.com/repos/openclaw/openclaw/issues/70825 | [acpx] formatDoctorFailureMessage serializes non-string details as "[object Object]" | ## Summary
`formatDoctorFailureMessage` in `extensions/acpx/src/service.ts` (lines 82–85) assumes `report.details` is `string[]`, but the upstream `acpx` npm package's `doctor()` can return non-string values (Error instances, plain objects). `Array.prototype.join` coerces those via `String(x)`, which yields the litera... | closed | completed | true | 1 | [] | [] | 2026-04-23T23:02:09Z | 2026-04-27T04:33:34Z | 2026-04-25T04:15:18Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | hexsprite | 23,538 | MDQ6VXNlcjIzNTM4 | User | false |
openclaw/openclaw | 4,319,425,097 | I_kwDOQb6kR88AAAABAXUySQ | 70,830 | https://github.com/openclaw/openclaw/issues/70830 | https://api.github.com/repos/openclaw/openclaw/issues/70830 | control ui: replace assistant-media ?token auth with scoped media tickets | ## Problem
Control UI local assistant attachments currently use `assistant-media?...&token=...` URLs.
That `token` can be a real shared gateway credential or a paired device token. This works, but it means a real auth credential appears in browser-visible URLs.
## Why this is follow-up work
PR #70741 keeps this beh... | open | null | false | 0 | [
"maintainer"
] | [] | 2026-04-23T23:17:44Z | 2026-04-23T23:17:55Z | null | CONTRIBUTOR | null | 20260424T000030Z | 2026-04-24T00:00:30Z | Patrick-Erichsen | 20,157,849 | MDQ6VXNlcjIwMTU3ODQ5 | User | false |
openclaw/openclaw | 4,319,390,918 | I_kwDOQb6kR88AAAABAXSsxg | 70,827 | https://github.com/openclaw/openclaw/issues/70827 | https://api.github.com/repos/openclaw/openclaw/issues/70827 | [Bug]: Venice live models can be rejected by stale configured allowlist | ## Bug type
Behavior bug (incorrect output/state without crash)
## Beta release blocker
No
## Summary
OpenClaw rejects newly advertised Venice text models such as `venice/kimi-k2-6` and `venice/qwen-3-6-plus` at the local allowed-model check even though Venice's live `/models` endpoint advertises those IDs.
## St... | closed | completed | true | 1 | [] | [] | 2026-04-23T23:07:24Z | 2026-04-28T04:47:20Z | 2026-04-25T19:28:17Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Jackten | 2,895,479 | MDQ6VXNlcjI4OTU0Nzk= | User | false |
openclaw/openclaw | 4,104,429,027 | I_kwDOQb6kR870pJ3j | 50,659 | https://github.com/openclaw/openclaw/issues/50659 | https://api.github.com/repos/openclaw/openclaw/issues/50659 | [Bug]: agents.defaults.imageGenerationModel unrecognized by config validation / doctor in 2026.3.13 (61d171a) – key gets stripped on --fix | ### Bug type
Regression (worked before, now fails)
### Summary
The config key agents.defaults.imageGenerationModel (used for setting primary/fallbacks for the image_generate tool and image generation in agents) is rejected as an "Unrecognized key" during config validation (or gateway restarts). Running openclaw doct... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-03-19T21:59:52Z | 2026-04-26T04:29:49Z | 2026-04-23T23:24:25Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | whoisheartbreak | 165,145,263 | U_kgDOCdfqrw | User | false |
openclaw/openclaw | 4,149,742,704 | I_kwDOQb6kR873WAxw | 55,601 | https://github.com/openclaw/openclaw/issues/55601 | https://api.github.com/repos/openclaw/openclaw/issues/55601 | Bug: Google Image Generation Provider throws 'No image-generation provider registered for google' | ### Describe the bug
When attempting to use the `image_generate` tool with a Google Gemini model configured as the `imageGenerationModel`, the system throws the following error:
```
Error: No image-generation provider registered for google
```
However, inspecting the codebase (e.g., in `plugin-google` / `pi-embedded... | closed | completed | true | 1 | [] | [] | 2026-03-27T06:09:04Z | 2026-04-26T04:29:48Z | 2026-04-23T23:24:26Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | Liu-XinYuan | 20,487,589 | MDQ6VXNlcjIwNDg3NTg5 | User | false |
openclaw/openclaw | 4,205,327,414 | I_kwDOQb6kR876qDQ2 | 60,959 | https://github.com/openclaw/openclaw/issues/60959 | https://api.github.com/repos/openclaw/openclaw/issues/60959 | [Bug]: Google image generation returns HTTP 404 when provider baseUrl includes /openai suffix | ## Description
The `image_generate` tool returns HTTP 404 when using a Google provider configured with the OpenAI-compatible base URL (`https://generativelanguage.googleapis.com/v1beta/openai`).
## Root Cause
In `extensions/google/image-generation-provider.ts`, `resolveGoogleBaseUrl()` calls `normalizeGoogleApiBaseU... | closed | completed | true | 1 | [] | [] | 2026-04-04T16:43:17Z | 2026-04-26T04:29:48Z | 2026-04-23T23:24:28Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | Max-Resilient | 64,356,730 | MDQ6VXNlcjY0MzU2NzMw | User | false |
openclaw/openclaw | 4,145,602,240 | I_kwDOQb6kR873GN7A | 55,277 | https://github.com/openclaw/openclaw/issues/55277 | https://api.github.com/repos/openclaw/openclaw/issues/55277 | [Bug]: Image tool fails with "No media-understanding provider registered for openrouter" (regression in v2026.3.24) | **Regression** (worked before, now fails)
After upgrading to v2026.3.24, the `image` tool fails for all OpenRouter-hosted models with:
```
[tools] image failed: All image models failed (N):
openrouter/google/gemini-2.5-flash: No media-understanding provider registered for openrouter
openrouter/openai/gpt-5-mini: No me... | closed | completed | true | 1 | [] | [] | 2026-03-26T16:41:25Z | 2026-04-26T04:29:46Z | 2026-04-23T23:24:40Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | illavia | 22,727,025 | MDQ6VXNlcjIyNzI3MDI1 | User | false |
openclaw/openclaw | 4,319,445,292 | I_kwDOQb6kR88AAAABAXWBLA | 70,833 | https://github.com/openclaw/openclaw/issues/70833 | https://api.github.com/repos/openclaw/openclaw/issues/70833 | Muninn fails to create feather from current chat context; should capture current delimited chat verbatim like Yggdrasill note-taking | ## Summary
During UAT on the canonical Muninn branch, `mundu` does not create a feather from the current chat context when invoked with a hint. Instead, the runtime fails with `Error: captured context is empty.`
## Observed behavior
Environment / canonical surface used:
- Branch: `/home/ubuntu/.openclaw/workspace/yggd... | closed | completed | true | 1 | [] | [] | 2026-04-23T23:24:42Z | 2026-04-26T04:29:45Z | 2026-04-23T23:26:35Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | dlucarelli-insight | 138,123,385 | U_kgDOCDuYeQ | User | false |
openclaw/openclaw | 4,102,376,456 | I_kwDOQb6kR870hUwI | 50,485 | https://github.com/openclaw/openclaw/issues/50485 | https://api.github.com/repos/openclaw/openclaw/issues/50485 | Support OpenRouter models in image tool (agents.defaults.imageModel) | ## Problem
The `image` tool rejects `openrouter/xiaomi/mimo-v2-omni` with `Unknown model` even though Mimo v2 Omni supports vision on OpenRouter.
## Root Cause
The image tool uses a **media-understanding provider registry** with only these providers:
- Hardcoded: `groq`, `deepgram`
- Extensions: `anthropic`, `goog... | closed | duplicate | true | 4 | [] | [] | 2026-03-19T15:14:03Z | 2026-04-26T04:29:43Z | 2026-04-23T23:27:05Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | mculp | 129,983 | MDQ6VXNlcjEyOTk4Mw== | User | false |
openclaw/openclaw | 4,192,015,561 | I_kwDOQb6kR8753RTJ | 59,486 | https://github.com/openclaw/openclaw/issues/59486 | https://api.github.com/repos/openclaw/openclaw/issues/59486 | [Bug]: Google Gemini Image Generation Failure | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
The `image_generate` tool returns a `fetch failed` error when calling the Google Gemini image generation model, but direct API calls to Google work successfully.
### Error Message
```
All image generation models failed (2):
... | closed | duplicate | true | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-02T06:22:58Z | 2026-04-26T04:29:42Z | 2026-04-23T23:27:08Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | Jageri | 14,195,779 | MDQ6VXNlcjE0MTk1Nzc5 | User | false |
openclaw/openclaw | 4,046,397,549 | I_kwDOQb6kR87xLyBt | 41,210 | https://github.com/openclaw/openclaw/issues/41210 | https://api.github.com/repos/openclaw/openclaw/issues/41210 | Image model config (google/gemini) ignored - hardcoded openai/gpt-5-mini fallback | [issue-image-model.md](https://github.com/user-attachments/files/25847069/issue-image-model.md) | closed | completed | true | 1 | [] | [] | 2026-03-09T15:53:26Z | 2026-04-26T04:29:42Z | 2026-04-23T23:27:09Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | jingyi22 | 37,988,577 | MDQ6VXNlcjM3OTg4NTc3 | User | false |
openclaw/openclaw | 4,211,244,774 | I_kwDOQb6kR877An7m | 61,828 | https://github.com/openclaw/openclaw/issues/61828 | https://api.github.com/repos/openclaw/openclaw/issues/61828 | [Bug]: Gemini image generation fails | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
After upgrading to the latest version of openclaw today, the Gemini image generation no longer works. Calling Gemini for text generation works fine (my day-to-day model is Kimi 2.5), but the image generation do... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-06T11:05:35Z | 2026-04-26T04:29:26Z | 2026-04-24T01:06:16Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | ajensenwaud | 5,827,908 | MDQ6VXNlcjU4Mjc5MDg= | User | false |
openclaw/openclaw | 4,319,455,262 | I_kwDOQb6kR88AAAABAXWoHg | 70,834 | https://github.com/openclaw/openclaw/issues/70834 | https://api.github.com/repos/openclaw/openclaw/issues/70834 | Add support for GPT-5.5 in openai-codex provider (released 2026-04-23) | ## Summary
GPT-5.5 was released today (April 23, 2026) and is available on ChatGPT and Codex for all paid subscribers. However, the `openai-codex` provider in OpenClaw does not recognize `gpt-5.5` as a valid model.
## Current Behavior
- Setting `openai-codex/gpt-5.5` as primary model results in: `Unknown model: open... | closed | completed | true | 1 | [] | [] | 2026-04-23T23:27:49Z | 2026-04-26T04:29:28Z | 2026-04-24T00:49:22Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | cianbyrne1010 | 272,894,325 | U_kgDOEEQJdQ | User | false |
openclaw/openclaw | 4,319,207,339 | I_kwDOQb6kR88AAAABAXHfqw | 70,806 | https://github.com/openclaw/openclaw/issues/70806 | https://api.github.com/repos/openclaw/openclaw/issues/70806 | [Bug]: multiple "[object Object]" in chat | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Quite often, I get multiple "[object Object]" in answers.
### Steps to reproduce
1. use OpenClaw 2026.4.22 (00bd2cf)
2. Setup model Mistral small or large (perhaps happen with other models) without thinking
... | closed | completed | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-04-23T22:17:47Z | 2026-04-29T04:34:30Z | 2026-04-26T18:57:57Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | fmatray | 8,267,716 | MDQ6VXNlcjgyNjc3MTY= | User | false |
openclaw/openclaw | 4,319,488,212 | I_kwDOQb6kR88AAAABAXYo1A | 70,835 | https://github.com/openclaw/openclaw/issues/70835 | https://api.github.com/repos/openclaw/openclaw/issues/70835 | [Bug]: Stable 2026.4.22 cannot expose/configure openai-codex GPT-5.5 despite docs/account availability | ## Bug type
Behavior bug
## Summary
On stable OpenClaw 2026.4.22 with working `openai-codex` OAuth, `gpt-5.5` is documented as supported and became available in the ChatGPT account, but the local OpenClaw model catalog/config path does not expose a way to add or select it.
## Steps to reproduce
1. Run OpenClaw `2026.... | closed | not_planned | true | 1 | [] | [] | 2026-04-23T23:36:54Z | 2026-04-26T04:29:34Z | 2026-04-24T00:23:20Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | barbarhan | 80,099,286 | MDQ6VXNlcjgwMDk5Mjg2 | User | false |
openclaw/openclaw | 4,319,424,509 | I_kwDOQb6kR88AAAABAXUv_Q | 70,829 | https://github.com/openclaw/openclaw/issues/70829 | https://api.github.com/repos/openclaw/openclaw/issues/70829 | fix: propagate timeoutMs to guarded dispatchers (local LLM 60s timeout) | ## Problem
Local LLM requests via `openai-completions` (and other provider transports) are killed at exactly **60 seconds**, despite having `timeoutSeconds: 900` (15 min) configured in `openclaw.json`.
## Root Cause
Timeout propagation gap in the guarded fetch path.
OpenClaw correctly sets a global undici dispatche... | closed | completed | true | 3 | [
"dedupe:parent"
] | [
"vincentkoc"
] | 2026-04-23T23:17:30Z | 2026-04-26T04:28:05Z | 2026-04-24T03:34:12Z | CONTRIBUTOR | null | 20260426T060034Z | 2026-04-26T06:00:34Z | DranboFieldston | 258,684,645 | U_kgDOD2s25Q | User | false |
openclaw/openclaw | 4,319,498,725 | I_kwDOQb6kR88AAAABAXZR5Q | 70,836 | https://github.com/openclaw/openclaw/issues/70836 | https://api.github.com/repos/openclaw/openclaw/issues/70836 | [Bug]: openclaw memory index fails with 'Unknown memory embedding provider: local' while gateway memory is vector ready | ## Summary
`openclaw memory index --verbose` fails with `Unknown memory embedding provider: local` even when local embeddings are configured correctly, `node-llama-cpp` is installed, and the running gateway reports memory vector support as ready.
There also appears to be a separate native stability problem in the loca... | closed | completed | true | 1 | [] | [] | 2026-04-23T23:39:36Z | 2026-04-26T04:29:12Z | 2026-04-24T02:09:51Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | mogglemoss | 10,966,195 | MDQ6VXNlcjEwOTY2MTk1 | User | false |
openclaw/openclaw | 4,319,511,093 | I_kwDOQb6kR88AAAABAXaCNQ | 70,837 | https://github.com/openclaw/openclaw/issues/70837 | https://api.github.com/repos/openclaw/openclaw/issues/70837 | Bug: tavily extension fails to load - typebox missing collapse.mjs | ## Bug Report
**OpenClaw version:** 2026.4.22
**Error:** [tools] read failed: EISDIR / avily failed to load
## Description
The avily extension fails to load at startup with the following error:
`
[plugins] tavily failed to load from C:\Users\Administrator\AppData\Local\nvm\v22.22.2\node_modules\openclaw\dist\exte... | closed | completed | true | 1 | [] | [] | 2026-04-23T23:42:57Z | 2026-04-28T04:47:19Z | 2026-04-25T19:28:23Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | visionzhang-prog | 266,700,237 | U_kgDOD-WFzQ | User | false |
openclaw/openclaw | 4,075,969,640 | I_kwDOQb6kR87y8lxo | 46,255 | https://github.com/openclaw/openclaw/issues/46255 | https://api.github.com/repos/openclaw/openclaw/issues/46255 | Bug: Images not passed to OpenRouter models — input_image format not recognized | ## Bug Description
When using OpenRouter as a provider with image-capable models (e.g., `openrouter/healer-alpha`), images sent through the `read` tool or as inbound message attachments are not recognized by the model. The model responds with "I don't see an image attached".
## Root Cause
OpenClaw's `contentToOpenAI... | closed | completed | true | 2 | [] | [] | 2026-03-14T14:53:37Z | 2026-04-26T04:29:40Z | 2026-04-23T23:46:48Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | witnesssun | 111,419,902 | U_kgDOBqQh_g | User | false |
openclaw/openclaw | 4,319,537,015 | I_kwDOQb6kR88AAAABAXbndw | 70,838 | https://github.com/openclaw/openclaw/issues/70838 | https://api.github.com/repos/openclaw/openclaw/issues/70838 | [Bug]: Config write corrupts sensitive fields to __OPENCLAW_REDACTED__ sentinel (memorySearch, qqbot) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
**GitHub Issue 草稿：**
---
**Config write replaces valid secrets with `__OPENCLAW_REDACTED__` sentinel, causing field corruption**
**Description**
When certain config write operations run (e.g. partial updates via Control ... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-23T23:50:02Z | 2026-04-28T04:47:18Z | 2026-04-25T19:28:30Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | iamyhzhao | 277,945,921 | U_kgDOEJEeQQ | User | false |
openclaw/openclaw | 4,221,864,091 | I_kwDOQb6kR877pIib | 62,879 | https://github.com/openclaw/openclaw/issues/62879 | https://api.github.com/repos/openclaw/openclaw/issues/62879 | [Bug]: image_generate blocks local OpenAI-compatible endpoint on private IP (SSRF) even with browser.ssrfPolicy.dangerouslyAllowPrivateNetwork=true | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
## OpenClaw version
OpenClaw 2026.4.5 (3e72c03)
## What I’m trying to do
Use `image_generate` with a local OpenAI-compatible image endpoint (LocalAI) on LAN:
- Text model: vLLM (`http://192.168.1.15:8000/v1`)
... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-08T03:24:42Z | 2026-04-26T04:29:38Z | 2026-04-23T23:50:54Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | seitzbg | 171,158 | MDQ6VXNlcjE3MTE1OA== | User | false |
openclaw/openclaw | 4,319,540,831 | I_kwDOQb6kR88AAAABAXb2Xw | 70,839 | https://github.com/openclaw/openclaw/issues/70839 | https://api.github.com/repos/openclaw/openclaw/issues/70839 | [Bug]: acpx@0.5.3 runtime dep not staged on npm global install of 2026.4.22 — ACP plugin fails to load with 'Cannot find module acpx/runtime' | ## Summary
After upgrading to `openclaw@2026.4.22` via `npm install -g openclaw`, the bundled **ACPX Runtime** plugin fails to load with:
```
[plugins] acpx failed to load from /usr/lib/node_modules/openclaw/dist/extensions/acpx/index.js: Error: Cannot find module 'acpx/runtime'
Require stack:
- /usr/lib/node_modules... | closed | completed | true | 2 | [] | [] | 2026-04-23T23:51:14Z | 2026-04-26T04:29:10Z | 2026-04-24T02:28:13Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | Dalton-AI-open | 259,261,635 | U_kgDOD3QEww | User | false |
openclaw/openclaw | 4,068,719,172 | I_kwDOQb6kR87yg7pE | 44,648 | https://github.com/openclaw/openclaw/issues/44648 | https://api.github.com/repos/openclaw/openclaw/issues/44648 | Image tool ignores configured imageModel, uses first vision-capable model in catalog instead | ## Summary
The image analysis tool does not use `agents.defaults.imageModel` when selecting which model to call. Instead it appears to pick the first vision-capable model it finds in the model catalog. This makes `imageModel` configuration effectively non-functional when any other vision-capable model appears earlier ... | closed | completed | true | 3 | [] | [] | 2026-03-13T04:14:02Z | 2026-04-26T04:29:37Z | 2026-04-23T23:52:19Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | svankirk | 3,734,797 | MDQ6VXNlcjM3MzQ3OTc= | User | false |
openclaw/openclaw | 4,319,564,777 | I_kwDOQb6kR88AAAABAXdT6Q | 70,840 | https://github.com/openclaw/openclaw/issues/70840 | https://api.github.com/repos/openclaw/openclaw/issues/70840 | [Bug]: Control UI/TUI can inject false "Bootstrap pending" wrapper even when workspace bootstrap is complete | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
In a workspace where `BOOTSTRAP.md` explicitly says bootstrap is complete, new turns/sessions in Control UI/TUI can still prepend a `Bootstrap pending` wrapper telling the agent to read and complete `BOOTSTRAP.... | closed | completed | true | 4 | [] | [
"Patrick-Erichsen"
] | 2026-04-23T23:58:39Z | 2026-04-27T04:49:25Z | 2026-04-24T22:41:12Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | GSUCAS | 201,416,672 | U_kgDODAFf4A | User | false |
openclaw/openclaw | 4,319,590,097 | I_kwDOQb6kR88AAAABAXe20Q | 70,841 | https://github.com/openclaw/openclaw/issues/70841 | https://api.github.com/repos/openclaw/openclaw/issues/70841 | [Bug]: Discord plugin: post-restart gateway wedges at "awaiting gateway readiness" on macOS (no 15s timeout fires) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Summary
On macOS, restarting the OpenClaw gateway often leaves the Discord plugin permanently stuck at `awaiting gateway readiness`. The underlying TCP/TLS connection to `gateway.discord.gg` is established, but the Disco... | closed | completed | true | 1 | [
"bug",
"duplicate",
"regression"
] | [] | 2026-04-24T00:06:39Z | 2026-04-27T04:55:20Z | 2026-04-24T21:43:07Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | HeilbronAILabs | 274,369,335 | U_kgDOEFqLNw | User | false |
openclaw/openclaw | 4,319,594,737 | I_kwDOQb6kR88AAAABAXfI8Q | 70,842 | https://github.com/openclaw/openclaw/issues/70842 | https://api.github.com/repos/openclaw/openclaw/issues/70842 | memory-wiki bridge mode: 0 artifacts with QMD on 2026.4.21 (regression from #65976 #69019) | # Bug Report: memory-wiki bridge mode still reports 0 artifacts on OpenClaw 2026.4.21
## Summary
memory-wiki in bridge mode continues to report `bridgePublicArtifactCount: 0` and imports 0 artifacts on OpenClaw **2026.4.21** (latest stable as of 2026-04-24), despite healthy QMD-backed memory and memory-core plugin be... | closed | completed | true | 2 | [
"dedupe:child"
] | [
"vincentkoc"
] | 2026-04-24T00:08:12Z | 2026-04-29T04:29:17Z | 2026-04-26T08:23:53Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | vanhemn | 5,332,392 | MDQ6VXNlcjUzMzIzOTI= | User | false |
openclaw/openclaw | 4,319,606,442 | I_kwDOQb6kR88AAAABAXf2qg | 70,843 | https://github.com/openclaw/openclaw/issues/70843 | https://api.github.com/repos/openclaw/openclaw/issues/70843 | channels.discord: broaden voice-note transcription beyond mention-required channels |
### Problem
Discord voice-note attachments only get pre-flight transcribed in a very narrow case. The condition lives at [`extensions/discord/src/monitor/preflight-audio.ts`](https://github.com/openclaw/openclaw/blob/main/extensions/discord/src/monitor/preflight-audio.ts) (compiled form shown here):
```js
const need... | closed | completed | true | 1 | [] | [] | 2026-04-24T00:11:39Z | 2026-04-30T04:49:06Z | 2026-04-25T05:53:47Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | bigmodizel93 | 278,831,245 | U_kgDOEJ6gjQ | User | false |
openclaw/openclaw | 4,175,167,482 | I_kwDOQb6kR8742__6 | 58,059 | https://github.com/openclaw/openclaw/issues/58059 | https://api.github.com/repos/openclaw/openclaw/issues/58059 | OpenRouter model context window not auto-detected from provider catalog | ## Issue
When a model is listed in `agents.defaults.models` (the allowlist/catalog), OpenClaw does not automatically resolve its `contextWindow` from the provider's API. Instead, it falls back to a default (200k).
### Reproduction
1. Add a model to `agents.defaults.models`:
```json
"openrouter/qwen/qwen3.6-plu... | closed | completed | true | 2 | [] | [] | 2026-03-31T02:06:29Z | 2026-04-26T04:29:33Z | 2026-04-24T00:40:55Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | barbarhan | 80,099,286 | MDQ6VXNlcjgwMDk5Mjg2 | User | false |
openclaw/openclaw | 4,167,548,152 | I_kwDOQb6kR874Z7z4 | 57,498 | https://github.com/openclaw/openclaw/issues/57498 | https://api.github.com/repos/openclaw/openclaw/issues/57498 | Bug: Image recognition bypasses OpenRouter attribution headers, causing unexpected charges | ## Summary
When OpenClaw processes images (via `describeImageWithModel`) or performs context compaction (`/compact`), the OpenRouter attribution headers (`HTTP-Referer`, `X-OpenRouter-Title`, `X-OpenRouter-Categories`) are **not sent**. This causes OpenRouter to treat these requests as unidentified, resulting in charg... | closed | completed | true | 1 | [] | [] | 2026-03-30T05:32:37Z | 2026-04-26T04:29:31Z | 2026-04-24T00:48:21Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | deepvoce | 147,256,182 | U_kgDOCMbzdg | User | false |
openclaw/openclaw | 4,154,325,868 | I_kwDOQb6kR873nfts | 55,914 | https://github.com/openclaw/openclaw/issues/55914 | https://api.github.com/repos/openclaw/openclaw/issues/55914 | [Feature]: openclaw invite — shareable invite codes for mobile pairing | ## Summary
Add an `openclaw invite` command that generates a short invite code (e.g., `CLAW-7X9K`) and registers it with a configurable cloud backend, so mobile users can discover and connect to a Gateway by typing a code instead of scanning a QR.
## Problem to solve
QR code pairing (`openclaw pair`) requires the mo... | open | null | false | 3 | [] | [] | 2026-03-27T16:44:01Z | 2026-05-03T07:44:30Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | ovai-felix | 185,858,100 | U_kgDOCxP4NA | User | false |
openclaw/openclaw | 4,319,746,928 | I_kwDOQb6kR88AAAABAXobcA | 70,853 | https://github.com/openclaw/openclaw/issues/70853 | https://api.github.com/repos/openclaw/openclaw/issues/70853 | Telegram DM session continuity lost after session reset/rebind on stable session key | ## Summary
Telegram DM session continuity can break when OpenClaw resets/rebinds a live session for an existing stable session key. In my case, the same Telegram direct-chat session key was rebound to a fresh `sessionId`, and the new live transcript began at a later inbound message instead of preserving or replaying pr... | closed | completed | true | 0 | [] | [] | 2026-04-24T00:53:03Z | 2026-05-01T04:42:44Z | 2026-04-28T17:01:17Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | clawdjons-code | 270,555,536 | U_kgDOECBZkA | User | false |
openclaw/openclaw | 4,319,758,335 | I_kwDOQb6kR88AAAABAXpH_w | 70,854 | https://github.com/openclaw/openclaw/issues/70854 | https://api.github.com/repos/openclaw/openclaw/issues/70854 | Add gpt-5.5 catalog support (OpenAI in-house model, released 2026-04-23) | ## Summary
OpenAI released **GPT-5.5** on 2026-04-23, available through the Codex OAuth flow for ChatGPT Plus/Pro/Business/Enterprise subscribers. As of 2026-04-24, **OpenClaw v2026.4.22** (bundled `@mariozechner/pi-ai` 0.69.0) still does not include `gpt-5.5` in its model catalog (`models.generated.d.ts`), which caus... | closed | completed | true | 5 | [] | [] | 2026-04-24T00:56:36Z | 2026-04-27T04:30:07Z | 2026-04-24T05:21:37Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | mjt0926 | 212,995,131 | U_kgDODLIMOw | User | false |
openclaw/openclaw | 4,319,796,963 | I_kwDOQb6kR88AAAABAXre4w | 70,857 | https://github.com/openclaw/openclaw/issues/70857 | https://api.github.com/repos/openclaw/openclaw/issues/70857 | [Bug]: Windows startup and reply latency, session lock held for 191s on sessions.json.lock | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
### Environment
- OpenClaw version: `2026.4.15 (041266a)`
- OS: Windows
- Install method: npm global install
- Gateway service: Scheduled Task
- State dir: `C:\Users\Wapapapow\.openclaw`
- Workspace: `F:\openclaw\workspace`
... | closed | completed | true | 6 | [
"bug",
"regression"
] | [] | 2026-04-24T01:07:42Z | 2026-05-05T04:27:17Z | 2026-05-02T13:02:07Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | Kanorin-chan | 27,182,052 | MDQ6VXNlcjI3MTgyMDUy | User | false |
openclaw/openclaw | 4,319,823,757 | I_kwDOQb6kR88AAAABAXtHjQ | 70,860 | https://github.com/openclaw/openclaw/issues/70860 | https://api.github.com/repos/openclaw/openclaw/issues/70860 | [Bug]: ACP Claude sessions leave orphaned processes after completion (memory leak) | ## Description
When `sessions_spawn` is called with `runtime: "acp"` (using Claude Code as the agent), the spawned `claude` and `claude-agent-acp` processes remain running after the session completes. OpenClaw's `subagents list` shows 0 active sessions, but the processes are still alive in the OS process table.
Over ... | closed | not_planned | true | 1 | [] | [] | 2026-04-24T01:13:38Z | 2026-04-29T04:41:50Z | 2026-04-26T07:24:54Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | swiser-nexa | 269,324,650 | U_kgDOEA2Rag | User | false |
openclaw/openclaw | 4,319,778,979 | I_kwDOQb6kR88AAAABAXqYow | 70,856 | https://github.com/openclaw/openclaw/issues/70856 | https://api.github.com/repos/openclaw/openclaw/issues/70856 | [Bug]: WhatsApp listener repeatedly disconnects/stalls on Windows, causing "No active WhatsApp Web listener" and missed messages | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
### Environment
- OpenClaw version: `2026.4.15 (041266a)`
- OS: Windows
- Install method: npm global install
- Gateway service: Scheduled Task
- State dir: `C:\Users\Wapapapow\.openclaw`
- Workspace: `F:\openclaw\workspace`
... | open | null | false | 5 | [
"bug",
"regression"
] | [] | 2026-04-24T01:03:24Z | 2026-05-02T05:38:17Z | null | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | Kanorin-chan | 27,182,052 | MDQ6VXNlcjI3MTgyMDUy | User | false |
openclaw/openclaw | 4,079,114,113 | I_kwDOQb6kR87zIleB | 47,614 | https://github.com/openclaw/openclaw/issues/47614 | https://api.github.com/repos/openclaw/openclaw/issues/47614 | [Bug] External vision model (tools.media.image) skipped when primary model has native vision in built-in catalog | ## Summary
When `tools.media.image.models` is configured with an external vision model (e.g. Gemini via OpenRouter), the model is never called if the primary model is listed as supporting vision natively in the `@mariozechner/pi-coding-agent` built-in model catalog.
## Environment
- OpenClaw: 2026.3.13
- macOS 14.8.... | closed | completed | true | 4 | [] | [] | 2026-03-15T20:17:21Z | 2026-04-26T04:29:20Z | 2026-04-24T01:23:43Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | caali7116-sudo | 259,935,283 | U_kgDOD35MMw | User | false |
openclaw/openclaw | 4,069,247,178 | I_kwDOQb6kR87yi8jK | 44,762 | https://github.com/openclaw/openclaw/issues/44762 | https://api.github.com/repos/openclaw/openclaw/issues/44762 | [Bug]: minimax-portal image handling routes MiniMax chat models incorrectly instead of dedicated VLM path | ## Summary
`minimax-portal` image understanding appears miswired in current OpenClaw.
Two separate issues combine into a bad UX:
1. `extensions/minimax-portal-auth/index.ts` registers `MiniMax-M2.5`, `MiniMax-M2.5-highspeed`, and `MiniMax-M2.5-Lightning` as `input: ["text"]`.
2. media-understanding active-model fallb... | closed | completed | true | 2 | [] | [] | 2026-03-13T06:37:43Z | 2026-04-26T04:29:17Z | 2026-04-24T01:27:46Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | MazzaWill | 32,673,411 | MDQ6VXNlcjMyNjczNDEx | User | false |
openclaw/openclaw | 4,319,884,288 | I_kwDOQb6kR88AAAABAXw0AA | 70,868 | https://github.com/openclaw/openclaw/issues/70868 | https://api.github.com/repos/openclaw/openclaw/issues/70868 | Plugin runtime: prependSystemContext via before_prompt_build causes occasional empty model outputs | ## Bug Report
**Severity:** High — causes silent session corruption
**Summary:**
Using `prependSystemContext` via the `before_prompt_build` hook causes occasional empty model outputs (`stop=stop`, `content=[]`). The issue was reproduced twice in the same day after the hook was enabled.
**Affected versions:** Confirm... | closed | completed | true | 2 | [] | [] | 2026-04-24T01:29:43Z | 2026-04-29T04:36:56Z | 2026-04-26T10:53:35Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Zen310 | 114,278,746 | U_kgDOBs_BWg | User | false |
openclaw/openclaw | 4,319,867,451 | I_kwDOQb6kR88AAAABAXvyOw | 70,865 | https://github.com/openclaw/openclaw/issues/70865 | https://api.github.com/repos/openclaw/openclaw/issues/70865 | [Bug]: QQbot在定时任务中用不了 cron 工具 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
测试了飞书是可以调用的，控制台ui 创建 也是正常的，唯独在 qq 机器人里用不了。
虽然通过exec也能正常创建。但是还是想知道原因。
AI回复：
你说得对，我应该按 skill 流程走的。但问题是——qqbot_remind 返回的 _instruction 说要用 "cron 工具"，但我查了一圈，OpenClaw 并没有单独提供 cron 这个函数工具，只能通过 exec 调用 openclaw cron add 命令。
所以实... | closed | completed | true | 0 | [
"bug",
"regression"
] | [] | 2026-04-24T01:25:10Z | 2026-04-28T04:45:47Z | 2026-04-25T23:15:30Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | wwcdg | 48,498,475 | MDQ6VXNlcjQ4NDk4NDc1 | User | false |
openclaw/openclaw | 4,319,885,209 | I_kwDOQb6kR88AAAABAXw3mQ | 70,869 | https://github.com/openclaw/openclaw/issues/70869 | https://api.github.com/repos/openclaw/openclaw/issues/70869 | delivery-mirror: writes role:assistant into parent session without paired user message, violating Anthropic alternation | ## Bug Report
**Severity:** High — causes silent session corruption and API rejection
**Summary:**
When a cron job or subagent session uses the `message` tool to send outbound messages (e.g., Telegram alerts), OpenClaw writes the sent text back into the **parent/originating session** transcript as `role: assistant` w... | closed | completed | true | 1 | [] | [] | 2026-04-24T01:30:00Z | 2026-04-29T04:33:35Z | 2026-04-26T22:19:42Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Zen310 | 114,278,746 | U_kgDOBs_BWg | User | false |
openclaw/openclaw | 4,319,886,344 | I_kwDOQb6kR88AAAABAXw8CA | 70,870 | https://github.com/openclaw/openclaw/issues/70870 | https://api.github.com/repos/openclaw/openclaw/issues/70870 | LCM: no invalidation or recovery path from .jsonl edits; DB accumulates structurally invalid messages permanently | ## Bug Report / Feature Request
**Severity:** Medium — no data loss, but no recovery path once corruption accumulates
**Summary:**
`~/.openclaw/lcm.db` accumulates every failed attempt, empty assistant message, delivery-mirror entry, and malformed message permanently. There is no mechanism to invalidate the LCM DB ba... | closed | not_planned | true | 1 | [] | [] | 2026-04-24T01:30:17Z | 2026-04-28T04:47:16Z | 2026-04-25T19:28:37Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Zen310 | 114,278,746 | U_kgDOBs_BWg | User | false |
openclaw/openclaw | 4,319,899,660 | I_kwDOQb6kR88AAAABAXxwDA | 70,871 | https://github.com/openclaw/openclaw/issues/70871 | https://api.github.com/repos/openclaw/openclaw/issues/70871 | Memory plugin auto-recall embeds full conversation context, causing context length errors and timeouts | ## Summary
Auto-recall for memory plugins (specifically `openclaw-mem0` in open-source mode) fails on conversations with large context because the full conversation is passed to the embedding model instead of a focused query.
## Environment
- **OpenClaw:** 2026.4.15
- **Plugin:** openclaw-mem0 1.0.10
- **Mode:** ope... | closed | not_planned | true | 1 | [] | [] | 2026-04-24T01:33:38Z | 2026-04-29T04:37:40Z | 2026-04-26T10:22:37Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | mcderms2-droid | 229,391,140 | U_kgDODaw7JA | User | false |
openclaw/openclaw | 4,319,899,964 | I_kwDOQb6kR88AAAABAXxxPA | 70,872 | https://github.com/openclaw/openclaw/issues/70872 | https://api.github.com/repos/openclaw/openclaw/issues/70872 | [Bug]: Synology MCP tool calls fail with "params: must be object" — OpenClaw passes params as string | ## Summary
When calling Synology MCP tools through OpenClaw's native MCP adapter, all tool calls fail with a validation error:
```
Validation failed for tool "synology__synology_test_connection":
- params: must be object
Received arguments:
{
"params": "{}"
}
```
The same tools work correctly when called throug... | closed | not_planned | true | 1 | [] | [] | 2026-04-24T01:33:44Z | 2026-04-29T04:34:29Z | 2026-04-26T18:58:06Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | slideshow-dingo | 258,996,445 | U_kgDOD2_43Q | User | false |
openclaw/openclaw | 4,319,926,438 | I_kwDOQb6kR88AAAABAXzYpg | 70,874 | https://github.com/openclaw/openclaw/issues/70874 | https://api.github.com/repos/openclaw/openclaw/issues/70874 | [Bug]: /tools does not list mcp.servers tools even though the agent can see and call them | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
In a normal OpenClaw session, tools from `mcp.servers` are usable by the agent, but `/tools` and `/tools verbose` do not list them.
This makes it look like MCP tools are not injected, even though they actually are available... | closed | not_planned | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-24T01:40:02Z | 2026-04-29T04:36:55Z | 2026-04-26T10:53:40Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | bighitbiker3 | 15,604,901 | MDQ6VXNlcjE1NjA0OTAx | User | false |
openclaw/openclaw | 4,319,928,172 | I_kwDOQb6kR88AAAABAXzfbA | 70,875 | https://github.com/openclaw/openclaw/issues/70875 | https://api.github.com/repos/openclaw/openclaw/issues/70875 | lmstudio plugin fails to load: Cannot find module './providers/faux.js' | ## Description
The built-in `lmstudio` extension fails to load on gateway startup with:
```
[plugins] lmstudio failed to load from /usr/lib/node_modules/openclaw/dist/extensions/lmstudio/index.js:
Error: Cannot find module './providers/faux.js'
Require stack:
- /usr/lib/node_modules/openclaw/dist/extensions/lmstudio/... | closed | completed | true | 1 | [] | [] | 2026-04-24T01:40:29Z | 2026-04-28T04:47:15Z | 2026-04-25T19:28:43Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | aspalagin | 263,096,044 | U_kgDOD66G7A | User | false |
openclaw/openclaw | 4,066,225,308 | I_kwDOQb6kR87yXayc | 44,276 | https://github.com/openclaw/openclaw/issues/44276 | https://api.github.com/repos/openclaw/openclaw/issues/44276 | [Bug] imageModel not invoked when primary model lacks vision capability | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
I configured agents.defaults.imageModel to use a vision-capable model (qwen3.5-plus) because my primary model (GLM-5) does not support image input. However, when sending images in Control UI, the imageModel is never invoked - the image is s... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-03-12T17:39:48Z | 2026-04-26T04:29:17Z | 2026-04-24T01:40:48Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | nsplnpbjy | 22,773,962 | MDQ6VXNlcjIyNzczOTYy | User | false |
openclaw/openclaw | 4,319,985,031 | I_kwDOQb6kR88AAAABAX29hw | 70,880 | https://github.com/openclaw/openclaw/issues/70880 | https://api.github.com/repos/openclaw/openclaw/issues/70880 | [Feature]: Improve MCP server tool discoverability and search precision for email agents | ## Summary
MCP server tools have generic descriptions that make it hard for LLM agents to select the right tool and use it effectively. Key issues: duplicate `search_emails` vs `search_all_emails`, no way to filter by account/folder/date, and raw verbose JSON responses that overwhelm context windows.
## Problem to so... | closed | not_planned | true | 1 | [] | [] | 2026-04-24T01:55:20Z | 2026-04-28T04:47:14Z | 2026-04-25T19:28:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lfbento | 36,018,041 | MDQ6VXNlcjM2MDE4MDQx | User | false |
openclaw/openclaw | 4,319,990,438 | I_kwDOQb6kR88AAAABAX3Spg | 70,881 | https://github.com/openclaw/openclaw/issues/70881 | https://api.github.com/repos/openclaw/openclaw/issues/70881 | Feature Request: Add configurable stopwords/filter for REM phase theme extraction | ## Problem Description
The REM phase in dreaming extracts themes from session transcripts, but it currently includes structural role markers like `assistant`, `user`, and `system` as "themes" because they appear frequently in the conversation corpus.
**Example:**
```
• 反思主题：assistant 在 1171 条记忆中持续出现（confidence: 1.00）... | closed | completed | true | 1 | [] | [] | 2026-04-24T01:56:57Z | 2026-04-26T04:29:13Z | 2026-04-24T02:01:46Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | chouti | 27,688 | MDQ6VXNlcjI3Njg4 | User | false |
openclaw/openclaw | 4,319,982,772 | I_kwDOQb6kR88AAAABAX20tA | 70,879 | https://github.com/openclaw/openclaw/issues/70879 | https://api.github.com/repos/openclaw/openclaw/issues/70879 | Feishu plugin lacks retry logic for API rate limit errors | Feishu's custom bot has a per-minute API call rate limit (100 requests/min). When the sending frequency approaches this limit, Feishu server returns error codes like 2200 (internal error) or 11232 (flow control).
Currently, OpenClaw's Feishu plugin treats these errors as regular exceptions — it logs them and continues... | open | null | false | 1 | [] | [] | 2026-04-24T01:54:43Z | 2026-05-02T11:16:45Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | AxelHu | 12,809,280 | MDQ6VXNlcjEyODA5Mjgw | User | false |
openclaw/openclaw | 3,930,201,103 | I_kwDOQb6kR87qQhwP | 14,494 | https://github.com/openclaw/openclaw/issues/14494 | https://api.github.com/repos/openclaw/openclaw/issues/14494 | [Feature Request] Per-agent temperature and maxTokens in agents.list[].model | ## Summary
Allow `temperature` and `maxTokens` to be configured per-agent in `agents.list[].model`, alongside the existing `primary` and `fallbacks` fields.
Currently, the `agents.list[].model` schema only accepts:
```json
{
"primary": "provider/model-id",
"fallbacks": ["provider/fallback-model"]
}
```
Adding `t... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-12T07:03:10Z | 2026-04-30T04:51:05Z | 2026-04-25T05:37:44Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | jbovard2016 | 16,016,292 | MDQ6VXNlcjE2MDE2Mjky | User | false |
openclaw/openclaw | 4,068,285,307 | I_kwDOQb6kR87yfRt7 | 44,528 | https://github.com/openclaw/openclaw/issues/44528 | https://api.github.com/repos/openclaw/openclaw/issues/44528 | [Control UI] Internal task completion event incorrectly delivered to chat | ## Description
When using the Control UI (built-in WebSocket client) and spawning sub-agents via `sessions_spawn`, the internal task completion event is being incorrectly delivered to the Control UI chat interface, exposing internal runtime details that should not be visible to end users.
## Environment
- OpenClaw: ... | closed | completed | false | 3 | [] | [] | 2026-03-13T01:53:37Z | 2026-04-25T08:18:45Z | 2026-04-25T08:18:45Z | NONE | null | 20260425T090015Z | 2026-04-25T09:00:15Z | Simple6K | 51,316,109 | MDQ6VXNlcjUxMzE2MTA5 | User | false |
openclaw/openclaw | 4,320,085,495 | I_kwDOQb6kR88AAAABAX9F9w | 70,885 | https://github.com/openclaw/openclaw/issues/70885 | https://api.github.com/repos/openclaw/openclaw/issues/70885 | [Bug]: openai/gpt-image-2 via Codex OAuth fails with missing image scope | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
OpenAI image generation with `openai/gpt-image-2` fails through Codex OAuth because the OAuth token lacks `api.model.images.request`, and the shipped OAuth client is rejected when that scope is requested.
###... | closed | completed | true | 5 | [
"bug",
"bug:behavior"
] | [] | 2026-04-24T02:24:15Z | 2026-04-27T05:00:42Z | 2026-04-24T19:28:36Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | lanciter | 159,670,295 | U_kgDOCYRgFw | User | false |
openclaw/openclaw | 4,320,091,408 | I_kwDOQb6kR88AAAABAX9dEA | 70,887 | https://github.com/openclaw/openclaw/issues/70887 | https://api.github.com/repos/openclaw/openclaw/issues/70887 | tasks audit: 700+ inconsistent_timestamps warnings on cron tasks (startedAt < createdAt by 1ms) | ## Summary
After upgrading to **2026.4.22 (00bd2cf)** and running `openclaw tasks audit`, 710 of 711 findings are `inconsistent_timestamps` warnings on cron-runtime tasks where `startedAt` is recorded as 1–4ms **earlier** than `createdAt`.
## Environment
- openclaw `2026.4.22 (00bd2cf)`
- Linux x86_64, Node 22.22.2
... | closed | completed | true | 1 | [] | [] | 2026-04-24T02:26:19Z | 2026-04-29T04:41:49Z | 2026-04-26T07:26:49Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ketwi15 | 20,687,695 | MDQ6VXNlcjIwNjg3Njk1 | User | false |
openclaw/openclaw | 4,319,612,826 | I_kwDOQb6kR88AAAABAXgPmg | 70,844 | https://github.com/openclaw/openclaw/issues/70844 | https://api.github.com/repos/openclaw/openclaw/issues/70844 | [Bug]: v2026.4.22 official Docker image — 18 bundled plugins fail runtime-dep install with EUNSUPPORTEDPROTOCOL: "workspace:*" | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
On `ghcr.io/openclaw/openclaw:2026.4.22`, bundled plugins that install runtime deps at gateway startup fail with `npm error code EUNSUPPORTEDPROTOCOL — Unsupported URL Type "workspace:": workspace:*`; 18 plugins failed in ou... | closed | completed | true | 2 | [] | [] | 2026-04-24T00:13:32Z | 2026-04-26T04:29:11Z | 2026-04-24T02:27:05Z | CONTRIBUTOR | null | 20260426T060034Z | 2026-04-26T06:00:34Z | simonemacario | 2,116,609 | MDQ6VXNlcjIxMTY2MDk= | User | false |
openclaw/openclaw | 4,320,087,895 | I_kwDOQb6kR88AAAABAX9PVw | 70,886 | https://github.com/openclaw/openclaw/issues/70886 | https://api.github.com/repos/openclaw/openclaw/issues/70886 | Image tool returns 404 for ZAI vision models (zai/glm-4.6v) | ## Bug
The built-in `image` tool returns **404 "Not Found"** when `agents.defaults.imageModel` is set to a ZAI vision model (e.g. `zai/glm-4.6v`).
The model works correctly via direct API calls (curl/SDK), so this is an internal routing issue within OpenClaw's image tool pipeline.
## Reproduction
1. Set `agents.def... | open | null | false | 2 | [] | [] | 2026-04-24T02:25:04Z | 2026-05-02T12:13:14Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | hebridean-tech | 248,799,456 | U_kgDODtRg4A | User | false |
openclaw/openclaw | 4,320,137,241 | I_kwDOQb6kR88AAAABAYAQGQ | 70,889 | https://github.com/openclaw/openclaw/issues/70889 | https://api.github.com/repos/openclaw/openclaw/issues/70889 | agents delete --force --json emits non-JSON 'Config overwrite:' prefix line to stdout | # `openclaw agents delete --force --json` emits non-JSON prefix line to stdout
**Version affected:** OpenClaw 2026.4.14 (323493f) on macOS
## Summary
`openclaw agents delete <id> --force --json` is documented (and commonly used) as emitting a machine-readable JSON summary to stdout. In practice, the command emits a ... | closed | completed | true | 0 | [] | [] | 2026-04-24T02:39:44Z | 2026-04-27T04:42:29Z | 2026-04-25T01:22:07Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | newestway | 965,626 | MDQ6VXNlcjk2NTYyNg== | User | false |
openclaw/openclaw | 4,320,137,555 | I_kwDOQb6kR88AAAABAYARUw | 70,890 | https://github.com/openclaw/openclaw/issues/70890 | https://api.github.com/repos/openclaw/openclaw/issues/70890 | agents delete recursively removes workspace dir — silent data-loss when workspace is shared | # `openclaw agents delete` recursively removes the workspace directory, including shared/template state
**Version affected:** OpenClaw 2026.4.14 (323493f) on macOS
**Severity:** data-loss (silent, unbounded blast radius when workspaces are shared across agents)
## Summary
`openclaw agents delete <id> --force` remove... | closed | completed | true | 0 | [] | [] | 2026-04-24T02:39:51Z | 2026-04-27T04:42:27Z | 2026-04-25T01:22:08Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | newestway | 965,626 | MDQ6VXNlcjk2NTYyNg== | User | false |
openclaw/openclaw | 4,320,148,850 | I_kwDOQb6kR88AAAABAYA9cg | 70,892 | https://github.com/openclaw/openclaw/issues/70892 | https://api.github.com/repos/openclaw/openclaw/issues/70892 | Per-section gates for moot L1 prompt blocks (A6/A13/A14) + available_skills filter | ## Context
Running a single-purpose trading pod on `ghcr.io/openclaw/openclaw:2026.4.9`. The base-image system prompt is ~35k chars of static preamble injected on every cold-start invocation. On a concrete measurement (captured via cllama's `/internal/context/<agent>/snapshot`), four sections contribute chars that our... | open | null | false | 1 | [] | [] | 2026-04-24T02:43:06Z | 2026-05-02T11:16:47Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | mostlydev | 67,087 | MDQ6VXNlcjY3MDg3 | User | false |
openclaw/openclaw | 4,320,167,851 | I_kwDOQb6kR88AAAABAYCHqw | 70,894 | https://github.com/openclaw/openclaw/issues/70894 | https://api.github.com/repos/openclaw/openclaw/issues/70894 | [Bug]: Regression: Webhook agent runs always start a new session regardless of sessionKey / session config | ### Summary
[Original issue](https://github.com/openclaw/openclaw/issues/29518) was closed without fix
Webhook-triggered agent runs always start a new session and lose conversation history/context, even when providing the same stable sessionKey and allowing it in config.
### Steps to reproduce
1. Configure gateway... | open | null | false | 1 | [] | [] | 2026-04-24T02:48:34Z | 2026-05-02T14:00:29Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | dannylin108 | 72,641,467 | MDQ6VXNlcjcyNjQxNDY3 | User | false |
openclaw/openclaw | 4,320,169,947 | I_kwDOQb6kR88AAAABAYCP2w | 70,896 | https://github.com/openclaw/openclaw/issues/70896 | https://api.github.com/repos/openclaw/openclaw/issues/70896 | [Feature]: Log files are created but never used for writing (no effective log rotation) | ### Summary
Log files can be created but logger does not switch write target, lacking effective log rotation
### Problem to solve
Currently, the gateway may generate new log files (e.g., with date-based filenames), but the logger continues writing to the originally opened file.
Even when a new log file exists (such... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-04-24T02:49:03Z | 2026-04-29T04:41:48Z | 2026-04-26T07:27:00Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Narlea | 163,425,703 | U_kgDOCb2tpw | User | false |
openclaw/openclaw | 4,320,240,021 | I_kwDOQb6kR88AAAABAYGhlQ | 70,901 | https://github.com/openclaw/openclaw/issues/70901 | https://api.github.com/repos/openclaw/openclaw/issues/70901 | MCP HTTP server headers don't expand ${ENV_VAR} syntax, forcing tokens inline in config | ## Summary
When configuring an HTTP MCP server (e.g., Supabase MCP), `${ENV_VAR}` substitution in the `headers` block is not expanded, forcing users to put tokens inline in `openclaw.json`.
## Environment
- **OpenClaw:** 2026.4.22
- **OS:** Windows 11
- **Use case:** Supabase MCP (`https://mcp.supabase.com/mcp`)
##... | closed | completed | true | 2 | [] | [] | 2026-04-24T03:09:00Z | 2026-04-29T04:36:01Z | 2026-04-26T12:21:25Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | mcderms2-droid | 229,391,140 | U_kgDODaw7JA | User | false |
openclaw/openclaw | 4,320,244,809 | I_kwDOQb6kR88AAAABAYG0SQ | 70,903 | https://github.com/openclaw/openclaw/issues/70903 | https://api.github.com/repos/openclaw/openclaw/issues/70903 | Persistent file-based provider cooldown blocks user for hours after billing recovery | ## Summary
When an Anthropic (or any) provider returns a 402 billing error, OpenClaw writes a `disabledUntil` timestamp to the agent auth-state file that persists across gateway restarts. On repeated failures the timestamp extends forward, which means even after the user tops up credit, they remain locked out for hour... | open | null | false | 2 | [] | [] | 2026-04-24T03:10:41Z | 2026-05-02T14:00:31Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | mattglover11 | 8,943,960 | MDQ6VXNlcjg5NDM5NjA= | User | false |
openclaw/openclaw | 4,320,246,025 | I_kwDOQb6kR88AAAABAYG5CQ | 70,904 | https://github.com/openclaw/openclaw/issues/70904 | https://api.github.com/repos/openclaw/openclaw/issues/70904 | [Bug]: thinkingLevel is silently cosmetic for openai/openai-codex Responses — createOpenAIThinkingLevelWrapper misses the existingReasoning === undefined branch | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
`createOpenAIThinkingLevelWrapper` in `proxy-stream-wrappers.ts` never injects `body.reasoning` when pi-ai leaves `payloadObj.reasoning` as `undefined`, so `thinkingDefault` / `thinkingLevel` is silently cosmet... | closed | not_planned | true | 2 | [] | [] | 2026-04-24T03:11:10Z | 2026-04-27T05:01:36Z | 2026-04-24T19:34:33Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | douglferreira977 | 271,555,325 | U_kgDOEC-a_Q | User | false |
openclaw/openclaw | 4,320,246,127 | I_kwDOQb6kR88AAAABAYG5bw | 70,905 | https://github.com/openclaw/openclaw/issues/70905 | https://api.github.com/repos/openclaw/openclaw/issues/70905 | [Feature]: Expose agent id in ChannelOutboundContext | ### Summary
Add an optional agentId field to ChannelOutboundContext so channel plugins can attribute outbound sends to a specific agent.
### Problem to solve
Channel plugins that model each agent as a distinct remote identity (one agent per user/member row on the remote side, rather
than a single shared bot wit... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-24T03:11:12Z | 2026-05-02T14:00:32Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | enzobjmendonca | 204,745,656 | U_kgDODDQruA | User | false |
openclaw/openclaw | 4,320,264,229 | I_kwDOQb6kR88AAAABAYIAJQ | 70,909 | https://github.com/openclaw/openclaw/issues/70909 | https://api.github.com/repos/openclaw/openclaw/issues/70909 | Claude CLI runner ignores user-configured `mcp.servers` from openclaw.json | ## Summary
Claude CLI runner silently drops user-registered `mcp.servers` from `openclaw.json`. Servers registered via `openclaw mcp set` appear in `openclaw mcp list`, and Pi-embedded runs pick them up, but Claude CLI sessions never see their tools — the generated `/tmp/openclaw-cli-mcp-*/mcp.json` only contains bund... | closed | completed | true | 0 | [] | [] | 2026-04-24T03:17:22Z | 2026-04-27T04:44:13Z | 2026-04-25T00:56:24Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | gm4xgvkx7g-droid | 277,947,168 | U_kgDOEJEjIA | User | false |
openclaw/openclaw | 4,320,280,642 | I_kwDOQb6kR88AAAABAYJAQg | 70,914 | https://github.com/openclaw/openclaw/issues/70914 | https://api.github.com/repos/openclaw/openclaw/issues/70914 | Bundler: unhashed runtime-model-auth.runtime.js stub missing from dist/, breaks image.generate + any capability using runtime model auth | ## Summary
Builds from current `main` (as of 2026-04-23, tip `f1ad5e27e0`) no longer emit the unhashed `dist/runtime-model-auth.runtime.{js,ts,mjs,...}` stub that `resolveRuntimeModelAuthModuleHref()` in `dist/provider-auth-runtime-*.js` walks a candidate list to locate. Only hashed chunks exist, so the resolver exhau... | closed | completed | true | 1 | [] | [] | 2026-04-24T03:21:56Z | 2026-04-26T04:29:07Z | 2026-04-24T03:41:10Z | CONTRIBUTOR | null | 20260426T060034Z | 2026-04-26T06:00:34Z | lesaai | 260,982,214 | U_kgDOD45Fxg | User | false |
openclaw/openclaw | 4,224,865,166 | I_kwDOQb6kR8770lOO | 63,173 | https://github.com/openclaw/openclaw/issues/63173 | https://api.github.com/repos/openclaw/openclaw/issues/63173 | [Bug]: Kimi Code (Moonshot) Provider enters an infinite loop during tasks | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
When using kimi-code (Kimi API) as the LLM provider in OpenClaw, the agent frequently gets stuck in an infinite loop. It repeatedly generates the same response text followed by a tool call with empty arguments (e.g., read: ... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-08T13:28:01Z | 2026-04-28T05:04:18Z | 2026-04-25T15:31:24Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Jojo0123 | 60,039,088 | MDQ6VXNlcjYwMDM5MDg4 | User | false |
openclaw/openclaw | 4,320,352,716 | I_kwDOQb6kR88AAAABAYNZzA | 70,919 | https://github.com/openclaw/openclaw/issues/70919 | https://api.github.com/repos/openclaw/openclaw/issues/70919 | [Feature]: During the conversation chat, automatically identify the key words and trigger the model switch based on the input key words. | ### Summary
During the conversation chat, automatically identify the key words and trigger the model switch based on the input key words.
### Problem to solve
Dear leader, now openclaw does not support automatic detection to switch models according to keywords in the dialogue chat. I hope the official support is to ... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-24T03:43:19Z | 2026-04-28T04:47:13Z | 2026-04-25T19:29:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | xuyancan | 19,484,293 | MDQ6VXNlcjE5NDg0Mjkz | User | false |
openclaw/openclaw | 4,320,362,931 | I_kwDOQb6kR88AAAABAYOBsw | 70,921 | https://github.com/openclaw/openclaw/issues/70921 | https://api.github.com/repos/openclaw/openclaw/issues/70921 | Block-streaming channels deliver assembled final payload in addition to per-chunk streaming, causing full-reply duplication |
## Summary
When `blockStreaming` is enabled for a channel (Signal, WhatsApp, iMessage/BlueBubbles — default for most IM channels), assistant replies containing structured markdown (tables, lists, code blocks) are delivered **twice** in the same turn:
1. First as per-chunk streamed payloads during LLM generation (`on... | closed | completed | true | 1 | [] | [] | 2026-04-24T03:47:02Z | 2026-04-28T04:47:12Z | 2026-04-25T19:29:09Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | joeykrug | 5,925,937 | MDQ6VXNlcjU5MjU5Mzc= | User | false |
openclaw/openclaw | 4,320,271,144 | I_kwDOQb6kR88AAAABAYIbKA | 70,912 | https://github.com/openclaw/openclaw/issues/70912 | https://api.github.com/repos/openclaw/openclaw/issues/70912 | [Bug]: 2026.4.22 Slack MPIM/group: internal "Working…" tool-trace messages leak into group chats as user-visible messages | ### Summary
In 2026.4.22, the runtime-level "Working… / tool: … / tool args …" progress messages (normally hidden in group chats) are leaking into **Slack group DMs (MPIMs)** as real, user-visible messages. This exposes internal tool names and arguments (including fetched URLs and message/reaction targets) to group pa... | closed | completed | true | 2 | [] | [] | 2026-04-24T03:19:38Z | 2026-04-26T04:28:04Z | 2026-04-24T03:52:31Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | jrex-jooni | 266,595,103 | U_kgDOD-PrHw | User | false |
openclaw/openclaw | 4,320,390,769 | I_kwDOQb6kR88AAAABAYPucQ | 70,927 | https://github.com/openclaw/openclaw/issues/70927 | https://api.github.com/repos/openclaw/openclaw/issues/70927 | qwen3.6-plus model incorrectly filtered out on DashScope Coding Plan (coding.dashscope.aliyuncs.com) | ## Problem
`qwen/qwen3.6-plus` model shows as `missing: true` with null metadata when using the DashScope Coding Plan base URL (`https://coding.dashscope.aliyuncs.com/v1`), even though the model is explicitly configured in `openclaw.json`.
### Model List Output
```json
{
"key": "qwen/qwen3.6-plus",
"name": "qwen... | closed | not_planned | true | 1 | [] | [] | 2026-04-24T03:56:03Z | 2026-04-29T04:41:48Z | 2026-04-26T07:27:05Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | qhenry | 1,560,317 | MDQ6VXNlcjE1NjAzMTc= | User | false |
openclaw/openclaw | 4,320,394,908 | I_kwDOQb6kR88AAAABAYP-nA | 70,928 | https://github.com/openclaw/openclaw/issues/70928 | https://api.github.com/repos/openclaw/openclaw/issues/70928 | [Bug]: `message_sent` / `before_message_write` hooks never fire for webchat and openclaw-tui outbound paths | # [Bug]: `message_sent` / `before_message_write` hooks never fire for webchat and openclaw-tui outbound paths
_Repo_: `openclaw/openclaw`
_Related_: #50126 (umbrella: inconsistent hook coverage across outbound paths), #47472 (plugin hook never fires for extension plugins)
---
## Summary
Outbound message lifecycle h... | open | null | false | 4 | [] | [] | 2026-04-24T03:57:36Z | 2026-05-02T14:00:33Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | ryota-murakami | 5,501,268 | MDQ6VXNlcjU1MDEyNjg= | User | false |
openclaw/openclaw | 4,320,390,304 | I_kwDOQb6kR88AAAABAYPsoA | 70,926 | https://github.com/openclaw/openclaw/issues/70926 | https://api.github.com/repos/openclaw/openclaw/issues/70926 | qwen3.6-plus model shows as 'missing' when using DashScope Coding Plan base URL | ## Problem
model shows as with null metadata when using the DashScope Coding Plan base URL (), even though the model is explicitly configured in .
### Model List Output
### User Configuration
## Root Cause
In the OpenClaw source code (), there is filtering logic:
This code **explicitly filters out ** when... | closed | duplicate | true | 0 | [] | [] | 2026-04-24T03:55:53Z | 2026-04-26T04:29:06Z | 2026-04-24T03:57:45Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | qhenry | 1,560,317 | MDQ6VXNlcjE1NjAzMTc= | User | false |
openclaw/openclaw | 4,320,412,646 | I_kwDOQb6kR88AAAABAYRD5g | 70,929 | https://github.com/openclaw/openclaw/issues/70929 | https://api.github.com/repos/openclaw/openclaw/issues/70929 | [Feature]: Add eNSP to kc.exe application whitelist | ### Summary
Add Huawei eNSP (Enterprise Network Simulation Platform) to kc.exe application whitelist so OpenClaw can control eNSP GUI devices.
### Problem to solve
The kc-gui skill's SKILL.md declares eNSP as supported, but eNSP_Client.exe and eNSP_VBoxServer.exe are not in the kc.exe binary whitelist. The whitelis... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-24T04:03:11Z | 2026-04-28T04:47:11Z | 2026-04-25T19:29:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | cyj-0077 | 203,324,320 | U_kgDODB57oA | User | false |
openclaw/openclaw | 4,320,274,139 | I_kwDOQb6kR88AAAABAYIm2w | 70,913 | https://github.com/openclaw/openclaw/issues/70913 | https://api.github.com/repos/openclaw/openclaw/issues/70913 | Windows: spawn codex ENOENT - codex/* models unusable due to raw spawn() not resolving PATHEXT | ## Summary
On Windows, the gateway fails to spawn the `codex` CLI when using a `codex/*` model (e.g. `codex/gpt-5.5`). Every request errors with `spawn codex ENOENT` and falls back to another model, even though `codex` is installed globally via `npm i -g @openai/codex` and is resolvable from the same shell that star... | closed | completed | true | 2 | [] | [] | 2026-04-24T03:20:20Z | 2026-04-26T04:29:06Z | 2026-04-24T04:05:02Z | NONE | null | 20260426T060034Z | 2026-04-26T06:00:34Z | ade5791 | 38,266,156 | MDQ6VXNlcjM4MjY2MTU2 | User | false |
openclaw/openclaw | 4,320,455,205 | I_kwDOQb6kR88AAAABAYTqJQ | 70,933 | https://github.com/openclaw/openclaw/issues/70933 | https://api.github.com/repos/openclaw/openclaw/issues/70933 | [Feature]: New Session 创建时增加确认提示 | ### Summary
New Session 创建时增加确认提示
### Problem to solve
点击 "New Session" 时没有任何确认机制，误点击会导致：
- 当前会话上下文丢失
- 未保存的工作进度丢失
- 需要重新解释任务背景
### Proposed solution
点击 New Session 后，弹出确认对话框：
- 提示："确定要创建新会话吗？"
- 说明影响："当前聊天内容将消失，之前未完成的任务需要重新开始"
- 提供"确认创建"和"取消"两个按钮
- 可选：增加"不再提示"的checkbox
### Alternatives considered
_No response_
... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-24T04:15:44Z | 2026-04-29T04:41:47Z | 2026-04-26T07:27:10Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | VossmanLeung | 182,184,462 | U_kgDOCtvqDg | User | false |
openclaw/openclaw | 4,320,473,406 | I_kwDOQb6kR88AAAABAYUxPg | 70,934 | https://github.com/openclaw/openclaw/issues/70934 | https://api.github.com/repos/openclaw/openclaw/issues/70934 | [Feature]: Show value before customization during first-run onboarding | ### Summary
Reduce first-time user friction by delaying non-essential personalization steps until after users see an initial useful result.
### Problem to solve
During first launch, users are asked to configure items such as agent name, reply style, symbols, or persona preferences before experiencing what OpenClaw c... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-24T04:21:12Z | 2026-05-02T14:00:34Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | johnsonbuilds | 48,879,766 | MDQ6VXNlcjQ4ODc5NzY2 | User | false |
openclaw/openclaw | 4,320,474,419 | I_kwDOQb6kR88AAAABAYU1Mw | 70,935 | https://github.com/openclaw/openclaw/issues/70935 | https://api.github.com/repos/openclaw/openclaw/issues/70935 | [Bug]: edit tool fails with "Expected ',' or ']'" error when edits array contains multi-line oldText/newText | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
edit Tool JSON Parameter Encoding Breaks with Multi-Line Text in Nested Arrays
### Steps to reproduce
1. Call the edit tool with an edits array containing multi-line text:
{
"path": "path/to/file.md",
"edits": [
{
... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-24T04:21:35Z | 2026-04-28T04:47:11Z | 2026-04-25T19:29:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | calistasherpa-png | 270,209,404 | U_kgDOEBsRfA | User | false |
openclaw/openclaw | 4,320,494,376 | I_kwDOQb6kR88AAAABAYWDKA | 70,939 | https://github.com/openclaw/openclaw/issues/70939 | https://api.github.com/repos/openclaw/openclaw/issues/70939 | Bug: scheduled cron agentTurn appends bundled MCP tools after toolsAllow filtering, wasting ~370k input tokens/day | ## Summary
Scheduled cron `agentTurn` runs do pass `toolsAllow`, but the allowlist is only applied to the base OpenClaw tool registry. Bundled MCP/LSP tools are materialized later and appended after policy filtering without reapplying `toolsAllow`, which allows an exec-only cron to compile with `exec + 45 Atlassian to... | closed | completed | true | 1 | [] | [] | 2026-04-24T04:28:33Z | 2026-04-28T04:47:10Z | 2026-04-25T19:29:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | scotcha1 | 4,205,933 | MDQ6VXNlcjQyMDU5MzM= | User | false |
openclaw/openclaw | 4,040,400,440 | I_kwDOQb6kR87w0544 | 39,439 | https://github.com/openclaw/openclaw/issues/39439 | https://api.github.com/repos/openclaw/openclaw/issues/39439 | DM replies fail with "Action send requires a target" - bot shows typing but never sends response | ## Bug Report
**Version:** 2026.3.2
**Description:**
When a user sends a DM to the bot, the bot starts typing but never sends a reply.
The agent processes the message but fails when trying to send the response.
**Error from logs:**
```
[tools] message failed: Action send requires a target.
```
The agent session is ... | closed | completed | true | 4 | [
"stale"
] | [] | 2026-03-08T04:05:08Z | 2026-04-27T04:59:49Z | 2026-04-24T20:31:13Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | milkr | 32,992 | MDQ6VXNlcjMyOTky | User | false |
openclaw/openclaw | 4,040,439,879 | I_kwDOQb6kR87w1DhH | 39,447 | https://github.com/openclaw/openclaw/issues/39447 | https://api.github.com/repos/openclaw/openclaw/issues/39447 | [Bug]: npm install failed for openclaw@latest | ### Bug type
Crash (process/app exits or hangs)
### Summary
Tried and failed to install open claw on a fresh ubuntu 24.04 from digital ocean
### Steps to reproduce
Install openclaw with 1 liner on ubuntu droplet from digitalocean
### Expected behavior
Installs correctly
### Actual behavior
Fails to install
##... | closed | completed | true | 12 | [
"bug",
"stale",
"bug:crash"
] | [
"vincentkoc"
] | 2026-03-08T04:22:09Z | 2026-04-30T04:41:39Z | 2026-04-26T04:54:02Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | jonyuhu-ux | 266,420,699 | U_kgDOD-FB2w | User | false |
openclaw/openclaw | 4,040,444,403 | I_kwDOQb6kR87w1Enz | 39,450 | https://github.com/openclaw/openclaw/issues/39450 | https://api.github.com/repos/openclaw/openclaw/issues/39450 | Control UI: main:local session disappears from dropdown after switching to main:main and refreshing | ### Environment
- OpenClaw Control UI (webchat)
- Channel: webchat
### Steps to reproduce
1. Open `main:local` session
2. Switch to `main:main`
3. Refresh page
4. Open session dropdown
### Expected behavior
`main:local` should still appear in dropdown.
### Actual behavior
`main:local` disappears from dropdown after ... | closed | completed | true | 3 | [
"stale"
] | [] | 2026-03-08T04:25:36Z | 2026-04-28T04:43:09Z | 2026-04-26T02:00:55Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Vigoose | 5,359,336 | MDQ6VXNlcjUzNTkzMzY= | User | false |
openclaw/openclaw | 4,040,445,867 | I_kwDOQb6kR87w1E-r | 39,454 | https://github.com/openclaw/openclaw/issues/39454 | https://api.github.com/repos/openclaw/openclaw/issues/39454 | fix: Anthropic SDK crashes when proxy/relay retransmits SSE stream — patch & workaround | ## Summary
When using OpenClaw with a third-party API proxy/relay (e.g., nexus-style relay), the bot intermittently fails to respond on complex queries (tool calls, web search). The root cause is in `@anthropic-ai/sdk@0.73.0`'s `MessageStream._accumulateMessage()`, which throws a fatal error when the relay retransmits... | closed | not_planned | true | 4 | [
"stale"
] | [] | 2026-03-08T04:27:00Z | 2026-04-27T04:59:49Z | 2026-04-24T20:31:16Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | zhoumo-creator | 113,785,378 | U_kgDOBsg6Ig | User | false |
openclaw/openclaw | 4,040,455,692 | I_kwDOQb6kR87w1HYM | 39,461 | https://github.com/openclaw/openclaw/issues/39461 | https://api.github.com/repos/openclaw/openclaw/issues/39461 | [Bug]：路径错误 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
## Bug: QQBot 插件未读取 QQBOT_DATA_DIR 环境变量
### 问题描述
设置 `QQBOT_DATA_DIR` 环境变量后，QQBot 插件仍使用中文路径 `C:\Users\张家豪\.openclaw\qqbot`，导致警告提示。
### 环境
- 操作系统: Windows 10
- OpenClaw 版本: 2026.3.2
- Node 版本: v24.14.0
### 复现步骤
1. 设置环境变量 `QQBOT_DATA_DIR=D... | open | null | false | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-03-08T04:34:01Z | 2026-05-04T07:01:21Z | null | NONE | null | 20260504T090105Z | 2026-05-04T09:01:05Z | jiahao-0111 | 240,259,976 | U_kgDODlITiA | User | false |
openclaw/openclaw | 4,040,467,435 | I_kwDOQb6kR87w1KPr | 39,466 | https://github.com/openclaw/openclaw/issues/39466 | https://api.github.com/repos/openclaw/openclaw/issues/39466 | [Bug] Webchat: Assistant text vanishes when tool calls follow text output in same response | ## Bug Description
When the assistant sends a response containing both text output and tool calls, **the visible text disappears from the webchat Control UI** if tool calls are made **after** the text content in the same response turn.
The text remains visible in the TUI interface — this is webchat-specific.
## Step... | closed | completed | true | 3 | [
"stale"
] | [] | 2026-03-08T04:40:50Z | 2026-04-27T04:59:47Z | 2026-04-24T20:31:24Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | navendugoyal19 | 132,033,603 | U_kgDOB96sQw | User | false |
openclaw/openclaw | 4,040,470,389 | I_kwDOQb6kR87w1K91 | 39,468 | https://github.com/openclaw/openclaw/issues/39468 | https://api.github.com/repos/openclaw/openclaw/issues/39468 | [Bug]: Gateway crashes when receiving group messages — auto-writes invalid 'agents.subagents' config field (v2026.3.2) | ## Bug Description
OpenClaw v2026.3.2 has a critical bug: when the gateway receives group messages from Feishu (Lark), it automatically writes an invalid `agents.subagents` field to `openclaw.json`, causing config validation to fail and the gateway to crash.
## Steps to Reproduce
1. Configure Feishu channel with mul... | closed | completed | true | 3 | [
"stale"
] | [] | 2026-03-08T04:42:44Z | 2026-04-27T04:59:46Z | 2026-04-24T20:31:27Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | fengshao312 | 36,143,264 | MDQ6VXNlcjM2MTQzMjY0 | User | false |
openclaw/openclaw | 4,040,476,646 | I_kwDOQb6kR87w1Mfm | 39,473 | https://github.com/openclaw/openclaw/issues/39473 | https://api.github.com/repos/openclaw/openclaw/issues/39473 | [Bug]: Web control UI leaks internal control tokens into visible chat messages | ### Bug type
Regression (worked before, now fails)
### Summary
In the OpenClaw web control UI, internal control outputs such as NO_REPLY are sometimes rendered as normal visible assistant messages.
### Steps to reproduce
1. Use the web-based openclaw-control-ui chat surface.
2. Trigger maintenance/system turns suc... | closed | completed | true | 4 | [
"bug",
"stale",
"regression"
] | [] | 2026-03-08T04:45:13Z | 2026-04-27T04:59:45Z | 2026-04-24T20:31:34Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | men357475-code | 245,711,358 | U_kgDODqVB_g | User | false |
openclaw/openclaw | 4,040,480,903 | I_kwDOQb6kR87w1NiH | 39,476 | https://github.com/openclaw/openclaw/issues/39476 | https://api.github.com/repos/openclaw/openclaw/issues/39476 | A2A sessions_send: target agent can call sessions_send back, causing duplicate messages | ## Summary
When Agent A calls `sessions_send` to Agent B, Agent B can use `sessions_send` to call **back** to Agent A with its response. This causes duplicate messages in the requester's channel:
1. The original `sessions_send` tool returns `primaryReply` → Agent A posts it
2. Agent B independently calls `sessions_se... | open | null | false | 8 | [] | [] | 2026-03-08T04:47:13Z | 2026-05-04T07:01:22Z | null | NONE | null | 20260504T090105Z | 2026-05-04T09:01:05Z | SYRS-AI | 240,142,975 | U_kgDODlBKfw | User | false |
openclaw/openclaw | 4,040,490,081 | I_kwDOQb6kR87w1Pxh | 39,484 | https://github.com/openclaw/openclaw/issues/39484 | https://api.github.com/repos/openclaw/openclaw/issues/39484 | memory_search returns 0 hits for specific terms despite indexed content | ## Summary
OpenClaw memory_search API returns 0 hits for specific terms despite content being verified present and indexed.
## Environment
- OpenClaw version: 2026.3.2 (85377a2)
- OS: macOS 26.3.0 (arm64)
- Provider: openai/text-embedding-3-small
- Agent: tex
- Index status: 7/7 files, 45 chunks indexed
## Issue Deta... | closed | completed | true | 3 | [
"stale"
] | [] | 2026-03-08T04:53:35Z | 2026-04-27T04:59:44Z | 2026-04-24T20:31:37Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | joshuakeithpa-sudo | 262,737,445 | U_kgDOD6kOJQ | User | false |
openclaw/openclaw | 4,040,491,344 | I_kwDOQb6kR87w1QFQ | 39,486 | https://github.com/openclaw/openclaw/issues/39486 | https://api.github.com/repos/openclaw/openclaw/issues/39486 | [Bug] Feishu mention-gated pending history can be lost after health-monitor auto-restart | ## Summary
In Feishu group chats with `requireMention=true`, pending pre-mention history can be lost when channel health monitoring triggers an automatic restart.
User-visible symptom: after sending messages `1..6` (without mention) and then `@bot 7`, injected history may be incomplete (e.g., only `5,6`).
## Repro t... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-08T04:54:12Z | 2026-04-27T04:59:44Z | 2026-04-24T20:31:40Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | CyberWenShiFu | 58,394,779 | MDQ6VXNlcjU4Mzk0Nzc5 | User | false |
openclaw/openclaw | 4,040,492,588 | I_kwDOQb6kR87w1QYs | 39,488 | https://github.com/openclaw/openclaw/issues/39488 | https://api.github.com/repos/openclaw/openclaw/issues/39488 | [Bug] Feishu health-check restarts can clear pending mention history context | ## Summary
In Feishu mention-gated group chats (`requireMention=true`), pre-mention history can be missing on the trigger turn after channel auto-restarts.
Validated chain:
health monitor restart -> in-memory pending history cleared -> mention-trigger context injection incomplete
## Validation
- `gateway.channelHe... | closed | not_planned | true | 2 | [
"stale"
] | [] | 2026-03-08T04:54:50Z | 2026-04-27T04:59:43Z | 2026-04-24T20:31:46Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | CyberWenShiFu | 58,394,779 | MDQ6VXNlcjU4Mzk0Nzc5 | User | false |
openclaw/openclaw | 4,040,499,211 | I_kwDOQb6kR87w1SAL | 39,491 | https://github.com/openclaw/openclaw/issues/39491 | https://api.github.com/repos/openclaw/openclaw/issues/39491 | feat(mattermost): chunk text before final streaming edit to respect post size limits | ## Problem
In the Mattermost block-streaming path, when the final deliver call patches the streamed placeholder via updateMattermostPost, the full response text is passed without chunking. This bypasses the chunkMarkdownTextWithMode flow used by the non-streaming path.
If a response exceeds Mattermosts post size limi... | closed | completed | true | 2 | [
"stale"
] | [] | 2026-03-08T04:58:24Z | 2026-04-27T04:54:55Z | 2026-04-24T22:05:50Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | teconomix | 6,959,299 | MDQ6VXNlcjY5NTkyOTk= | User | false |
openclaw/openclaw | 4,040,512,363 | I_kwDOQb6kR87w1VNr | 39,497 | https://github.com/openclaw/openclaw/issues/39497 | https://api.github.com/repos/openclaw/openclaw/issues/39497 | [Bug]: read tool path guard rejects valid /agent/ mount paths in workspaceAccess: "ro" sandboxed sessions | ### Bug type
Crash (process/app exits or hangs)
### Summary
When workspaceAccess is set to "ro", the agent workspace is mounted read-only at /agent/ inside the container. However, the gateway-side wrapSandboxPathGuard in pi-tools.read.ts does not map /agent/ paths back to the host workspace root, causing all read ... | open | null | false | 5 | [
"bug",
"bug:crash"
] | [] | 2026-03-08T05:03:03Z | 2026-05-04T07:01:31Z | null | NONE | null | 20260504T090105Z | 2026-05-04T09:01:05Z | teosborne | 4,041,297 | MDQ6VXNlcjQwNDEyOTc= | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.