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 | 4,340,701,584 | I_kwDOQb6kR88AAAABArnZkA | 73,285 | https://github.com/openclaw/openclaw/issues/73285 | https://api.github.com/repos/openclaw/openclaw/issues/73285 | [Bug] Feishu channel crashes on Windows v2026.4.24: jiti tryNative bypasses pathToFileURL, and mirror dist/node_modules missing openclaw package | ## Bug Report
**Feishu channel (and potentially other bundled channels) crash on Windows v2026.4.24** with two distinct but related issues:
1. `jiti` `tryNative` path bypasses `pathToFileURL()` conversion, passing raw `C:\...` paths to `import()`
2. Mirror `dist/node_modules` directory lacks `openclaw` package symlin... | closed | completed | true | 1 | [] | [] | 2026-04-28T04:58:45Z | 2026-05-01T04:45:26Z | 2026-04-28T05:25:44Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | chris721us-hue | 275,222,342 | U_kgDOEGePRg | User | false |
openclaw/openclaw | 4,340,809,579 | I_kwDOQb6kR88AAAABArt_aw | 73,300 | https://github.com/openclaw/openclaw/issues/73300 | https://api.github.com/repos/openclaw/openclaw/issues/73300 | [Feature]: Allow disabling multiple skills without immediate restart of gateway. | ### Summary
Can we change multiple skill switches and then commit all of them at once with a save button vs immediate restart of gateway with each skill toggle?
### Problem to solve
Can we change multiple skill switches and then commit all of them at once with a save button vs immediate restart of gateway with each ... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-28T05:27:17Z | 2026-05-03T04:36:35Z | 2026-04-30T14:58:15Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | chsbusch-dot | 265,684,486 | U_kgDOD9YGBg | User | false |
openclaw/openclaw | 4,340,727,595 | I_kwDOQb6kR88AAAABAro_Kw | 73,288 | https://github.com/openclaw/openclaw/issues/73288 | https://api.github.com/repos/openclaw/openclaw/issues/73288 | Support OpenRouter provider routing params (only, order, allow, avoid) | ## Feature Request: Pass through OpenRouter provider routing params
**What:** Support OpenRouter's provider routing parameters in model config.
**Why:** OpenRouter's API supports a \`provider\` object in request bodies for intelligent routing:
- \`provider.only\` — restrict requests to specific providers
- \`provider... | closed | completed | true | 1 | [] | [] | 2026-04-28T05:06:14Z | 2026-05-01T04:45:21Z | 2026-04-28T05:27:39Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | astro-15-alive | 273,241,066 | U_kgDOEElT6g | User | false |
openclaw/openclaw | 4,340,728,053 | I_kwDOQb6kR88AAAABArpA9Q | 73,289 | https://github.com/openclaw/openclaw/issues/73289 | https://api.github.com/repos/openclaw/openclaw/issues/73289 | Support OpenRouter provider routing params (only, order, allow, avoid) | ## Feature Request: Pass through OpenRouter provider routing params
**What:** Support OpenRouter's provider routing parameters in model config.
**Why:** OpenRouter's API supports a \`provider\` object in request bodies for intelligent routing:
- \`provider.only\` — restrict requests to specific providers
- \`provider... | closed | completed | true | 1 | [] | [] | 2026-04-28T05:06:22Z | 2026-05-01T04:45:20Z | 2026-04-28T05:27:42Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | astro-15-alive | 273,241,066 | U_kgDOEElT6g | User | false |
openclaw/openclaw | 4,340,652,989 | I_kwDOQb6kR88AAAABArkbvQ | 73,276 | https://github.com/openclaw/openclaw/issues/73276 | https://api.github.com/repos/openclaw/openclaw/issues/73276 | Cold-boot freeze (~3 min) on K8s when only one channel plugin is allowlisted; gateway /readyz returns 504 throughout | ## Summary
On `openclaw 2026.4.26` running in a Kubernetes pod with a single custom channel plugin and a single proxy-style model provider configured (`plugins.allow = ["my-channel", "my-proxy-provider"]`), the gateway takes **~2m42s** between `[gateway] starting channels and sidecars...` and the gateway calling the c... | closed | completed | true | 3 | [] | [] | 2026-04-28T04:46:59Z | 2026-05-01T04:44:31Z | 2026-04-28T08:31:31Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | rudy-ihealth | 232,279,462 | U_kgDODdhNpg | User | false |
openclaw/openclaw | 4,069,176,167 | I_kwDOQb6kR87yirNn | 44,734 | https://github.com/openclaw/openclaw/issues/44734 | https://api.github.com/repos/openclaw/openclaw/issues/44734 | [Bug]: Memory Multimodal Indexing - Gemini API rejects OGG/Opus audio files | ### Bug type
Regression (worked before, now fails)
### Summary
Memory multimodal indexing fails when audio files are OGG/Opus format - Gemini API only supports MP3/WAV
### Steps to reproduce
1. Configure extraPaths in memorySearch to point to a folder with OGG/Opus audio files
2. Run openclaw memory index --force
... | open | null | false | 3 | [
"bug",
"regression"
] | [] | 2026-03-13T06:19:03Z | 2026-05-04T00:11:50Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | olmirr | 30,650,652 | MDQ6VXNlcjMwNjUwNjUy | User | false |
openclaw/openclaw | 4,069,213,349 | I_kwDOQb6kR87yi0Sl | 44,749 | https://github.com/openclaw/openclaw/issues/44749 | https://api.github.com/repos/openclaw/openclaw/issues/44749 | [Bug]: Concurrent allow-always approvals silently lose allowlist entries (last-write-wins race in exec-approvals.json) | # [Bug]: Concurrent `allow-always` approvals silently lose allowlist entries (last-write-wins race in exec-approvals.json)
## Bug type
Data loss / Race condition
## Summary
When multiple exec approval requests are resolved with `allow-always` concurrently, only the last write to `~/.openclaw/exec-approvals.json` pers... | open | null | false | 4 | [] | [] | 2026-03-13T06:29:05Z | 2026-05-04T13:18:47Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | aukei | 26,145,869 | MDQ6VXNlcjI2MTQ1ODY5 | User | false |
openclaw/openclaw | 4,069,230,955 | I_kwDOQb6kR87yi4lr | 44,759 | https://github.com/openclaw/openclaw/issues/44759 | https://api.github.com/repos/openclaw/openclaw/issues/44759 | [Security] Browser screenshot tool auto-delivers sensitive on-screen data to chat channels | ## Summary
When the agent uses `browser screenshot`, the resulting image is **automatically delivered to the conversation channel** (Discord, Telegram, WhatsApp, etc.) via the `MEDIA:` tag embedded in the tool result. This happens at the tool-result processing layer — before the agent writes its reply — and the agent ... | open | null | false | 3 | [] | [] | 2026-03-13T06:33:37Z | 2026-05-04T13:18:48Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | sunshineo | 1,072,766 | MDQ6VXNlcjEwNzI3NjY= | User | false |
openclaw/openclaw | 4,069,351,988 | I_kwDOQb6kR87yjWI0 | 44,789 | https://github.com/openclaw/openclaw/issues/44789 | https://api.github.com/repos/openclaw/openclaw/issues/44789 | [Bug]: openclaw 2026.03.11 partially config litellm | ### Bug type
Regression (worked before, now fails)
### Summary
Installed the lates openclaw 2026.03.11 and trying to `openclaw onboard` to config models with litellm. It automatically skipped litellm url configuration and only have one option in models selection. It worked before but now fails. Please kindly check a... | open | null | false | 3 | [
"bug",
"regression"
] | [] | 2026-03-13T07:05:20Z | 2026-05-04T13:18:50Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | Hylance | 8,049,185 | MDQ6VXNlcjgwNDkxODU= | User | false |
openclaw/openclaw | 4,069,428,847 | I_kwDOQb6kR87yjo5v | 44,810 | https://github.com/openclaw/openclaw/issues/44810 | https://api.github.com/repos/openclaw/openclaw/issues/44810 | Codex ACP run can stall silently when gateway child environment lacks proxy/network access | ## Summary
We debugged a real OpenClaw -> ACP -> Codex specialist flow where `sessions_spawn` succeeded, the Codex child session started, but the run never produced output and eventually surfaced as:
- `codex has produced no output for 60s. It may be waiting for interactive input.`
In our case, the root cause was no... | open | null | false | 4 | [] | [] | 2026-03-13T07:25:25Z | 2026-05-04T13:18:51Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | allan0509 | 5,134,900 | MDQ6VXNlcjUxMzQ5MDA= | User | false |
openclaw/openclaw | 4,069,453,746 | I_kwDOQb6kR87yju-y | 44,820 | https://github.com/openclaw/openclaw/issues/44820 | https://api.github.com/repos/openclaw/openclaw/issues/44820 | Session files not being cleaned up (disk bloat) | OpenClaw is creating session files with .reset.*, .deleted.*, .tmp, and .bak.* extensions but never deleting them. Over time this causes significant disk bloat.
**Steps to reproduce:**
1. Use OpenClaw normally for several days
2. Check ~/.openclaw/agents/main/sessions/ — observe accumulation of stale files
**Expected... | open | null | false | 3 | [] | [] | 2026-03-13T07:31:48Z | 2026-05-04T00:43:14Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | Jragyn | 7,897,461 | MDQ6VXNlcjc4OTc0NjE= | User | false |
openclaw/openclaw | 4,340,963,108 | I_kwDOQb6kR88AAAABAr3XJA | 73,324 | https://github.com/openclaw/openclaw/issues/73324 | https://api.github.com/repos/openclaw/openclaw/issues/73324 | test(cli): logs-cli.test.ts has vitest mock hoisting pattern that mirrors #73177 | ## Summary
`src/cli/logs-cli.test.ts` uses the same pre-existing vitest mock hoisting pattern that caused #73177 (qr-cli.test.ts). Top-level `const` declarations for mock functions are referenced inside `vi.mock()` factory functions. When vitest hoists those factories to the top of the file, the variables are not yet ... | closed | completed | true | 0 | [] | [] | 2026-04-28T06:05:08Z | 2026-05-01T04:45:08Z | 2026-04-28T06:06:33Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | DoncicX | 13,766,184 | MDQ6VXNlcjEzNzY2MTg0 | User | false |
openclaw/openclaw | 4,340,784,926 | I_kwDOQb6kR88AAAABArsfHg | 73,297 | https://github.com/openclaw/openclaw/issues/73297 | https://api.github.com/repos/openclaw/openclaw/issues/73297 | [Bug] Telegram unusable: heartbeat creates isolated session, user loses conversation context | ## Bug Description
When heartbeat fires, it creates an isolated session as expected — but the user's Telegram connection gets routed into the heartbeat session instead of staying in the original conversation session. The user's messages are effectively "hijacked" into the isolated heartbeat session, losing all prior c... | closed | not_planned | true | 2 | [] | [] | 2026-04-28T05:21:04Z | 2026-05-01T04:45:08Z | 2026-04-28T05:50:15Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | ttomiczek | 265,786,127 | U_kgDOD9eTDw | User | false |
openclaw/openclaw | 4,340,983,669 | I_kwDOQb6kR88AAAABAr4ndQ | 73,327 | https://github.com/openclaw/openclaw/issues/73327 | https://api.github.com/repos/openclaw/openclaw/issues/73327 | Cron: per-job configurable stuck-session threshold (currently hardcoded 120s causes false positives on legitimately slow jobs) | ## Summary
The stuck-session detector emits `stuck session: ... age=Ns queueDepth=0` warnings when a session's `state=processing` exceeds **120 seconds**. The threshold appears hardcoded (or at least not user-configurable per-job), which generates false positives for cron jobs that are legitimately long-running by des... | open | null | false | 2 | [] | [] | 2026-04-28T06:09:37Z | 2026-05-03T01:22:53Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | durang | 14,999,651 | MDQ6VXNlcjE0OTk5NjUx | User | false |
openclaw/openclaw | 4,340,994,027 | I_kwDOQb6kR88AAAABAr5P6w | 73,328 | https://github.com/openclaw/openclaw/issues/73328 | https://api.github.com/repos/openclaw/openclaw/issues/73328 | amazon-bedrock discovery: contextWindow defaults to 32000 for Claude 3.x/4.x; need per-model override | ## Summary
`extensions/amazon-bedrock/discovery.ts`'s `toModelDefinition` function applies a single `defaults.contextWindow` to every model returned by `ListFoundationModelsCommand`. The default in current openclaw releases is 32000. Anthropic Claude 3.x and 4.x models have a 200K context window per [Anthropic's publi... | open | null | false | 1 | [] | [] | 2026-04-28T06:12:14Z | 2026-05-03T08:39:50Z | null | CONTRIBUTOR | null | 20260503T090142Z | 2026-05-03T09:01:42Z | markthebest12 | 6,416,874 | MDQ6VXNlcjY0MTY4NzQ= | User | false |
openclaw/openclaw | 4,341,014,164 | I_kwDOQb6kR88AAAABAr6elA | 73,331 | https://github.com/openclaw/openclaw/issues/73331 | https://api.github.com/repos/openclaw/openclaw/issues/73331 | v2026.4.26: Gateway busy-loops on bundled openai SDK directory walk; stops accepting connections | # v2026.4.26: Gateway busy-loops on bundled `openai` SDK directory walk; stops accepting connections
## Summary
After upgrading to `v2026.4.26`, the gateway boots, briefly serves `/health`, and within minutes becomes unresponsive while pegging one CPU core. `strace` shows an inescapable directory walk over `/usr/lib/... | closed | completed | true | 1 | [] | [] | 2026-04-28T06:16:31Z | 2026-05-01T04:45:07Z | 2026-04-28T06:22:04Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | haumanto | 72,009,799 | MDQ6VXNlcjcyMDA5Nzk5 | User | false |
openclaw/openclaw | 4,341,024,005 | I_kwDOQb6kR88AAAABAr7FBQ | 73,333 | https://github.com/openclaw/openclaw/issues/73333 | https://api.github.com/repos/openclaw/openclaw/issues/73333 | [Bug]: Acpx broken after 2025.4.26 update when openclaw home is on remote file systems | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
acpx plugin fails to start when OPENCLAW_STATE_DIR (or equivalent) points at a directory on a remote or special filesystem that rejects chmod (e.g. some NFS/SMB/cloud mounts). Startup fails while preparing the Codex ACP wrap... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-28T06:18:33Z | 2026-05-01T04:45:06Z | 2026-04-28T06:30:28Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | david-garcia-garcia | 9,599,352 | MDQ6VXNlcjk1OTkzNTI= | User | false |
openclaw/openclaw | 4,340,876,746 | I_kwDOQb6kR88AAAABAryFyg | 73,306 | https://github.com/openclaw/openclaw/issues/73306 | https://api.github.com/repos/openclaw/openclaw/issues/73306 | [Bug] Active Memory plugin times out on every run (15s timeout, 0 chars output) - v2026.4.26 | ## Summary
Active Memory plugin consistently times out on every run since upgrading to v2026.4.26 (also occurred on v2026.4.25). The plugin starts, hits the 15s timeout, returns 0 summary chars, and the embedded run fails over with `surface_error`.
Related issues: #68825, #66708, #66157, #65159
## Environment
- Open... | closed | completed | true | 14 | [] | [] | 2026-04-28T05:43:13Z | 2026-05-04T04:33:23Z | 2026-05-01T22:31:44Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | Johannes0402 | 207,033,718 | U_kgDODFcVdg | User | false |
openclaw/openclaw | 4,341,063,728 | I_kwDOQb6kR88AAAABAr9gMA | 73,337 | https://github.com/openclaw/openclaw/issues/73337 | https://api.github.com/repos/openclaw/openclaw/issues/73337 | [Bug] macOS Companion App: Canvas A2UI push returns ok but renders completely blank (regression of #7143) | ## Summary
On macOS Companion App Canvas, pushing A2UI content via `canvas a2ui_push` returns `{ ok: true }` from the gateway, but the Canvas (A2UI) window stays **completely blank** — no rendered content, no "Waiting for A2UI message" placeholder, just the empty gradient background.
Reproducible on **OpenClaw 2026.4... | closed | completed | true | 1 | [] | [
"ngutman"
] | 2026-04-28T06:27:18Z | 2026-05-03T04:37:42Z | 2026-04-30T11:21:08Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Gr4via | 279,906,021 | U_kgDOEK8G5Q | User | false |
openclaw/openclaw | 4,340,882,260 | I_kwDOQb6kR88AAAABArybVA | 73,307 | https://github.com/openclaw/openclaw/issues/73307 | https://api.github.com/repos/openclaw/openclaw/issues/73307 | [Bug]: PluginEntrySchema zod schema missing allowConversationAccess — non-bundled plugins cannot use conversation hooks | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Non-bundled plugins (e.g., openclaw-observability) are blocked from using conversation-related typed hooks (like llm_input, llm_output) because the zod schema for PluginEntrySchema does not include allowConversationAccess, d... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-28T05:44:37Z | 2026-05-01T04:44:12Z | 2026-04-28T06:41:39Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | weblily588 | 266,917,648 | U_kgDOD-jXEA | User | false |
openclaw/openclaw | 4,341,005,573 | I_kwDOQb6kR88AAAABAr59BQ | 73,329 | https://github.com/openclaw/openclaw/issues/73329 | https://api.github.com/repos/openclaw/openclaw/issues/73329 | [Bug]: Gateway hard-couples to OpenRouter + LiteLLM pricing fetches at boot; no opt-out | ## Summary
Gateway boot is hard-coupled to two external HTTP endpoints (OpenRouter `/models` and the LiteLLM pricing JSON on raw.githubusercontent.com) for the model-pricing cache. There is no config flag, env var, or plugin disable to opt out. On networks where either endpoint is slow or unreachable, the fetches stac... | closed | not_planned | true | 3 | [] | [] | 2026-04-28T06:14:45Z | 2026-05-01T04:45:02Z | 2026-04-28T06:20:13Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | rjdjohnston | 293,686 | MDQ6VXNlcjI5MzY4Ng== | User | false |
openclaw/openclaw | 4,341,117,123 | I_kwDOQb6kR88AAAABAsAwww | 73,343 | https://github.com/openclaw/openclaw/issues/73343 | https://api.github.com/repos/openclaw/openclaw/issues/73343 | [Feature]: Expose channel user identity (SenderId) to skill/tool execution environment | ## Summary
Allow plugins/skills to access the channel user's identity (SenderId) during tool/script execution, so downstream APIs can identify which user triggered the action.
## Problem to solve
When OpenClaw calls external company APIs on behalf of a user (via skills or tools that execute shell scripts), the downs... | open | null | false | 2 | [] | [] | 2026-04-28T06:39:21Z | 2026-05-03T09:49:19Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | eightHundreds | 18,695,431 | MDQ6VXNlcjE4Njk1NDMx | User | false |
openclaw/openclaw | 4,067,926,199 | I_kwDOQb6kR87yd6C3 | 44,458 | https://github.com/openclaw/openclaw/issues/44458 | https://api.github.com/repos/openclaw/openclaw/issues/44458 | [Feature Request] Add agent:error event hook support for auto-recovery mechanisms | ## Problem
Currently, OpenClaw hooks system supports events like command:new, command:reset, gateway:startup, etc. However, there's no event for when the agent encounters errors (like API errors, 401 authentication failures, timeouts, etc.).
## Use Case
Users want to implement auto-recovery mechanisms that:
1. Detec... | closed | completed | true | 2 | [
"stale"
] | [] | 2026-03-12T23:53:19Z | 2026-05-01T04:45:00Z | 2026-04-28T06:45:08Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | jiaguoxinzhi | 13,130,772 | MDQ6VXNlcjEzMTMwNzcy | User | false |
openclaw/openclaw | 4,068,332,019 | I_kwDOQb6kR87yfdHz | 44,539 | https://github.com/openclaw/openclaw/issues/44539 | https://api.github.com/repos/openclaw/openclaw/issues/44539 | openclaw doctor reports 'Telegram: not configured' despite Telegram channel working correctly | ## Description
`openclaw doctor` reports `Telegram: not configured` in the channel probe section, even though the Telegram channel is fully operational.
## Current Behavior
- `openclaw doctor` → `Telegram: not configured`
- `openclaw status` → `Telegram │ ON │ OK │ token config×2 (8450…_SqQ · len 46) · accounts 2/2`... | closed | completed | true | 6 | [
"stale"
] | [] | 2026-03-13T02:09:57Z | 2026-05-01T04:44:59Z | 2026-04-28T06:45:14Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | tamaosamu | 5,938,440 | MDQ6VXNlcjU5Mzg0NDA= | User | false |
openclaw/openclaw | 4,068,402,283 | I_kwDOQb6kR87yfuRr | 44,555 | https://github.com/openclaw/openclaw/issues/44555 | https://api.github.com/repos/openclaw/openclaw/issues/44555 | [Bug]: Discord attachments sent via message tool do not arrive (buffer reports success, path inputs fail) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Discord attachments sent through the OpenClaw message tool do not arrive correctly: buffer-based sends can report success without a visible attachment, while path-based sends fail.
### Steps to reproduce
1. Configure a working Discord cha... | closed | not_planned | true | 2 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-13T02:33:12Z | 2026-05-01T04:44:58Z | 2026-04-28T06:45:20Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | canadaeee | 235,809,206 | U_kgDODg4ptg | User | false |
openclaw/openclaw | 4,068,460,093 | I_kwDOQb6kR87yf8Y9 | 44,565 | https://github.com/openclaw/openclaw/issues/44565 | https://api.github.com/repos/openclaw/openclaw/issues/44565 | [Bug]: Interactive browser tool fails despite healthy gateway (Docker / Hostinger image) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
Interactive Browser Tool Fails Despite Healthy Gateway (Docker / Hostinger Image)
Environment
Image: ghcr.io/hostinger/hvps-openclaw:latest
OpenClaw version: 2026.3.8 (3caab92)
Deployment: Docker (Hostinger VPS)
Model: openai/gpt-5.2
O... | closed | completed | true | 2 | [
"bug",
"stale",
"bug:behavior"
] | [] | 2026-03-13T02:50:04Z | 2026-05-01T04:44:58Z | 2026-04-28T06:45:25Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | jeanmi873 | 267,818,436 | U_kgDOD_aVxA | User | false |
openclaw/openclaw | 4,069,379,664 | I_kwDOQb6kR87yjc5Q | 44,794 | https://github.com/openclaw/openclaw/issues/44794 | https://api.github.com/repos/openclaw/openclaw/issues/44794 | Feature Request: gateway.preStartCommand / gateway.wrapperScript for LaunchAgent/systemd | ## Summary
Add configuration options in `openclaw.json` for specifying a pre-start command or wrapper script for the gateway service, so custom startup logic persists across `openclaw gateway install` upgrades.
## Problem
When running OpenClaw Gateway as a LaunchAgent (macOS) or systemd service (Linux), there is no ... | closed | completed | true | 4 | [
"stale"
] | [] | 2026-03-13T07:11:29Z | 2026-05-01T04:44:02Z | 2026-04-28T06:45:32Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | wulonchia-pro | 267,232,346 | U_kgDOD-2kWg | User | false |
openclaw/openclaw | 4,341,121,274 | I_kwDOQb6kR88AAAABAsBA-g | 73,344 | https://github.com/openclaw/openclaw/issues/73344 | https://api.github.com/repos/openclaw/openclaw/issues/73344 | [Bug]: /new on openai-codex/* (Responses API) returns "One of input/previous_response_id/prompt/conversation_id must be provided" — regression in 2026.4.26 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
### Bug type
Regression (worked before, now fails)
### Summary
After updating to OpenClaw 2026.4.26, `/new` on `openai-codex/gpt-5.5`
immediately fails with:
`One of "input" or "previous_respo... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-28T06:40:15Z | 2026-05-01T04:44:57Z | 2026-04-28T06:45:35Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | bruhsh | 114,047,165 | U_kgDOBsw4vQ | User | false |
openclaw/openclaw | 4,341,141,839 | I_kwDOQb6kR88AAAABAsCRTw | 73,348 | https://github.com/openclaw/openclaw/issues/73348 | https://api.github.com/repos/openclaw/openclaw/issues/73348 | openclaw update can succeed while the running gateway stays on an older version, causing plugin/config version skew | ## Summary
`openclaw update` can succeed, but the running gateway/app service can still come back on an older runtime. In our case, the update path and the service path appeared to drift apart, which left the system in a broken mixed-version state.
## What happened
- The machine successfully updated OpenClaw via CLI... | closed | completed | true | 1 | [] | [] | 2026-04-28T06:44:46Z | 2026-05-01T04:44:54Z | 2026-04-28T06:50:27Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | mattskii89 | 143,202,885 | U_kgDOCIkaRQ | User | false |
openclaw/openclaw | 4,338,155,719 | I_kwDOQb6kR88AAAABApMAxw | 72,985 | https://github.com/openclaw/openclaw/issues/72985 | https://api.github.com/repos/openclaw/openclaw/issues/72985 | Cron isolated-agent: fallback chain silently delivers incomplete-turn warning, and status=error rows mark delivered=true | # Cron isolated-agent: provider-fallback failures silently deliver "Agent couldn't generate a response" + status=error rows mark delivered=true
## TL;DR
When a cron job's primary provider (e.g. `anthropic-proxy`) is temporarily unavailable, OpenClaw's cron isolated-agent runner falls back to the configured backup pro... | open | null | false | 1 | [] | [] | 2026-04-27T18:53:51Z | 2026-05-03T07:05:19Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Allenbluff | 131,441,706 | U_kgDOB9WkKg | User | false |
openclaw/openclaw | 4,338,183,751 | I_kwDOQb6kR88AAAABApNuRw | 72,991 | https://github.com/openclaw/openclaw/issues/72991 | https://api.github.com/repos/openclaw/openclaw/issues/72991 | [Feature]: Expose machine-wide hook policies | ### Summary
Add an admin-managed, system-wide hook discovery tier whose entries always load, run before user-controlled hooks, and cannot be disabled by **_openclaw hooks disable_**.
### Problem to solve
Enterprises self-deploying OpenClaw across their fleet need an admin governance surface, and today there isn't on... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-27T18:59:04Z | 2026-05-03T09:13:17Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | amirkutcher | 244,961,889 | U_kgDODpnSYQ | User | true |
openclaw/openclaw | 4,338,236,726 | I_kwDOQb6kR88AAAABApQ9Ng | 72,995 | https://github.com/openclaw/openclaw/issues/72995 | https://api.github.com/repos/openclaw/openclaw/issues/72995 | [Bug]: web_search tool aborts silently with Gemini provider — instantaneous abort, no err.log entry | ## Summary
`web_search` with the Gemini provider (google_search grounding) intermittently aborts instantly with the message **"This operation was aborted"**. The abort is near-instantaneous (<1s), ruling out timeout. The error is swallowed by the tool layer and never appears in `gateway.err.log`.
## Environment
- **... | closed | completed | true | 2 | [] | [] | 2026-04-27T19:07:44Z | 2026-05-05T04:28:59Z | 2026-05-02T04:36:17Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | RoseKongPS | 270,016,663 | U_kgDOEBgglw | User | false |
openclaw/openclaw | 4,338,261,040 | I_kwDOQb6kR88AAAABApScMA | 72,997 | https://github.com/openclaw/openclaw/issues/72997 | https://api.github.com/repos/openclaw/openclaw/issues/72997 | Feature request: channel reload <provider> command | ## Feature Request
### Problem
When a channel plugin encounters an error or needs its config refreshed, the only current option is a full gateway restart (`openclaw gateway restart`). This is disruptive — it kills all active sessions and introduces ~30s downtime — when the intent is only to reload a single channel.
#... | open | null | false | 1 | [] | [] | 2026-04-27T19:12:04Z | 2026-05-03T08:39:45Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | axonrelaybot | 275,827,808 | U_kgDOEHDMYA | User | false |
openclaw/openclaw | 4,074,484,699 | I_kwDOQb6kR87y27Pb | 45,686 | https://github.com/openclaw/openclaw/issues/45686 | https://api.github.com/repos/openclaw/openclaw/issues/45686 | Compaction circuit breaker: fallback model + force-truncate after repeated failures | ## Problem
When the primary model (e.g. Opus) is degraded or overloaded, context compaction can enter a death spiral:
1. Context hits the limit, compaction triggers
2. Compaction sends full context to the model, which times out (10min timeout)
3. Failure kills Telegram polling ("Unsubscribed during compaction")
4. Ne... | open | null | false | 3 | [] | [] | 2026-03-14T03:12:40Z | 2026-05-04T00:12:53Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | apiekar | 114,312,144 | U_kgDOBtBD0A | User | false |
openclaw/openclaw | 4,341,014,684 | I_kwDOQb6kR88AAAABAr6gnA | 73,332 | https://github.com/openclaw/openclaw/issues/73332 | https://api.github.com/repos/openclaw/openclaw/issues/73332 | v2026.4.25: Agents with no fallbacks array auto-fail-over to every registered models.providers.* entry | # v2026.4.25: Agents with no `fallbacks` array auto-fail-over to every registered `models.providers.*` entry
## Summary
Starting in `v2026.4.25`, when an agent's primary model fails, the gateway will attempt every model registered under `models.providers.*` as a fallback — even when the agent's `model` config has **n... | closed | completed | true | 3 | [] | [] | 2026-04-28T06:16:39Z | 2026-05-01T04:44:53Z | 2026-04-28T07:02:17Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | haumanto | 72,009,799 | MDQ6VXNlcjcyMDA5Nzk5 | User | false |
openclaw/openclaw | 4,340,839,693 | I_kwDOQb6kR88AAAABArv1DQ | 73,303 | https://github.com/openclaw/openclaw/issues/73303 | https://api.github.com/repos/openclaw/openclaw/issues/73303 | v2026.4.26: gateway restart can hang ~3-4 min before new process starts | ## Summary
On macOS (LaunchAgent mode), restarting Gateway on v2026.4.26 can sometimes take ~3-4 minutes before the new process starts, even though normal startup is only a few seconds.
## Environment
- OpenClaw: `2026.4.26 (be8c246)`
- OS: macOS 15/26 (arm64)
- LaunchAgent label: `ai.openclaw.gateway`
- Active channe... | open | null | false | 12 | [] | [] | 2026-04-28T05:34:18Z | 2026-05-02T16:56:20Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | githubdulong | 59,136,159 | MDQ6VXNlcjU5MTM2MTU5 | User | false |
openclaw/openclaw | 4,341,212,568 | I_kwDOQb6kR88AAAABAsGlmA | 73,352 | https://github.com/openclaw/openclaw/issues/73352 | https://api.github.com/repos/openclaw/openclaw/issues/73352 | MCP Client: Add OAuth2 Authorization Code Flow support for HTTP-based MCP servers | ### Summary
OAuth2 Authorization Code Flow support for HTTP-based MCP servers
### Problem to solve
OpenClaw cannot connect to MCP servers that use OAuth2 Authorization Code Flow + PKCE for authentication. When configured in Cursor, the flow is: browser opens OAuth2 login page → user authenticates with corporate acco... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-28T07:00:07Z | 2026-05-01T04:44:52Z | 2026-04-28T07:05:24Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | feng9875412 | 66,237,368 | MDQ6VXNlcjY2MjM3MzY4 | User | false |
openclaw/openclaw | 4,340,865,520 | I_kwDOQb6kR88AAAABArxZ8A | 73,305 | https://github.com/openclaw/openclaw/issues/73305 | https://api.github.com/repos/openclaw/openclaw/issues/73305 | openclaw auth list hangs indefinitely with no output | # `openclaw auth list` hangs indefinitely with no output
## Summary
The `openclaw auth list` subcommand hangs indefinitely and never produces output, regardless of TTY/non-TTY context, `--help`, or `--json` flags. This makes it impossible to inspect configured auth profiles via the CLI.
## Environment
- **OpenClaw ve... | open | null | false | 3 | [] | [] | 2026-04-28T05:40:20Z | 2026-05-03T06:07:42Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | hcostaBR | 67,788,052 | MDQ6VXNlcjY3Nzg4MDUy | User | false |
openclaw/openclaw | 4,341,214,414 | I_kwDOQb6kR88AAAABAsGszg | 73,353 | https://github.com/openclaw/openclaw/issues/73353 | https://api.github.com/repos/openclaw/openclaw/issues/73353 | 60s startup hang in sidecars.channels — synchronous plugin manifest re-discovery on every cold start (v2026.4.26) | ## Environment
- **OpenClaw:** 2026.4.26 (be8c246)
- **Node:** v24.15.0
- **OS:** Ubuntu 24.04 (6.8.0-110-generic, x86_64)
- **Deployment:** `openclaw-gateway.service` via systemd user unit
## Symptom
`openclaw gateway start` takes ~67s from systemd start to `[gateway] ready`. The 55-60s window is silent (no logs). ... | closed | completed | true | 1 | [] | [] | 2026-04-28T07:00:30Z | 2026-05-01T04:44:49Z | 2026-04-28T07:06:18Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | chsbusch-dot | 265,684,486 | U_kgDOD9YGBg | User | false |
openclaw/openclaw | 4,139,572,813 | I_kwDOQb6kR872vN5N | 54,786 | https://github.com/openclaw/openclaw/issues/54786 | https://api.github.com/repos/openclaw/openclaw/issues/54786 | skills.load.extraDirs: support recursive scanning or glob patterns | ## Problem
`extraDirs` only scans one level deep (`<dir>/*/SKILL.md`). This makes it incompatible with plugin/marketplace directory structures that nest skills deeper, e.g.:
```
plugins/
├── gemini/skills/gemini/SKILL.md
├── grok/skills/grok/SKILL.md
└── monitor-pr/skills/monitor-pr/SKILL.md
```
Currently the only w... | closed | completed | true | 1 | [] | [] | 2026-03-26T00:58:41Z | 2026-05-01T04:44:48Z | 2026-04-28T07:11:46Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | shengdade | 16,719,907 | MDQ6VXNlcjE2NzE5OTA3 | User | false |
openclaw/openclaw | 4,197,824,563 | I_kwDOQb6kR876Nbgz | 59,998 | https://github.com/openclaw/openclaw/issues/59998 | https://api.github.com/repos/openclaw/openclaw/issues/59998 | Feature Request: Add tools.exec.shell config to specify default shell executable | ## Feature Request: tools.exec.shell configuration option
**Is your feature request related to a problem? Please describe.**
On Windows, the exec tool defaults to PowerShell/cmd. Users who prefer bash (e.g., Git Bash from MSYS2, Cygwin, or Fork git's bundled bash) have no way to change this default behavior globally.
... | closed | not_planned | true | 1 | [] | [] | 2026-04-03T01:59:03Z | 2026-05-01T04:44:48Z | 2026-04-28T07:11:53Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | PheobeAI | 267,847,712 | U_kgDOD_cIIA | User | false |
openclaw/openclaw | 4,199,870,344 | I_kwDOQb6kR876VO-I | 60,226 | https://github.com/openclaw/openclaw/issues/60226 | https://api.github.com/repos/openclaw/openclaw/issues/60226 | Non-image file attachments (PDF/Word/etc) not injected into agent context after upload via webchat | ## Summary
When a non-image file (PDF, Word, Excel, etc.) is uploaded via webchat, the file is saved to media/inbound/ but the agent never receives the file path in its context.
Images work (passed as image_url blocks). Non-image files have no equivalent injection mechanism.
## Steps to Reproduce
1. Upload a PDF via... | closed | completed | true | 1 | [] | [] | 2026-04-03T10:29:57Z | 2026-05-01T04:44:47Z | 2026-04-28T07:12:02Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | FamilyLiOpenClaw | 273,331,171 | U_kgDOEEqz4w | User | false |
openclaw/openclaw | 4,337,084,686 | I_kwDOQb6kR88AAAABAoKpDg | 72,899 | https://github.com/openclaw/openclaw/issues/72899 | https://api.github.com/repos/openclaw/openclaw/issues/72899 | Empty-text ContentBlock persisted to transcript permanently bricks a session (Bedrock validation error: messages.N.content.0 is blank) | ## Summary
An empty-text user `ContentBlock` can be persisted to a session transcript, permanently poisoning the session. Every subsequent LLM call fails with a Bedrock/Anthropic validation error and the session becomes unusable until the transcript is manually edited.
```
Validation error: The text field in the Cont... | closed | completed | true | 2 | [] | [] | 2026-04-27T15:58:53Z | 2026-05-01T04:44:44Z | 2026-04-28T07:13:01Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | jowhee327 | 67,151,378 | MDQ6VXNlcjY3MTUxMzc4 | User | false |
openclaw/openclaw | 4,337,951,573 | I_kwDOQb6kR88AAAABAo_jVQ | 72,975 | https://github.com/openclaw/openclaw/issues/72975 | https://api.github.com/repos/openclaw/openclaw/issues/72975 | webchat shows each user message twice (chat.history dumps all JSONL lines instead of walking the active branch) | ### Symptom
In the webchat (`openclaw-control-ui`), every user message renders twice in the chat history: once as the LLM-facing copy (with the `Sender (untrusted metadata):` block prepended), and once as the cleaned transcript copy (without the wrapper). Reloading the page or reconnecting reproduces it 100%.
### Vers... | closed | completed | true | 2 | [] | [] | 2026-04-27T18:18:18Z | 2026-05-01T04:40:25Z | 2026-04-28T07:13:13Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | dmagdici | 269,990,928 | U_kgDOEBe8EA | User | false |
openclaw/openclaw | 4,340,934,976 | I_kwDOQb6kR88AAAABAr1pQA | 73,319 | https://github.com/openclaw/openclaw/issues/73319 | https://api.github.com/repos/openclaw/openclaw/issues/73319 | openclaw channels list only shows auth providers, not configured chat channel accounts | ## Summary
openclaw channels list displays only Auth providers and does not list configured chat channel accounts (e.g., Telegram, Discord), even though the channels are properly configured and operational.
## Steps to reproduce
1. Configure Telegram accounts in channels.telegram.accounts
2. Run openclaw channels li... | closed | completed | true | 2 | [] | [] | 2026-04-28T05:57:59Z | 2026-05-01T04:44:42Z | 2026-04-28T07:33:41Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | mlaihk | 25,972,362 | MDQ6VXNlcjI1OTcyMzYy | User | false |
openclaw/openclaw | 4,340,939,351 | I_kwDOQb6kR88AAAABAr16Vw | 73,322 | https://github.com/openclaw/openclaw/issues/73322 | https://api.github.com/repos/openclaw/openclaw/issues/73322 | openclaw channels list only shows auth providers, not configured chat channel accounts | ## Summary
`openclaw channels list` displays only Auth providers and does not list configured chat channel accounts (e.g., Telegram, Discord), even though the channels are properly configured and operational.
## Steps to reproduce
1. Configure Telegram accounts in `channels.telegram.accounts`
2. Run `openclaw channe... | closed | completed | true | 2 | [] | [] | 2026-04-28T05:59:07Z | 2026-05-01T04:44:42Z | 2026-04-28T07:33:50Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | mlaihk | 25,972,362 | MDQ6VXNlcjI1OTcyMzYy | User | false |
openclaw/openclaw | 4,341,266,867 | I_kwDOQb6kR88AAAABAsJ5sw | 73,360 | https://github.com/openclaw/openclaw/issues/73360 | https://api.github.com/repos/openclaw/openclaw/issues/73360 | [Bug]: text RangeError: Maximum call stack size exceeded | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
在 OpenClaw Control UI 中上传 PDF 或长图片并发送给 WebChat 分析时，请求在 `chat.send` 阶段失败，返回：
```text
RangeError: Maximum call stack size exceeded
### Steps to reproduce
[openclaw-control-ui-attachment-stack-overflow-report.... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-28T07:11:29Z | 2026-05-01T04:39:59Z | 2026-04-28T07:34:05Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | hejunhui-73 | 206,196,442 | U_kgDODEpO2g | User | false |
openclaw/openclaw | 4,341,331,346 | I_kwDOQb6kR88AAAABAsN1kg | 73,376 | https://github.com/openclaw/openclaw/issues/73376 | https://api.github.com/repos/openclaw/openclaw/issues/73376 | [Bug] CLI命令(openclaw status/cron list)卡住不动 | ># OpenClaw CLI 命令卡住问题
>
>## 问题描述
>- `openclaw status` 命令会卡住不动
>- `openclaw cron list` 等其他命令也会卡住
>- Gateway Dashboard 正常运行
>- 核心功能（飞书聊天等）正常
>
>## 环境
>- OpenClaw: 2026.4.26
>- OS: WSL2 (linux 6.6.87.2-microsoft-standard-WSL2)
>- Node: v22.22.2
>
>## 测试结果
>- `openclaw --version` ✅ 正常
>- `openclaw health` ✅ 正常
>- `curl ht... | closed | completed | true | 3 | [] | [] | 2026-04-28T07:24:41Z | 2026-05-03T04:35:38Z | 2026-04-30T18:16:21Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | LittleRollo | 256,484,611 | U_kgDOD0mlAw | User | false |
openclaw/openclaw | 4,221,667,245 | I_kwDOQb6kR877oYet | 62,847 | https://github.com/openclaw/openclaw/issues/62847 | https://api.github.com/repos/openclaw/openclaw/issues/62847 | [Bug] Main session can brick on arcee/trinity-large-thinking: repeated connection errors, TUI spam, Telegram nonresponsive until delayed fallback | ## Summary
Switching the **main session** model to `arcee/trinity-large-thinking` can brick the session: the TUI repeatedly shows `connection error`, Telegram becomes nonresponsive, and the main lane stalls until fallback finally fires ~20-24s later. The same model can work in other contexts, which makes this a dangero... | closed | completed | true | 1 | [] | [] | 2026-04-08T02:23:08Z | 2026-05-01T04:44:38Z | 2026-04-28T07:42:45Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Adam-Researchh | 261,827,313 | U_kgDOD5sq8Q | User | false |
openclaw/openclaw | 4,341,403,030 | I_kwDOQb6kR88AAAABAsSNlg | 73,381 | https://github.com/openclaw/openclaw/issues/73381 | https://api.github.com/repos/openclaw/openclaw/issues/73381 | Gateway agent invocation blocked by session sendPolicy even when not delivering | ## Bug Description
On OpenClaw `2026.4.25`, a gateway `agent` request can be rejected with:
```text
GatewayClientRequestError: send blocked by session policy
```
This happens for a non-delivery quick-agent smoke invocation, after a clean maintenance-window canary with no bus watcher and no local `openclaw-agent` pro... | closed | completed | false | 1 | [] | [] | 2026-04-28T07:39:00Z | 2026-05-03T09:57:13Z | 2026-05-03T09:57:13Z | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | marcusglee11 | 200,859,841 | U_kgDOC_jgwQ | User | false |
openclaw/openclaw | 4,341,420,635 | I_kwDOQb6kR88AAAABAsTSWw | 73,387 | https://github.com/openclaw/openclaw/issues/73387 | https://api.github.com/repos/openclaw/openclaw/issues/73387 | Discord: channels block should be additive overrides, not an implicit allowlist | ## Problem
When a Discord guild config includes a `channels` block to set per-channel overrides (e.g., `requireMention: true` on specific channels), **all non-listed channels are implicitly denied**. This is documented behavior ("if a guild has `channels` configured, non-listed channels are denied"), but it creates an... | closed | not_planned | true | 1 | [] | [] | 2026-04-28T07:42:27Z | 2026-05-05T04:26:10Z | 2026-05-02T22:21:52Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | NOVA-Openclaw | 258,138,572 | U_kgDOD2LhzA | User | false |
openclaw/openclaw | 4,338,390,058 | I_kwDOQb6kR88AAAABApaUKg | 73,002 | https://github.com/openclaw/openclaw/issues/73002 | https://api.github.com/repos/openclaw/openclaw/issues/73002 | [BUG] 2026.4.25: bundled runtime-deps staging blocks gateway for minutes, causing handshake timeouts | ## Bug Description
Running `openclaw status` (or any CLI command) triggers bundled runtime-deps staging for plugins that are not actively enabled or configured. Each failed plugin install can block for **76-366 seconds**. During this window:
- Gateway HTTP/WS handshakes timeout
- `/healthz` can timeout
- CPU stays at... | closed | completed | true | 2 | [] | [] | 2026-04-27T19:35:29Z | 2026-04-30T04:30:31Z | 2026-04-27T20:55:23Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | Schnup03 | 145,503,562 | U_kgDOCKw1Sg | User | false |
openclaw/openclaw | 4,340,882,752 | I_kwDOQb6kR88AAAABArydQA | 73,308 | https://github.com/openclaw/openclaw/issues/73308 | https://api.github.com/repos/openclaw/openclaw/issues/73308 | infer model run should expose a flag to skip agent-context wrapping for raw provider testing | ## Problem
`openclaw infer model run --local --model <provider>/<model> --prompt "..."` is documented as "Run a one-shot model turn" and the source path in `src/cli/capability-cli.ts:580-680` looks like a thin builder that would dispatch only `[{ role: "user", content: <prompt> }]`. In practice the embedded runtime it... | closed | completed | true | 2 | [] | [] | 2026-04-28T05:44:46Z | 2026-05-01T04:44:06Z | 2026-04-28T09:22:28Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | ScientificProgrammer | 4,348,294 | MDQ6VXNlcjQzNDgyOTQ= | User | false |
openclaw/openclaw | 4,341,524,549 | I_kwDOQb6kR88AAAABAsZoRQ | 73,407 | https://github.com/openclaw/openclaw/issues/73407 | https://api.github.com/repos/openclaw/openclaw/issues/73407 | [Bug]: The libsignal component is under the GPLv3 open source license and has a risk of infection | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
WhatsApp introduces libsignal component, which is an open source GPLv3 license and poses a risk of infection；
### Steps to reproduce
none
### Expected behavior
none
### Actual behavior
none
### OpenClaw version
2026.... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-28T08:02:10Z | 2026-05-05T04:26:16Z | 2026-05-02T20:59:50Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | chchliang | 23,158,339 | MDQ6VXNlcjIzMTU4MzM5 | User | false |
openclaw/openclaw | 4,340,960,954 | I_kwDOQb6kR88AAAABAr3Oug | 73,323 | https://github.com/openclaw/openclaw/issues/73323 | https://api.github.com/repos/openclaw/openclaw/issues/73323 | [Bug]: Gateway runtime degradation: pricing fetch 60s timeouts, Telegram polling stalls, slow RPC — chronic across 4.23/4.25/4.26 on Windows 11 + Node 24 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Gateway long-running Node process exhibits multi-subsystem network/timer degradation (model-pricing fetch 60s timeouts, Telegram polling stalls 127–266s, RPC slowdowns 8–83s) reproducible across 2026.4.23, 2026.4.25, and 202... | open | null | false | 15 | [
"bug"
] | [] | 2026-04-28T06:04:39Z | 2026-05-03T20:25:10Z | null | NONE | null | 20260503T210034Z | 2026-05-03T21:00:34Z | maoruilun0411-del | 276,904,174 | U_kgDOEIE47g | User | false |
openclaw/openclaw | 4,341,315,204 | I_kwDOQb6kR88AAAABAsM2hA | 73,370 | https://github.com/openclaw/openclaw/issues/73370 | https://api.github.com/repos/openclaw/openclaw/issues/73370 | LINE: Add inbound media persistence to ~/.openclaw/media/inbound/ | # Issue: LINE inbound media missing persistence compared to WhatsApp
## Problem Description
The LINE channel handles inbound images differently from the WhatsApp channel:
- **WhatsApp**: Uses `saveMediaBuffer(..., "inbound", ...)` to persist images to `~/.openclaw/media/inbound/`
- **LINE**: Only saves to `/tmp/openc... | closed | completed | true | 2 | [] | [] | 2026-04-28T07:21:20Z | 2026-05-01T04:44:35Z | 2026-04-28T08:15:03Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | hijirii | 160,108,038 | U_kgDOCYsOBg | User | false |
openclaw/openclaw | 4,341,170,109 | I_kwDOQb6kR88AAAABAsD_vQ | 73,350 | https://github.com/openclaw/openclaw/issues/73350 | https://api.github.com/repos/openclaw/openclaw/issues/73350 | MiniMax-only configuration still invokes claude-cli/codex-cli and can hang TUI/model execution | ## Summary
With OpenClaw 2026.4.26 configured to use only `minimax/MiniMax-M2.7`, the TUI / agent execution path still invokes `claude-cli` and later `codex-cli`. The run becomes extremely slow or hangs, and related OpenClaw processes can consume high CPU.
## Environment
- OpenClaw: 2026.4.26
- macOS: Darwin 25.4.0 ... | closed | completed | true | 2 | [] | [] | 2026-04-28T06:50:51Z | 2026-05-01T04:44:33Z | 2026-04-28T08:18:53Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | jasonlll | 12,613,346 | MDQ6VXNlcjEyNjEzMzQ2 | User | false |
openclaw/openclaw | 4,340,746,488 | I_kwDOQb6kR88AAAABArqI-A | 73,291 | https://github.com/openclaw/openclaw/issues/73291 | https://api.github.com/repos/openclaw/openclaw/issues/73291 | CLI `openclaw status` text mode extremely slow (~20-30s) due to `cache: false` hardcoded in manifest registry loader | ## Bug Report
### Environment
- OpenClaw version: 2026.4.25
- Node.js: v24.14.1
- OS: Ubuntu 24.04 LTS
- Installed plugins: 112 (19 enabled, 93 disabled)
### Problem
`openclaw status` in text mode takes **20-30 seconds** to produce output, appearing to hang. JSON mode (`--json`) completes in ~8 seconds. The process w... | closed | completed | true | 2 | [] | [] | 2026-04-28T05:11:24Z | 2026-05-01T04:44:32Z | 2026-04-28T08:18:55Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | slideshow-dingo | 258,996,445 | U_kgDOD2_43Q | User | false |
openclaw/openclaw | 4,337,860,163 | I_kwDOQb6kR88AAAABAo5-Qw | 72,967 | https://github.com/openclaw/openclaw/issues/72967 | https://api.github.com/repos/openclaw/openclaw/issues/72967 | [Bug] 2026.4.25 regression: bundled plugin deps missing after global npm update; doctor --fix fails; gateway 6min startup; Telegram channel non-functional (works fine on 2026.4.22) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After updating OpenClaw from 2026.4.22 to 2026.4.25 via `npm i -g openclaw@latest` on Windows 11, the installation leaves bundled plugin runtime dependencies missing. `openclaw doctor --fix` does not install them. The gatewa... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-04-27T18:02:09Z | 2026-04-30T04:30:31Z | 2026-04-27T20:11:51Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | centralpc | 59,361,530 | MDQ6VXNlcjU5MzYxNTMw | User | false |
openclaw/openclaw | 4,341,259,220 | I_kwDOQb6kR88AAAABAsJb1A | 73,358 | https://github.com/openclaw/openclaw/issues/73358 | https://api.github.com/repos/openclaw/openclaw/issues/73358 | [Bug]: v2026.4.26 ships `coding-agent` skill + `codex` provider with `openai/gpt-5.5` as silent default — breaks stacks without OpenAI configured | ## Bug
Upgrading to OpenClaw `2026.4.26` (verified `node[…]: 2026.4.26`) on a stack that **never had OpenAI configured** introduces:
1. A `codex` provider in the agent's `models.json` catalog with `gpt-5.5`, `gpt-5.4-mini`, `gpt-5.2`
2. A new bundled skill `coding-agent` (eligible by default because `claude`, `codex`... | closed | completed | true | 2 | [] | [] | 2026-04-28T07:09:59Z | 2026-05-01T04:44:32Z | 2026-04-28T08:24:57Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | LaFleurAdvertising | 14,811,991 | MDQ6VXNlcjE0ODExOTkx | User | false |
openclaw/openclaw | 4,341,540,818 | I_kwDOQb6kR88AAAABAsan0g | 73,411 | https://github.com/openclaw/openclaw/issues/73411 | https://api.github.com/repos/openclaw/openclaw/issues/73411 | [Bug]: Gateway becomes unresponsive after “starting channels and sidecars...” on v2026.4.26 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
On v2026.4.26, the gateway logs “http server listening” and then “starting channels and sidecars...”, but all internal HTTP routes time out and the container remains unhealthy; v2026.4.23 works in the same setup.
Observed o... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-04-28T08:05:17Z | 2026-05-01T04:39:09Z | 2026-04-28T08:31:37Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | eurojojo | 1,299,091 | MDQ6VXNlcjEyOTkwOTE= | User | false |
openclaw/openclaw | 4,341,296,549 | I_kwDOQb6kR88AAAABAsLtpQ | 73,366 | https://github.com/openclaw/openclaw/issues/73366 | https://api.github.com/repos/openclaw/openclaw/issues/73366 | [Bug]: [Regression]: Ollama thinking parameter always forced to false after 2026.4.26 (be8c246) update | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
**Version:** 2026.4.26 (`be8c246`)
**Regression introduced:** Worked correctly before 2026.4.15
---
After updating to `be8c246`, the `think` parameter for Ollama models is always sent as `false` regardless of any configur... | closed | completed | true | 8 | [
"bug",
"regression"
] | [] | 2026-04-28T07:17:42Z | 2026-05-01T04:39:46Z | 2026-04-28T08:34:40Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | cymise | 83,527,948 | MDQ6VXNlcjgzNTI3OTQ4 | User | false |
openclaw/openclaw | 4,341,676,028 | I_kwDOQb6kR88AAAABAsi3_A | 73,422 | https://github.com/openclaw/openclaw/issues/73422 | https://api.github.com/repos/openclaw/openclaw/issues/73422 | Sorry clanker posted in the wrong repo | null | closed | completed | true | 0 | [] | [] | 2026-04-28T08:29:40Z | 2026-05-01T04:44:30Z | 2026-04-28T08:33:37Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | NZF-JDWang | 31,868,402 | MDQ6VXNlcjMxODY4NDAy | User | false |
openclaw/openclaw | 4,341,516,073 | I_kwDOQb6kR88AAAABAsZHKQ | 73,405 | https://github.com/openclaw/openclaw/issues/73405 | https://api.github.com/repos/openclaw/openclaw/issues/73405 | [Bug][v2026.4.24] Control UI webchat: duplicate messages delivered (async command completion events) | - | closed | completed | true | 3 | [] | [] | 2026-04-28T08:00:39Z | 2026-05-01T04:42:50Z | 2026-04-28T16:46:45Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | 7inRivulet | 152,936,729 | U_kgDOCR2hGQ | User | false |
openclaw/openclaw | 4,341,297,328 | I_kwDOQb6kR88AAAABAsLwsA | 73,367 | https://github.com/openclaw/openclaw/issues/73367 | https://api.github.com/repos/openclaw/openclaw/issues/73367 | Bug: bare /reset falls through to model call with empty input | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Sending a bare `/reset` command in a Discord channel resets/rolls the session but then attempts normal model execution with an empty input, producing a provider error instead of a reset acknowledgement.
### St... | closed | completed | true | 4 | [] | [] | 2026-04-28T07:17:51Z | 2026-05-01T04:44:29Z | 2026-04-28T08:31:31Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | hoyanhan | 45,415,044 | MDQ6VXNlcjQ1NDE1MDQ0 | User | false |
openclaw/openclaw | 4,341,601,776 | I_kwDOQb6kR88AAAABAseV8A | 73,413 | https://github.com/openclaw/openclaw/issues/73413 | https://api.github.com/repos/openclaw/openclaw/issues/73413 | [Bug]: When the sub-agent compresses, the session yield incorrectly returns intermediate results, causing the system to fail to produce the correct outcome once compression is complete and execution proceeds. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
This is a troublesome bug that makes sub-agent calls highly unstable. I have to strictly control the overall context length of the sub-agent; otherwise, once the limit is reached, it will directly result in the sub-agent’s w... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-28T08:16:40Z | 2026-05-01T04:44:28Z | 2026-04-28T08:38:12Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Ask-sola | 88,783,257 | MDQ6VXNlcjg4NzgzMjU3 | User | false |
openclaw/openclaw | 4,341,089,006 | I_kwDOQb6kR88AAAABAr_C7g | 73,339 | https://github.com/openclaw/openclaw/issues/73339 | https://api.github.com/repos/openclaw/openclaw/issues/73339 | [Bug]: Gateway cold start is dominated by bundled runtime mirror work (not deps install) on Docker Desktop + WSL | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
## Bug type
Startup / performance bug
## Summary
On Windows 11 with Docker Desktop + WSL, `openclaw-gateway` cold start takes about 3+ minutes before reaching `http server listening`.
This does **not** look... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-28T06:32:48Z | 2026-05-01T04:40:01Z | 2026-04-28T08:39:38Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Ryuu-64 | 61,548,752 | MDQ6VXNlcjYxNTQ4NzUy | User | false |
openclaw/openclaw | 4,337,207,817 | I_kwDOQb6kR88AAAABAoSKCQ | 72,906 | https://github.com/openclaw/openclaw/issues/72906 | https://api.github.com/repos/openclaw/openclaw/issues/72906 | [Bug]: status channel summary is empty while Telegram health probe is OK on v2026.4.25 | ## Summary
After upgrading a Linux VPS from OpenClaw `2026.4.24` to `2026.4.25`, `openclaw status --deep` shows Telegram as healthy in the `Health` section, but the `Channels` table is empty.
This makes the status output look like no channels are enabled/running, even though Telegram is configured and the health prob... | closed | not_planned | true | 3 | [] | [] | 2026-04-27T16:17:53Z | 2026-05-03T04:36:58Z | 2026-04-30T14:32:42Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | RamaAditya49 | 213,913,142 | U_kgDODMAONg | User | false |
openclaw/openclaw | 4,341,661,895 | I_kwDOQb6kR88AAAABAsiAxw | 73,421 | https://github.com/openclaw/openclaw/issues/73421 | https://api.github.com/repos/openclaw/openclaw/issues/73421 | [Bug]: Regression after upgrading from 2026.4.23 to 2026.4.26: delayed reply delivery and websocket/control-ui/browser stalls in NAS Docker environment | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading from OpenClaw 2026.4.23 to 2026.4.26, reply delivery became noticeably slower in my NAS Docker environment.
This does not look like a pure model-inference slowdown or a Slack-only issue.
The stronger signal ... | closed | completed | true | 5 | [
"bug"
] | [] | 2026-04-28T08:27:28Z | 2026-05-02T04:26:38Z | 2026-04-30T02:04:27Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | realistic-guy | 59,161,385 | MDQ6VXNlcjU5MTYxMzg1 | User | false |
openclaw/openclaw | 4,341,689,805 | I_kwDOQb6kR88AAAABAsjtzQ | 73,424 | https://github.com/openclaw/openclaw/issues/73424 | https://api.github.com/repos/openclaw/openclaw/issues/73424 | image tool: 'Failed to optimize image' error in preprocessing pipeline | ## Bug Report: "Failed to optimize image" error in media understanding pipeline
### Description
The built-in `image` tool fails with "Failed to optimize image" when analyzing JPEG images, even though the configured VLM model (`nvidia/google/gemma-4-31b-it`) works correctly via direct API calls.
### Reproduction Steps... | open | null | false | 8 | [] | [] | 2026-04-28T08:32:00Z | 2026-05-03T08:39:52Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | hijirii | 160,108,038 | U_kgDOCYsOBg | User | false |
openclaw/openclaw | 4,341,524,869 | I_kwDOQb6kR88AAAABAsZphQ | 73,408 | https://github.com/openclaw/openclaw/issues/73408 | https://api.github.com/repos/openclaw/openclaw/issues/73408 | Feishu plugin: @_all (mention all) treated as bot mention, requireMention bypassed | ## Description
When `requireMention: true` is configured for a Feishu account, the plugin should only respond to messages that explicitly @mention the bot. However, messages with `@_all` (mention all) are incorrectly treated as bot mentions, causing the bot to respond when it should not.
## Root Cause
In `extensions... | closed | completed | true | 1 | [] | [] | 2026-04-28T08:02:14Z | 2026-05-01T04:44:27Z | 2026-04-28T08:41:37Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Tinkmagic | 48,941,734 | MDQ6VXNlcjQ4OTQxNzM0 | User | false |
openclaw/openclaw | 4,341,743,461 | I_kwDOQb6kR88AAAABAsm_ZQ | 73,427 | https://github.com/openclaw/openclaw/issues/73427 | https://api.github.com/repos/openclaw/openclaw/issues/73427 | Control UI Realtime Talk: CSP blocks OpenAI WebRTC and UI hides realtime audio errors | ## Summary
On OpenClaw 2026.4.26, the Control UI Realtime Talk flow can fail in two hard-to-debug ways when using the bundled OpenAI realtime browser provider.
1. The Control UI CSP only allowed `connect-src 'self' ws: wss:`. The browser WebRTC setup posts the SDP offer directly to `https://api.openai.com/v1/realtime... | closed | completed | true | 2 | [] | [] | 2026-04-28T08:40:31Z | 2026-05-05T04:28:12Z | 2026-05-02T07:47:59Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | lac5q | 7,991,102 | MDQ6VXNlcjc5OTExMDI= | User | false |
openclaw/openclaw | 4,341,637,215 | I_kwDOQb6kR88AAAABAsggXw | 73,417 | https://github.com/openclaw/openclaw/issues/73417 | https://api.github.com/repos/openclaw/openclaw/issues/73417 | DeepSeek V4: ensureDeepSeekV4ToolCallReasoningContent 未覆盖无 tool_calls 的 assistant 消息 | ## 复现步骤
1. 在会话中开启 DeepSeek V4 thinking mode（如 `/think high`）
2. 使用 `deepseek-v4-flash` 或 `deepseek-v4-pro` 模型
3. 正常对话（会产生普通 assistant 回复，不带 tool_calls）
4. 继续对话，API 返回 400 错误：
```
The `reasoning_content` in the thinking mode must be passed back to the API.
```
## 原因分析
`ensureDeepSeekV4ToolCallReasoningContent` 函数的当前... | closed | completed | true | 6 | [] | [] | 2026-04-28T08:23:03Z | 2026-05-01T04:44:08Z | 2026-04-28T09:16:53Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | 34262315716 | 62,001,009 | MDQ6VXNlcjYyMDAxMDA5 | User | false |
openclaw/openclaw | 4,341,645,791 | I_kwDOQb6kR88AAAABAshB3w | 73,418 | https://github.com/openclaw/openclaw/issues/73418 | https://api.github.com/repos/openclaw/openclaw/issues/73418 | memory-wiki bridge still imports 0 artifacts with QMD backend on 2026.4.26 | ## Summary
The documented QMD + `memory-wiki` bridge setup still reports `0 exported artifacts` on OpenClaw `2026.4.26`, even after QMD is installed, selected as the memory backend, indexed successfully, and `memory-core` is loaded as the active memory plugin.
This looks related to, or a still-reproducible variant of... | closed | not_planned | true | 1 | [] | [] | 2026-04-28T08:24:35Z | 2026-05-01T04:44:25Z | 2026-04-28T08:45:15Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | LittleJakub | 35,995,921 | MDQ6VXNlcjM1OTk1OTIx | User | false |
openclaw/openclaw | 4,341,750,455 | I_kwDOQb6kR88AAAABAsnatw | 73,428 | https://github.com/openclaw/openclaw/issues/73428 | https://api.github.com/repos/openclaw/openclaw/issues/73428 | [Bug]: Severe chat latency (30–90s) on Docker VPS while direct OpenAI completes in <1s — OpenClaw 2026.4.26 | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
Every chat turn through OpenClaw WebChat takes 30–90 seconds end-to-end on a 2 vCPU / 4 GB Docker VPS, while the same gpt-4o-mini call from inside the same container via curl returns in 0.83–1.73 seconds.
### Steps to reproduc... | closed | completed | true | 11 | [
"bug"
] | [] | 2026-04-28T08:41:41Z | 2026-05-04T04:36:40Z | 2026-05-01T11:44:40Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | Dimaoggg | 149,779,914 | U_kgDOCO11yg | User | false |
openclaw/openclaw | 4,171,808,259 | I_kwDOQb6kR874qL4D | 57,795 | https://github.com/openclaw/openclaw/issues/57795 | https://api.github.com/repos/openclaw/openclaw/issues/57795 | WebChat can leave parent agent stuck in typing state after subagent completion | ## Summary
When the main agent delegates to one or more subagents, the subagents can complete successfully and the main agent can post the follow-up summary, but the WebChat UI may still leave the parent agent in a typing / processing state until the page is refreshed.
This appears to be a frontend state reconciliati... | closed | completed | true | 1 | [] | [] | 2026-03-30T15:42:04Z | 2026-05-01T04:44:24Z | 2026-04-28T08:47:02Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | A1fred-AI | 269,621,014 | U_kgDOEBIXFg | User | false |
openclaw/openclaw | 4,341,806,261 | I_kwDOQb6kR88AAAABAsq0tQ | 73,431 | https://github.com/openclaw/openclaw/issues/73431 | https://api.github.com/repos/openclaw/openclaw/issues/73431 | [Bug]: Discord `message read/search` hang indefinitely and Discord channel plugin does not emit standard inbound hooks (`message_received` / `inbound_claim`) — possible regression of #31264 / #33038 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
On OpenClaw 2026.4.25, Discord `message read` and `message search` hang indefinitely with empty stdout, and the Discord channel plugin does not appear to emit the standard inbound hooks (`message_received` / `inbound_claim`)... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-28T08:52:06Z | 2026-05-01T04:43:40Z | 2026-04-28T11:46:07Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | ferchache | 279,961,573 | U_kgDOEK_f5Q | User | false |
openclaw/openclaw | 4,341,562,364 | I_kwDOQb6kR88AAAABAsb7_A | 73,412 | https://github.com/openclaw/openclaw/issues/73412 | https://api.github.com/repos/openclaw/openclaw/issues/73412 | [Bug]: /new causes model to monologue unprompted with leaked context from previous sessions (since 2026.4.24) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Since v2026.4.24 (continuing through 4.25 and 4.26), using the `/new` slash command causes the model to start **monologuing unprompted** — it generates responses without any user input, pulling random content fragments from ... | closed | completed | true | 1 | [] | [] | 2026-04-28T08:09:31Z | 2026-05-01T04:44:22Z | 2026-04-28T08:57:07Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | amdhelper | 232,188,533 | U_kgDODdbqdQ | User | false |
openclaw/openclaw | 4,109,522,309 | I_kwDOQb6kR8708lWF | 51,185 | https://github.com/openclaw/openclaw/issues/51185 | https://api.github.com/repos/openclaw/openclaw/issues/51185 | ensureDockerImage() silently overwrites custom sandbox image with plain debian | ## Summary
`ensureDockerImage()` in `dist/reply-Bm8VrLQh.js` has a hardcoded fallback that runs `docker tag debian:bookworm-slim openclaw-sandbox:bookworm-slim` whenever the sandbox image is missing. This silently replaces any custom-built image (with python3, jq, ripgrep, etc.) with plain debian — breaking Write/Edit... | closed | completed | true | 1 | [] | [] | 2026-03-20T18:41:44Z | 2026-05-01T04:44:20Z | 2026-04-28T08:57:46Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | dpalis | 170,011,854 | U_kgDOCiIszg | User | false |
openclaw/openclaw | 4,185,979,464 | I_kwDOQb6kR875gPpI | 58,956 | https://github.com/openclaw/openclaw/issues/58956 | https://api.github.com/repos/openclaw/openclaw/issues/58956 | [Bug]: Pre-compaction memory flush leaks into chat UI and blocks the user’s active turn in 2026.3.31 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to openclaw 2026.3.31, the internal Pre-compaction memory flush prompt is no longer silent. It is rendered directly in the chat window, even though this mechanism is supposed to be an internal housekeeping st... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-01T10:52:06Z | 2026-05-01T04:44:18Z | 2026-04-28T08:59:06Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | guoyuhang9 | 147,230,674 | U_kgDOCMaP0g | User | false |
openclaw/openclaw | 4,341,822,595 | I_kwDOQb6kR88AAAABAsr0gw | 73,432 | https://github.com/openclaw/openclaw/issues/73432 | https://api.github.com/repos/openclaw/openclaw/issues/73432 | [Bug]: qmd embedding is never triggered per memory.qmd.update.interval/embedInterval | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
I only started to configure qmd (2.1.0) with OpenClaw, and noticed the documents always show pending embedding in qmd status. Manual qmd embed works but every time document changes via OpenClaw, it becomes pend... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-28T08:55:03Z | 2026-05-03T04:44:00Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | danielngn | 2,748,966 | MDQ6VXNlcjI3NDg5NjY= | User | false |
openclaw/openclaw | 4,341,825,054 | I_kwDOQb6kR88AAAABAsr-Hg | 73,433 | https://github.com/openclaw/openclaw/issues/73433 | https://api.github.com/repos/openclaw/openclaw/issues/73433 | [Bug]: Gateway pegs single CPU thread at 100% immediately on boot in 2026.4.26 — TUI handshake timeouts, multi-minute message latency | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading from an earlier release to `2026.4.26`, `openclaw-gateway` immediately pins one CPU thread at 100% on boot and stays there indefinitely, while reporting `gateway ready` in the logs. The TUI cannot establish a... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-04-28T08:55:28Z | 2026-05-01T04:44:17Z | 2026-04-28T09:01:04Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | AnneVoss | 70,486,014 | MDQ6VXNlcjcwNDg2MDE0 | User | false |
openclaw/openclaw | 4,074,850,255 | I_kwDOQb6kR87y4UfP | 45,800 | https://github.com/openclaw/openclaw/issues/45800 | https://api.github.com/repos/openclaw/openclaw/issues/45800 | [Feature Request] Prevent accidental /new session reset — add confirmation dialog or relocate "+" button | ### Summary
The "+" button in the WebChat Dashboard UI triggers /new, which immediately destroys the current session. Due to its position (bottom-right, next to the send button), it is very easy to misclick. This causes permanent loss of unsaved conversation context and memory. A confirmation dialog or button relocati... | closed | completed | true | 5 | [
"enhancement"
] | [] | 2026-03-14T06:24:45Z | 2026-05-01T04:44:11Z | 2026-04-28T09:11:25Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | aethnova | 262,512,133 | U_kgDOD6WeBQ | User | false |
openclaw/openclaw | 4,161,620,254 | I_kwDOQb6kR874DUke | 56,611 | https://github.com/openclaw/openclaw/issues/56611 | https://api.github.com/repos/openclaw/openclaw/issues/56611 | Control UI: + button easily confused with attachment — add option to hide or reposition | ## Problem
The **+ (new session)** button in the Control UI webchat is positioned where users instinctively expect an **attachment button**. This leads to accidentally creating new sessions instead of attaching files.
## Suggestion
A few possible fixes (any would help):
1. **Add a config option** to hide the new se... | closed | completed | true | 1 | [] | [] | 2026-03-28T20:48:33Z | 2026-05-01T04:44:09Z | 2026-04-28T09:11:36Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | adambezemek | 1,010,175 | MDQ6VXNlcjEwMTAxNzU= | User | false |
openclaw/openclaw | 4,079,102,358 | I_kwDOQb6kR87zIimW | 47,610 | https://github.com/openclaw/openclaw/issues/47610 | https://api.github.com/repos/openclaw/openclaw/issues/47610 | config.patch returns full config in response — bloats agent context | ## Problem
`config.patch` and `config.get` return the **entire resolved config** in the tool response. For a typical setup (multiple auth profiles, browser profiles with long WSS URLs, 15+ plugins, multi-channel config), this is **~15,000 tokens per call**.
When an agent makes 5-6 config changes in a session, that's ... | closed | completed | true | 2 | [] | [] | 2026-03-15T20:10:33Z | 2026-05-01T04:44:05Z | 2026-04-28T09:34:41Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | HanenVit | 183,886,923 | U_kgDOCvXkSw | User | false |
openclaw/openclaw | 4,341,981,440 | I_kwDOQb6kR88AAAABAs1hAA | 73,446 | https://github.com/openclaw/openclaw/issues/73446 | https://api.github.com/repos/openclaw/openclaw/issues/73446 | Docs language switcher loses page path (Mintlify platform behavior) | ## Related
- Previous issue: #34021 (closed by stale bot, same problem)
- Attempted fix PRs: #73415, #73436 (closed — see explanation below)
## Problem
When switching languages via the Mintlify language dropdown (top-left), the site navigates to the target locale's homepage instead of the equivalent page.
**Reprodu... | open | null | false | 1 | [] | [] | 2026-04-28T09:21:12Z | 2026-05-03T01:38:20Z | null | CONTRIBUTOR | null | 20260503T030030Z | 2026-05-03T03:00:30Z | lanzhi-lee | 36,190,508 | MDQ6VXNlcjM2MTkwNTA4 | User | false |
openclaw/openclaw | 4,341,976,460 | I_kwDOQb6kR88AAAABAs1NjA | 73,444 | https://github.com/openclaw/openclaw/issues/73444 | https://api.github.com/repos/openclaw/openclaw/issues/73444 | Sub-agent model routing silently ignored - explicit model= falls back to primary model | # Sub-Agent Model Routing Ignored
## Summary
`buildDirectChildSessionPatch()` writes the model to session display metadata fields (`model`, `modelProvider`) but the execution engine reads override fields (`modelOverride`, `providerOverride`). Explicit `sessions_spawn(model="...")` calls and `agents.defaults.subagents... | closed | completed | true | 1 | [] | [] | 2026-04-28T09:20:19Z | 2026-05-01T04:44:03Z | 2026-04-28T09:42:31Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | aurora-openclaw | 273,858,899 | U_kgDOEFLBUw | User | false |
openclaw/openclaw | 4,341,998,844 | I_kwDOQb6kR88AAAABAs2k_A | 73,448 | https://github.com/openclaw/openclaw/issues/73448 | https://api.github.com/repos/openclaw/openclaw/issues/73448 | Bug: message tool asVoice parameter defined in schema but never read in handleSendAction | ## Bug Report
### Summary
The `message` tool's `asVoice` parameter is defined in the schema but **never read or passed through** in `handleSendAction`, so audio sent with `asVoice: true` is always delivered as a file attachment instead of a voice message.
### Root Cause
In `message-action-runner-CkqyWfJZ.js`, the `h... | closed | completed | true | 2 | [] | [] | 2026-04-28T09:24:04Z | 2026-05-03T04:34:11Z | 2026-05-01T02:47:10Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Bowerai | 17,424,365 | MDQ6VXNlcjE3NDI0MzY1 | User | false |
openclaw/openclaw | 4,342,102,092 | I_kwDOQb6kR88AAAABAs84TA | 73,461 | https://github.com/openclaw/openclaw/issues/73461 | https://api.github.com/repos/openclaw/openclaw/issues/73461 | [Bug]: Cerebras provider page is missing from the docs navigation | ## Summary
The Cerebras provider documentation page exists and is linked from the provider index, but `providers/cerebras` is missing from the Mintlify docs navigation provider list.
## Steps to reproduce
1. Inspect `docs/providers/cerebras.md`.
2. Inspect the Providers group in `docs/docs.json`.
3. Observe that `pr... | closed | completed | false | 3 | [] | [] | 2026-04-28T09:41:06Z | 2026-05-03T19:35:40Z | 2026-05-03T19:35:40Z | NONE | null | 20260503T210034Z | 2026-05-03T21:00:34Z | WadydX | 65,117,428 | MDQ6VXNlcjY1MTE3NDI4 | User | false |
openclaw/openclaw | 4,342,097,284 | I_kwDOQb6kR88AAAABAs8lhA | 73,457 | https://github.com/openclaw/openclaw/issues/73457 | https://api.github.com/repos/openclaw/openclaw/issues/73457 | openclaw update does not update systemd service path after version upgrade | ## Bug Description
After running `openclaw update`, the npm package is updated to the new version, but the systemd service file (`~/.config/systemd/user/openclaw-gateway.service`) still contains the hardcoded old version path, causing the gateway to fail to start.
## Steps to Reproduce
1. Install openclaw (e.g., ver... | closed | completed | true | 1 | [] | [] | 2026-04-28T09:40:14Z | 2026-05-01T04:44:02Z | 2026-04-28T09:45:49Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | YuchenNing | 36,885,113 | MDQ6VXNlcjM2ODg1MTEz | User | false |
openclaw/openclaw | 4,342,006,667 | I_kwDOQb6kR88AAAABAs3Diw | 73,450 | https://github.com/openclaw/openclaw/issues/73450 | https://api.github.com/repos/openclaw/openclaw/issues/73450 | [Bug]: Telegram /new can trigger empty Codex Responses request and expose provider error | ### Bug type
Regression / channel runtime error
### Summary
On OpenClaw `2026.4.26`, sending `/new` in a Telegram direct chat can reset/create the session but then trigger an empty Codex Responses API request. The raw provider error is delivered back to Telegram:
```text
One of "input" or "previous_response_id" or ... | closed | completed | true | 1 | [] | [] | 2026-04-28T09:25:13Z | 2026-05-01T04:44:01Z | 2026-04-28T09:45:52Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | geegewu | 129,384,489 | U_kgDOB7ZAKQ | User | false |
openclaw/openclaw | 4,341,888,328 | I_kwDOQb6kR88AAAABAsv1SA | 73,437 | https://github.com/openclaw/openclaw/issues/73437 | https://api.github.com/repos/openclaw/openclaw/issues/73437 | [Feature]: `ContextEngine.assemble` should fire per LLM call, not per user prompt | ### Summary
`ContextEngine.assemble` should fire per LLM call, not per user prompt
#### Current behavior
User prompt
→ assembleAttemptContextEngine() ← fires ONCE
→ activeSession.prompt() ← agentic loop
LLM call 1 ← ✅ assemble ran before this
→ tool result
→ LLM call 2 ← ❌ no as... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-28T09:06:01Z | 2026-05-03T12:31:47Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | HaotianChen616 | 52,230,263 | MDQ6VXNlcjUyMjMwMjYz | User | false |
openclaw/openclaw | 4,342,127,178 | I_kwDOQb6kR88AAAABAs-aSg | 73,464 | https://github.com/openclaw/openclaw/issues/73464 | https://api.github.com/repos/openclaw/openclaw/issues/73464 | [Bug] Webchat messages delivered twice — duplicate replies on same session | ## Bug Description
When sending messages via webchat (Chrome browser), the same message is sometimes delivered to the agent twice, resulting in two identical replies appearing in the chat UI.
## Steps to Reproduce
1. Open OpenClaw webchat in Chrome
2. Send a message (e.g., any plain text query)
3. Observe: two identic... | closed | completed | true | 1 | [] | [] | 2026-04-28T09:45:19Z | 2026-05-01T04:44:00Z | 2026-04-28T09:49:44Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | zyzsen | 189,302,767 | U_kgDOC0iH7w | User | false |
openclaw/openclaw | 4,342,164,571 | I_kwDOQb6kR88AAAABAtAsWw | 73,470 | https://github.com/openclaw/openclaw/issues/73470 | https://api.github.com/repos/openclaw/openclaw/issues/73470 | Feature: Mandatory completeness scan guardrail for read/search operations | null | open | null | false | 1 | [] | [] | 2026-04-28T09:51:33Z | 2026-05-03T01:38:22Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | cjboy007 | 64,246,343 | MDQ6VXNlcjY0MjQ2MzQz | User | false |
openclaw/openclaw | 4,342,165,799 | I_kwDOQb6kR88AAAABAtAxJw | 73,471 | https://github.com/openclaw/openclaw/issues/73471 | https://api.github.com/repos/openclaw/openclaw/issues/73471 | [Bug] doctor `--fix` archives historical session transcripts as "orphans," silently losing chat history; recommend CI hardening of every "Run openclaw doctor" recommendation | # [Bug] doctor `--fix` archives historical session transcripts as "orphans," causing silent loss of conversational history; recommend CI hardening of every "Run openclaw doctor" recommendation
## Summary
`openclaw doctor --fix` classifies all historical primary session transcripts (`<uuid>.jsonl`) as "orphans" and of... | open | null | false | 1 | [] | [] | 2026-04-28T09:51:46Z | 2026-05-03T08:39:54Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | injinj | 14,224,432 | MDQ6VXNlcjE0MjI0NDMy | User | false |
openclaw/openclaw | 4,342,169,331 | I_kwDOQb6kR88AAAABAtA-8w | 73,472 | https://github.com/openclaw/openclaw/issues/73472 | https://api.github.com/repos/openclaw/openclaw/issues/73472 | Session repair (repairSessionFileIfNeeded) does not sanitize empty text blocks in user messages | ## Bug Description
`repairSessionFileIfNeeded` in `session-file-repair.ts` only checks **assistant** messages for empty content (`isAssistantEntryWithEmptyContent`), but does not sanitize **user** messages containing empty text content blocks like `{"type": "text", "text": ""}`.
When such a block exists in a user mes... | closed | completed | true | 1 | [] | [] | 2026-04-28T09:52:22Z | 2026-05-01T04:43:59Z | 2026-04-28T09:56:55Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | waiT-Chi | 49,801,847 | MDQ6VXNlcjQ5ODAxODQ3 | User | false |
openclaw/openclaw | 4,342,172,704 | I_kwDOQb6kR88AAAABAtBMIA | 73,473 | https://github.com/openclaw/openclaw/issues/73473 | https://api.github.com/repos/openclaw/openclaw/issues/73473 | Image tool broken after upgrading to 2026.4.26 - Failed to optimize image | ## Problem
After upgrading from 2026.4.23 to 2026.4.26, the built-in `image` tool is completely broken.
## Steps to reproduce
1. Send an image via Control UI (webchat) attachment or paste
2. Try to use the `image` tool to analyze it
3. Error: `Failed to optimize image`
## Environment
- OpenClaw: 2026.4.26
- Node: v2... | closed | not_planned | true | 1 | [] | [] | 2026-04-28T09:52:53Z | 2026-05-01T04:43:59Z | 2026-04-28T09:57:36Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | tonyxiaogui-cpu | 266,668,221 | U_kgDOD-UIvQ | User | false |
openclaw/openclaw | 4,342,043,788 | I_kwDOQb6kR88AAAABAs5UjA | 73,454 | https://github.com/openclaw/openclaw/issues/73454 | https://api.github.com/repos/openclaw/openclaw/issues/73454 | [Bug] 2026.4.26 regression: bundled deps fix works but 14.5 min startup and Telegram stalls persist (follow-up to #72967) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After updating from 2026.4.22 to 2026.4.26, `doctor --fix` successfully installs bundled plugin deps in 9s but the gateway startup degrades from ~1 minute to ~14.5 minutes, Telegram polling stalls for 138+ seconds, and cron.... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-28T09:31:48Z | 2026-05-01T04:43:58Z | 2026-04-28T09:40:07Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | centralpc | 59,361,530 | MDQ6VXNlcjU5MzYxNTMw | User | false |
openclaw/openclaw | 4,342,198,508 | I_kwDOQb6kR88AAAABAtCw7A | 73,475 | https://github.com/openclaw/openclaw/issues/73475 | https://api.github.com/repos/openclaw/openclaw/issues/73475 | Anthropic API Migration Required: Sonnet 4.6 Model Name and Header Changes | ## Migration Required for Anthropic Sonnet 4.6
Anthropic is releasing Sonnet 4.6 with API changes that will affect OpenClaw's Anthropic provider.
### Changes Required:
1. **Model string format change**:
- Current: `anthropic/claude-sonnet-4-20250514`
- New: `claude-sonnet-4-6-[version]`
2. **Remove beta head... | closed | not_planned | true | 1 | [] | [] | 2026-04-28T09:57:25Z | 2026-05-03T04:35:37Z | 2026-04-30T18:16:30Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | MrCerq | 188,248,728 | U_kgDOCzhymA | User | false |
openclaw/openclaw | 4,342,101,634 | I_kwDOQb6kR88AAAABAs82gg | 73,459 | https://github.com/openclaw/openclaw/issues/73459 | https://api.github.com/repos/openclaw/openclaw/issues/73459 | [Bug]: openclaw migrate --no-backup is parsed but ignored | ## Summary
`openclaw migrate` and `openclaw migrate apply` register `--no-backup`, but the CLI action reads `opts.backup === false` instead of `opts.noBackup`, so the apply path still creates a pre-migration backup.
## Steps to reproduce
1. Run `openclaw migrate apply hermes --yes --force --no-backup`.
2. Observe th... | closed | completed | true | 2 | [] | [] | 2026-04-28T09:41:01Z | 2026-05-01T04:43:57Z | 2026-04-28T10:07:26Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | WadydX | 65,117,428 | MDQ6VXNlcjY1MTE3NDI4 | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.