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 | 4,160,823,532 | I_kwDOQb6kR874ASDs | 56,516 | https://github.com/openclaw/openclaw/issues/56516 | https://api.github.com/repos/openclaw/openclaw/issues/56516 | Subagent sessions persist in web UI after completion | ## Problem
When using `sessions_spawn` with `mode: "run"`, subagent sessions are correctly marked as "done" and `cleanupHandled: true` in `runs.json`. However, these completed subagent sessions continue to appear in the web UI session list.
The `openclaw sessions cleanup` command only prunes the agent session store (... | closed | not_planned | true | 1 | [] | [] | 2026-03-28T16:51:56Z | 2026-05-02T04:28:47Z | 2026-04-29T09:06:04Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | arbaleast | 57,560,360 | MDQ6VXNlcjU3NTYwMzYw | User | false |
openclaw/openclaw | 4,181,677,283 | I_kwDOQb6kR875P1Tj | 58,534 | https://github.com/openclaw/openclaw/issues/58534 | https://api.github.com/repos/openclaw/openclaw/issues/58534 | Session management performance degrades severely with subagent usage (100%+ CPU at ~400 sessions) | ## Problem
OpenClaw natively supports and encourages subagent spawning for task parallelism, but session accumulation severely impacts host performance, forcing users to sacrifice data retention for operational stability.
## Real-World Metrics (from our deployment)
### Before (447 sessions accumulated over ~3 days)
... | closed | not_planned | true | 1 | [] | [] | 2026-03-31T19:10:39Z | 2026-05-02T04:28:46Z | 2026-04-29T09:06:08Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | lucca-alma | 258,741,130 | U_kgDOD2wTig | User | false |
openclaw/openclaw | 4,349,188,529 | I_kwDOQb6kR88AAAABAztZsQ | 74,239 | https://github.com/openclaw/openclaw/issues/74239 | https://api.github.com/repos/openclaw/openclaw/issues/74239 | [Bug]: Bodyless Venice/Anthropic 400 misclassified as context-overflow → Pi triggers compaction recovery loop on sessions with nothing to compact | ## Environment
- **OpenClaw:** 2026.4.25 (aa36ee6)
- **Model:** `venice/claude-sonnet-4-6` (Anthropic Claude Sonnet 4.6 via Venice proxy)
- **Context window:** 200,000 tokens (as reported by Venice discovery)
- **Platform:** macOS Darwin 25.4.0 arm64, Node v22.22.2
- **Compaction config:** `mode: safeguard`, `reserveT... | closed | completed | true | 1 | [] | [] | 2026-04-29T09:02:42Z | 2026-05-02T04:28:43Z | 2026-04-29T09:08:20Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | axonrelaybot | 275,827,808 | U_kgDOEHDMYA | User | false |
openclaw/openclaw | 4,158,878,783 | I_kwDOQb6kR87343Q_ | 56,312 | https://github.com/openclaw/openclaw/issues/56312 | https://api.github.com/repos/openclaw/openclaw/issues/56312 | Telegram video files not being downloaded to media/inbound directory | ## Bug Description
When a video file is sent via Telegram to the bot, the message is received and logged by OpenClaw, but the actual video file is NOT downloaded to the \`media/inbound\` directory.
## Environment
- OpenClaw version: 2026.3.24
- OS: macOS (Darwin 25.4.0)
- Channel: Telegram
- Model: minimax/MiniMax-M... | closed | not_planned | true | 1 | [] | [] | 2026-03-28T09:20:36Z | 2026-05-02T04:28:42Z | 2026-04-29T09:08:41Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | ThisisPeggy | 213,046,103 | U_kgDODLLTVw | User | false |
openclaw/openclaw | 4,144,785,173 | I_kwDOQb6kR873DGcV | 55,231 | https://github.com/openclaw/openclaw/issues/55231 | https://api.github.com/repos/openclaw/openclaw/issues/55231 | delivery-recovery flushes stale pending messages as batch on every restart | ## Summary
When the gateway restarts (for any reason), `delivery-recovery` flushes ALL pending delivery entries at once. This causes stale transcript echo messages — sometimes hours old — to arrive as a batch in the user's chat, making it look like the bot is spamming.
## Reproduction
Observed on v2026.3.24 stable. ... | closed | not_planned | true | 1 | [] | [] | 2026-03-26T15:00:52Z | 2026-05-02T04:28:41Z | 2026-04-29T09:10:08Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | coolmanns | 117,685,070 | U_kgDOBwO7Tg | User | false |
openclaw/openclaw | 4,349,208,913 | I_kwDOQb6kR88AAAABAzupUQ | 74,241 | https://github.com/openclaw/openclaw/issues/74241 | https://api.github.com/repos/openclaw/openclaw/issues/74241 | [Bug] Agent timezone desync causes perception drift in multi-timezone setups | ## Bug: Agent timezone desync — container clock vs user local time
**Priority:** MEDIUM
**Problem:**
The OpenClaw agent runs inside a Docker container with its own system timezone (CEST/Europe). The user operates in a different timezone (BST/UK = UTC+1). This creates a **permanent 1-2 hour perception gap**.
**Real e... | closed | not_planned | true | 3 | [] | [] | 2026-04-29T09:06:17Z | 2026-05-03T04:35:03Z | 2026-04-30T20:37:58Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | AS76 | 22,101,267 | MDQ6VXNlcjIyMTAxMjY3 | User | false |
openclaw/openclaw | 4,349,272,960 | I_kwDOQb6kR88AAAABAzyjgA | 74,248 | https://github.com/openclaw/openclaw/issues/74248 | https://api.github.com/repos/openclaw/openclaw/issues/74248 | [Bug]: Discord image attachments fail when sharp is missing from package runtime | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Discord outbound image attachments fail with `Failed to optimize image` when `sharp` is not installed, because the Discord media send path always optimizes image `MEDIA:` attachments and the published OpenClaw ... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T09:17:12Z | 2026-05-02T04:28:39Z | 2026-04-29T09:21:53Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | hoyanhan | 45,415,044 | MDQ6VXNlcjQ1NDE1MDQ0 | User | false |
openclaw/openclaw | 4,349,257,192 | I_kwDOQb6kR88AAAABAzxl6A | 74,245 | https://github.com/openclaw/openclaw/issues/74245 | https://api.github.com/repos/openclaw/openclaw/issues/74245 | [Bug]: deepseek extension missing provider-policy-api.ts — contextWindow and cost default to wrong values | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Deepseek models have no provider-policy-api.ts, so materializeRuntimeConfig fills contextWindow with DEFAULT_CONTEXT_TOKENS (200k) and cost with zeros. v4-flash/v4-pro should be 1M with correct cost like other ... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-29T09:14:32Z | 2026-05-02T04:28:17Z | 2026-04-29T12:38:13Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | moonlightb | 39,666,627 | MDQ6VXNlcjM5NjY2NjI3 | User | false |
openclaw/openclaw | 4,349,305,086 | I_kwDOQb6kR88AAAABAz0g_g | 74,249 | https://github.com/openclaw/openclaw/issues/74249 | https://api.github.com/repos/openclaw/openclaw/issues/74249 | Per-channel timeout configuration (telegram/discord/webchat/cron) | ## Problem
Today `agents.defaults.timeoutSeconds` is a **single global knob** that applies to all interactive channels (Telegram, Webchat, Discord) plus is the default for cron jobs without explicit per-job override.
This forces a single trade-off across very different latency expectations:
| Channel | User expectat... | open | null | false | 1 | [] | [] | 2026-04-29T09:22:13Z | 2026-05-03T03:37:55Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | fertilejim | 4,189,272 | MDQ6VXNlcjQxODkyNzI= | User | false |
openclaw/openclaw | 4,349,388,652 | I_kwDOQb6kR88AAAABAz5nbA | 74,257 | https://github.com/openclaw/openclaw/issues/74257 | https://api.github.com/repos/openclaw/openclaw/issues/74257 | [Bug]: Telegram async exec followup leaks HEARTBEAT_OK/internal text after context overflow | ## Bug type
Behavior bug (incorrect output/state without crash)
## Beta release blocker
No
## Summary
In OpenClaw 2026.4.26, a Telegram topic session that hit context overflow during an approved exec/tool flow leaked heartbeat/internal async-completion text into the user-visible chat and then produced expired appr... | open | null | false | 1 | [] | [] | 2026-04-29T09:34:36Z | 2026-05-03T08:12:01Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Staruy | 2,365,702 | MDQ6VXNlcjIzNjU3MDI= | User | false |
openclaw/openclaw | 4,349,415,359 | I_kwDOQb6kR88AAAABAz7Pvw | 74,259 | https://github.com/openclaw/openclaw/issues/74259 | https://api.github.com/repos/openclaw/openclaw/issues/74259 | Update QA lab parity gate for GPT-5.5 vs Opus 4.7 and harden preflight | ## Problem
The QA lab parity gate and related tests are only partially updated for the current frontier model targets as of 2026-04-29. The OpenAI candidate lane has moved to `openai/gpt-5.5`, but several workflow names, artifact paths, mock-model fixtures, reports, scenario docs, and the Anthropic baseline still enco... | closed | duplicate | true | 2 | [] | [] | 2026-04-29T09:38:06Z | 2026-05-02T04:28:37Z | 2026-04-29T09:41:30Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | electricsheepops | 276,317,342 | U_kgDOEHhEng | User | false |
openclaw/openclaw | 4,349,437,446 | I_kwDOQb6kR88AAAABAz8mBg | 74,262 | https://github.com/openclaw/openclaw/issues/74262 | https://api.github.com/repos/openclaw/openclaw/issues/74262 | Update QA lab parity gate for GPT-5.5 vs Opus 4.7 and harden preflight | ## Problem
The QA lab parity gate and related tests are only partially updated for the current frontier model targets as of 2026-04-29. The OpenAI candidate lane has moved to `openai/gpt-5.5`, but several workflow names, artifact paths, mock-model fixtures, reports, scenario docs, and the Anthropic baseline still enco... | open | null | false | 1 | [] | [] | 2026-04-29T09:41:07Z | 2026-05-03T06:07:58Z | null | CONTRIBUTOR | null | 20260503T090142Z | 2026-05-03T09:01:42Z | 100yenadmin | 239,388,517 | U_kgDODkTHZQ | User | false |
openclaw/openclaw | 4,349,446,843 | I_kwDOQb6kR88AAAABAz9Kuw | 74,264 | https://github.com/openclaw/openclaw/issues/74264 | https://api.github.com/repos/openclaw/openclaw/issues/74264 | feat(memory): shared embedding cache across agents indexing same workspace — eliminates duplicate API calls | ## Summary
When multiple agents share the same workspace (e.g. `sonnet-working` and `opus-strategic`), identical document chunks are independently indexed and embedded in each agent's SQLite database, resulting in duplicate embedding API calls and wasted storage.
## Evidence (OpenClaw 2026.4.25, audit 2026-04-28)
Fr... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T09:42:30Z | 2026-05-02T04:28:37Z | 2026-04-29T09:46:39Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | chrismclarnon-ep | 275,793,450 | U_kgDOEHBGKg | User | false |
openclaw/openclaw | 4,349,205,471 | I_kwDOQb6kR88AAAABAzub3w | 74,240 | https://github.com/openclaw/openclaw/issues/74240 | https://api.github.com/repos/openclaw/openclaw/issues/74240 | 2026.4.27: Gateway main event loop blocks at 100% CPU on Telegram message due to repeated filesystem/runtime-deps work | ## Summary
After upgrading an OpenClaw Docker install to `2026.4.27`, each Telegram message can trigger a gateway main-thread CPU spike. During the spike the gateway becomes partially unresponsive: `/healthz` times out, Control UI websocket requests stall/disconnect, and Telegram replies are delayed.
The model run it... | closed | completed | true | 3 | [] | [] | 2026-04-29T09:05:40Z | 2026-05-02T04:25:42Z | 2026-04-29T09:10:07Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | crash2kx | 11,518,960 | MDQ6VXNlcjExNTE4OTYw | User | false |
openclaw/openclaw | 3,912,595,916 | I_kwDOQb6kR87pNXnM | 11,829 | https://github.com/openclaw/openclaw/issues/11829 | https://api.github.com/repos/openclaw/openclaw/issues/11829 | Security Roadmap: Protecting API Keys from Agent Access | ## Summary
OpenClaw currently has multiple vectors where API keys can leak to the LLM or be exposed in chat. This roadmap proposes a layered approach to comprehensively protect secrets.
## The Problem
1. **LLM Provider Keys** — Model catalog with resolved apiKey values is serialized into prompt context (#11202)
2. *... | open | null | false | 14 | [
"bug",
"security"
] | [] | 2026-02-08T11:52:12Z | 2026-04-29T09:52:56Z | null | NONE | null | 20260429T120040Z | 2026-04-29T12:00:40Z | jmkritt | 8,386,890 | MDQ6VXNlcjgzODY4OTA= | User | false |
openclaw/openclaw | 4,349,350,260 | I_kwDOQb6kR88AAAABAz3RdA | 74,251 | https://github.com/openclaw/openclaw/issues/74251 | https://api.github.com/repos/openclaw/openclaw/issues/74251 | [Bug]: [Bug]: npm install -g openclaw@latest crashes with V8 Fatal error in WSL2 Ubuntu 24.04 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
npm install -g openclaw@latest crashes with unreachable code / RepresentationChangerError during official install script in WSL2 Ubuntu 24.04.
### Steps to reproduce
1. Open WSL2 terminal (Ubuntu 24.04.4)
2. Run: curl -fsS... | open | null | false | 3 | [
"bug"
] | [] | 2026-04-29T09:29:14Z | 2026-05-03T04:44:05Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | wkerun | 189,182,099 | U_kgDOC0awkw | User | false |
openclaw/openclaw | 4,349,419,160 | I_kwDOQb6kR88AAAABAz7emA | 74,261 | https://github.com/openclaw/openclaw/issues/74261 | https://api.github.com/repos/openclaw/openclaw/issues/74261 | [Feature]: whatsapp create group tool for whatsapp-agent-tools | ### Summary
I need my AI Agent to create group for information storage only
### Problem to solve
As a property agent using OpenClaw on WhatsApp, I need my AI assistant to create WhatsApp groups automatically when a new property listing is confirmed.
Current whatsapp-agent-tools only has `whatsapp_send` — no group ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-29T09:38:35Z | 2026-05-03T04:44:06Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | adentan3131-lgtm | 273,148,919 | U_kgDOEEfr9w | User | false |
openclaw/openclaw | 4,349,572,159 | I_kwDOQb6kR88AAAABA0E0Pw | 74,268 | https://github.com/openclaw/openclaw/issues/74268 | https://api.github.com/repos/openclaw/openclaw/issues/74268 | Wake up — 92 hardcoded paths, 1.2GB of node_modules nobody asked for, and a token leak waiting to happen | Closed — was created in error. The actual PR about the Smart Router is separate. | closed | not_planned | true | 0 | [] | [] | 2026-04-29T10:01:10Z | 2026-05-02T04:28:35Z | 2026-04-29T10:04:23Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | gutchapa | 209,160,664 | U_kgDODHeJ2A | User | false |
openclaw/openclaw | 4,349,603,101 | I_kwDOQb6kR88AAAABA0GtHQ | 74,270 | https://github.com/openclaw/openclaw/issues/74270 | https://api.github.com/repos/openclaw/openclaw/issues/74270 | Smart Router: local-semantic triage layer cuts cloud cost from $5/day to pennies — with privacy override | Closed. Posted without understanding the full history correctly. | closed | not_planned | true | 0 | [] | [] | 2026-04-29T10:06:22Z | 2026-05-02T04:28:34Z | 2026-04-29T10:06:50Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | gutchapa | 209,160,664 | U_kgDODHeJ2A | User | false |
openclaw/openclaw | 4,349,593,113 | I_kwDOQb6kR88AAAABA0GGGQ | 74,269 | https://github.com/openclaw/openclaw/issues/74269 | https://api.github.com/repos/openclaw/openclaw/issues/74269 | [Bug]: Whatsapp starts to reply then stops and never sends | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Right after install and WhatsApp setup with QR code, openclaw is unable to reply on WhatsApp (works fine on Telegram), It starts replying with the 3 dots, then stops and never sends answer.
Tried to troubleshoo... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-29T10:04:43Z | 2026-05-03T04:44:07Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | tomcosta-git | 280,330,212 | U_kgDOELV_5A | User | false |
openclaw/openclaw | 4,349,628,802 | I_kwDOQb6kR88AAAABA0IRgg | 74,271 | https://github.com/openclaw/openclaw/issues/74271 | https://api.github.com/repos/openclaw/openclaw/issues/74271 | [Feature]: doctor/status should warn when OPENCLAW_GATEWAY_TOKEN in ~/.openclaw/.env overrides gateway.auth.token | ### Summary
Warn users when a stale OPENCLAW_GATEWAY_TOKEN in ~/.openclaw/.env overrides gateway.auth.token in local gateway mode.
### Problem to solve
After updating OpenClaw, my local gateway entered a confusing split-brain auth state.
I had gateway.auth.token configured in ~/.openclaw/openclaw.json, but ~/.openc... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-04-29T10:10:32Z | 2026-05-03T03:37:56Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | grzesiolpl | 95,895,603 | U_kgDOBbdAMw | User | false |
openclaw/openclaw | 4,349,643,717 | I_kwDOQb6kR88AAAABA0JLxQ | 74,272 | https://github.com/openclaw/openclaw/issues/74272 | https://api.github.com/repos/openclaw/openclaw/issues/74272 | [Bug]: openclaw gateway restart takes extremely long on Windows | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Description
The openclaw gateway restart command takes an extremely long time to complete on Windows. The gateway process itself starts successfully (plugins load, channels connect), but the restart command keeps hanging and... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-29T10:13:05Z | 2026-05-02T04:28:33Z | 2026-04-29T10:17:53Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | zym951223 | 51,111,325 | MDQ6VXNlcjUxMTExMzI1 | User | false |
openclaw/openclaw | 4,349,715,579 | I_kwDOQb6kR88AAAABA0Nkew | 74,277 | https://github.com/openclaw/openclaw/issues/74277 | https://api.github.com/repos/openclaw/openclaw/issues/74277 | [Bug]: npm install -g openclaw@latest crashes with V8 Fatal error in WSL2 Ubuntu 24.04 | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
npm install -g openclaw@latest crashes with unreachable code / RepresentationChangerError during official install script in WSL2 Ubuntu 24.04.
### Steps to reproduce
1. Open WSL2 terminal (Ubuntu 24.04.4)
2. Run: curl -fsSL h... | closed | not_planned | true | 2 | [
"bug",
"bug:crash"
] | [] | 2026-04-29T10:24:49Z | 2026-05-02T04:28:33Z | 2026-04-29T10:29:12Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | wkerun | 189,182,099 | U_kgDOC0awkw | User | false |
openclaw/openclaw | 4,349,780,099 | I_kwDOQb6kR88AAAABA0Rggw | 74,279 | https://github.com/openclaw/openclaw/issues/74279 | https://api.github.com/repos/openclaw/openclaw/issues/74279 | Regression: Windows gateway probe/connect timeout returns in 2026.4.26 | This looks like a regression of the Windows gateway probe / connect-timeout class of failures that was previously marked fixed around 2026.3.22.
Related locked/closed issues:
- #45560: gateway probe timeout while gateway healthy
- #49726: connect challenge / handshake timeout under gateway load
Environment:
- Windows... | closed | completed | true | 2 | [] | [] | 2026-04-29T10:34:53Z | 2026-05-02T04:28:09Z | 2026-04-29T13:52:43Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | joost-heijden | 29,427,775 | MDQ6VXNlcjI5NDI3Nzc1 | User | false |
openclaw/openclaw | 4,349,819,170 | I_kwDOQb6kR88AAAABA0T5Ig | 74,280 | https://github.com/openclaw/openclaw/issues/74280 | https://api.github.com/repos/openclaw/openclaw/issues/74280 | [Bug]: openclaw gateway restart takes extremely long on Windows | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
This question still exist
Confirmed that the local installation includes the https://github.com/openclaw/openclaw/commit/c85065eb7f396557efdc407dd16ec01639f889e3 fix (connect failed is in the allowlist).
openclaw --version
O... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-29T10:40:37Z | 2026-05-02T04:28:03Z | 2026-04-29T14:26:15Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | zym951223 | 51,111,325 | MDQ6VXNlcjUxMTExMzI1 | User | false |
openclaw/openclaw | 4,349,252,758 | I_kwDOQb6kR88AAAABAzxUlg | 74,244 | https://github.com/openclaw/openclaw/issues/74244 | https://api.github.com/repos/openclaw/openclaw/issues/74244 | [Bug]: Gemini API error (400): missing `thought_signature` | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Sustaining chain of thought after tool call hangs up randomly, due to missing `thought_signature` in the api call succeeding the tool output
### Steps to reproduce
> Not fully deterministic as this is not a h... | closed | completed | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-04-29T09:13:46Z | 2026-05-02T04:27:07Z | 2026-04-29T09:23:29Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | YouFoundJK | 72,195,985 | MDQ6VXNlcjcyMTk1OTg1 | User | false |
openclaw/openclaw | 4,349,881,595 | I_kwDOQb6kR88AAAABA0Xs-w | 74,283 | https://github.com/openclaw/openclaw/issues/74283 | https://api.github.com/repos/openclaw/openclaw/issues/74283 | delivery-recovery: post-approval follow-up dumps raw internal handler source to channel (related to #72184) | Same symptom family as #72184 (closed Apr 27 as fixed in v2026.4.24), but a **different code path** that the previous fix does not cover. The v2026.4.24 fix was scoped to config.patch/config.apply restart sentinels only; this manifestation is in the `/approve` post-approval follow-up delivery hop.
Filing fresh because... | open | null | false | 2 | [] | [] | 2026-04-29T10:50:19Z | 2026-05-03T03:37:57Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | xonaman | 45,240,351 | MDQ6VXNlcjQ1MjQwMzUx | User | false |
openclaw/openclaw | 4,349,886,187 | I_kwDOQb6kR88AAAABA0X-6w | 74,284 | https://github.com/openclaw/openclaw/issues/74284 | https://api.github.com/repos/openclaw/openclaw/issues/74284 | [Bug]: heartbeat session at agent:main:main pins model from creation time, ignores openclaw.json edits across runs (follow-up to #51677) | ## Summary
The autonomous heartbeat session keyed `agent:main:main` continues to use the model that was current at the session's creation time, even after `agents.defaults.model.primary` is changed in `openclaw.json` and the gateway is restarted. #51677 closed the equivalent issue on the reply path; this report is a h... | open | null | false | 1 | [] | [] | 2026-04-29T10:51:04Z | 2026-05-03T06:08:00Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | brtkwr | 2,181,426 | MDQ6VXNlcjIxODE0MjY= | User | false |
openclaw/openclaw | 4,349,910,817 | I_kwDOQb6kR88AAAABA0ZfIQ | 74,285 | https://github.com/openclaw/openclaw/issues/74285 | https://api.github.com/repos/openclaw/openclaw/issues/74285 | Telegram: Assistant reply silently dropped after long audio transcription + subagent spawn | # Telegram: Assistant reply silently dropped after long audio transcription + subagent spawn
## Environment
- **OpenClaw Version:** 2026.4.26 (incident occurred on 2026.4.24/2026.4.25 in production VPS)
- **Channel:** Telegram DM (direct message, `streamMode: "partial"`)
- **Model:** claude-opus-4-7 (primary)
- **OS:... | closed | completed | true | 1 | [] | [] | 2026-04-29T10:54:42Z | 2026-05-04T04:37:00Z | 2026-05-01T09:44:10Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | lolaopenclaw | 259,362,584 | U_kgDOD3WPGA | User | false |
openclaw/openclaw | 4,349,932,805 | I_kwDOQb6kR88AAAABA0a1BQ | 74,286 | https://github.com/openclaw/openclaw/issues/74286 | https://api.github.com/repos/openclaw/openclaw/issues/74286 | Bug: sub-agent announce completion causes parent session to generate out-of-context replies |
## Bug Description
When a sub-agent or cron job with \`delivery.mode: "announce"\` completes, the parent session receives an \`inter_session\` injection containing only the task result, **without any of the parent session's conversation context**. This causes the parent session to generate replies that are completely... | open | null | false | 4 | [] | [] | 2026-04-29T10:57:45Z | 2026-05-03T07:23:32Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | kansjwqwq | 264,942,807 | U_kgDOD8q01w | User | false |
openclaw/openclaw | 4,349,838,279 | I_kwDOQb6kR88AAAABA0VDxw | 74,281 | https://github.com/openclaw/openclaw/issues/74281 | https://api.github.com/repos/openclaw/openclaw/issues/74281 | [Bug]: Custom watchdog kills healthy gateway during slow startup (2026.4.26 + 20 plugin deps) | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
**Component:** `openclaw-ops` skill / `watchdog.sh`
**Date:** April 29, 2026
The watchdog script at `~/.openclaw/skills/openclaw-ops/scripts/watchdog.sh` uses a 5-second HTTP health check timeout (`curl --max-time 5`). In ... | closed | not_planned | true | 2 | [
"bug",
"bug:crash"
] | [] | 2026-04-29T10:43:40Z | 2026-05-02T04:28:28Z | 2026-04-29T10:59:00Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | drmarcopapa | 39,249,407 | MDQ6VXNlcjM5MjQ5NDA3 | User | false |
openclaw/openclaw | 4,350,076,116 | I_kwDOQb6kR88AAAABA0jk1A | 74,296 | https://github.com/openclaw/openclaw/issues/74296 | https://api.github.com/repos/openclaw/openclaw/issues/74296 | msteams: jwt.verify undefined under Node ESM-CJS interop — every Bot Framework activity silently rejected | ## Summary
In `extensions/msteams/src/sdk.ts:716`, `loadBotFrameworkJwtDeps` does:
```ts
botFrameworkJwtDepsPromise ??= Promise.all([import("jsonwebtoken"), import("jwks-rsa")]).then(
([jwt, { JwksClient }]) => ({ jwt, JwksClient }),
);
```
`jsonwebtoken` is a CJS module. When dynamically imported under Node ESM, ... | closed | completed | true | 3 | [] | [
"vincentkoc"
] | 2026-04-29T11:17:43Z | 2026-05-02T04:28:14Z | 2026-04-29T13:09:33Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | pfw-dean | 110,184,914 | U_kgDOBpFJ0g | User | false |
openclaw/openclaw | 4,350,080,379 | I_kwDOQb6kR88AAAABA0j1ew | 74,298 | https://github.com/openclaw/openclaw/issues/74298 | https://api.github.com/repos/openclaw/openclaw/issues/74298 | Migration: rsync silently strips TELEGRAM_BOT_TOKEN/DISCORD_BOT_TOKEN from .env | # Migration: rsync excludes `TELEGRAM_BOT_TOKEN` and `DISCORD_BOT_TOKEN` from `.env` when using standard patterns
## Environment
- **OpenClaw Version:** 2026.4.24–2026.4.26
- **Migration Context:** BLUE/GREEN VPS deployment (Ubuntu 24.04 → Ubuntu 26.04)
- **Tool:** rsync with standard exclude patterns
## Summary
Wh... | closed | completed | true | 2 | [] | [] | 2026-04-29T11:18:23Z | 2026-05-04T04:34:14Z | 2026-05-02T03:29:48Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | lolaopenclaw | 259,362,584 | U_kgDOD3WPGA | User | false |
openclaw/openclaw | 4,350,080,575 | I_kwDOQb6kR88AAAABA0j2Pw | 74,300 | https://github.com/openclaw/openclaw/issues/74300 | https://api.github.com/repos/openclaw/openclaw/issues/74300 | Migration: version mismatch between source/target causes silent state incompatibility | # Migration: version mismatch between source and target environments causes silent state incompatibility
## Environment
- **Source (BLUE):** OpenClaw 2026.4.25
- **Target (GREEN):** OpenClaw 2026.4.26
- **Migration Type:** BLUE/GREEN VPS deployment
## Summary
When migrating OpenClaw between environments with differ... | open | null | false | 2 | [] | [] | 2026-04-29T11:18:26Z | 2026-05-03T04:44:09Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | lolaopenclaw | 259,362,584 | U_kgDOD3WPGA | User | false |
openclaw/openclaw | 4,350,080,484 | I_kwDOQb6kR88AAAABA0j15A | 74,299 | https://github.com/openclaw/openclaw/issues/74299 | https://api.github.com/repos/openclaw/openclaw/issues/74299 | Telegram plugin reports "connected" without actually polling (false positive) | # Telegram plugin reports "connected" without polling (false positive)
## Environment
- **OpenClaw Version:** 2026.4.25–2026.4.26
- **Channel:** Telegram
- **OS:** Ubuntu 24.04 / 26.04 LTS
## Summary
After a VPS migration, the OpenClaw gateway logs showed:
```
[telegram] Telegram bot connected (polling enabled)
``... | closed | completed | true | 3 | [] | [] | 2026-04-29T11:18:25Z | 2026-05-02T04:28:00Z | 2026-04-29T14:45:42Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | lolaopenclaw | 259,362,584 | U_kgDOD3WPGA | User | false |
openclaw/openclaw | 4,350,103,724 | I_kwDOQb6kR88AAAABA0lQrA | 74,304 | https://github.com/openclaw/openclaw/issues/74304 | https://api.github.com/repos/openclaw/openclaw/issues/74304 | [Feature]: Support before_reset hook for automatic daily session reset | ### Summary
Extend before_reset hook to fire on automatic daily session resets (4 AM), enabling session-memory to auto-save context before session wipe.
### Problem to solve
OpenClaw's automatic daily session reset at 4 AM silently wipes conversation context with no pre-reset hook. The built-in session-memory hook o... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-29T11:21:50Z | 2026-05-02T04:28:23Z | 2026-04-29T11:26:33Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | kaxiaoge | 179,785,897 | U_kgDOCrdQqQ | User | false |
openclaw/openclaw | 4,350,111,125 | I_kwDOQb6kR88AAAABA0ltlQ | 74,305 | https://github.com/openclaw/openclaw/issues/74305 | https://api.github.com/repos/openclaw/openclaw/issues/74305 | [Bug]: ACPX Codex worker fails when model/thinking overrides are configured | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
Codex ACP worker runs succeed without internal model/thinking overrides, but fail with AcpRuntimeError: Internal error when model and model_reasoning_effort overrides are passed to codex-acp.
### Steps to reproduce
1. Start O... | open | null | false | 2 | [
"bug",
"bug:crash"
] | [] | 2026-04-29T11:22:52Z | 2026-05-03T01:23:11Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | SimSef | 56,121,054 | MDQ6VXNlcjU2MTIxMDU0 | User | false |
openclaw/openclaw | 4,350,160,379 | I_kwDOQb6kR88AAAABA0ot-w | 74,307 | https://github.com/openclaw/openclaw/issues/74307 | https://api.github.com/repos/openclaw/openclaw/issues/74307 | [Bug]: doctor --fix recreates plugin-runtime-deps plugin-sdk as directory instead of symlink on npm/macOS install | ### Bug type
Incorrect behavior without a crash
### Summary
On a global npm install on macOS, `openclaw doctor --fix` does not converge the bundled runtime-deps repair for `plugin-sdk`.
If `~/.openclaw/plugin-runtime-deps/<version>/dist/extensions/node_modules/openclaw/plugin-sdk` is moved aside before the repair, ... | closed | completed | true | 2 | [] | [] | 2026-04-29T11:29:57Z | 2026-05-02T04:26:25Z | 2026-04-30T03:14:34Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | Skeptomenos | 22,348,986 | MDQ6VXNlcjIyMzQ4OTg2 | User | false |
openclaw/openclaw | 4,350,173,843 | I_kwDOQb6kR88AAAABA0pikw | 74,309 | https://github.com/openclaw/openclaw/issues/74309 | https://api.github.com/repos/openclaw/openclaw/issues/74309 | [Feature Request] Show actual model identifier per reply when fallback or model switch occurs | ## Problem
When a model request fails (rate limit, billing error, timeout, etc.) and OpenClaw silently falls back to another provider/model, the user has no way to know which model actually generated the reply. This is especially problematic when:
- The primary model is down and a much weaker fallback model responds ... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T11:31:53Z | 2026-05-02T04:28:22Z | 2026-04-29T11:36:25Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | cloudpaw5216 | 267,859,441 | U_kgDOD_c18Q | User | false |
openclaw/openclaw | 4,350,217,130 | I_kwDOQb6kR88AAAABA0sLqg | 74,310 | https://github.com/openclaw/openclaw/issues/74310 | https://api.github.com/repos/openclaw/openclaw/issues/74310 | Bug Report: `normalizeProviderId` breaks provider-namespaced models like `kimi-coding/k2p5` | # Bug Report: `normalizeProviderId` breaks provider-namespaced models like `kimi-coding/k2p5`
**Component:** Core model resolution (`src/agents/provider-id.ts`)
**Severity:** Medium — breaks model selection for users with custom provider namespaces
**Affected versions:** 2026.4.26 (and likely earlier)
## Summary
`no... | open | null | false | 2 | [] | [] | 2026-04-29T11:38:36Z | 2026-05-03T01:23:12Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | vokasug | 280,127,440 | U_kgDOELJn0A | User | false |
openclaw/openclaw | 4,350,259,526 | I_kwDOQb6kR88AAAABA0uxRg | 74,312 | https://github.com/openclaw/openclaw/issues/74312 | https://api.github.com/repos/openclaw/openclaw/issues/74312 | claude-cli auth-epoch flips on token rotation, forcing session resets mid-conversation | ## Summary
On macOS, long-running Claude CLI agent sessions are hard-reset every few minutes during active conversations. The reset is triggered by `cli session reset: reason=auth-epoch` whenever Claude CLI rotates its OAuth tokens (which it does periodically and especially aggressively near token expiry). The user's ... | open | null | false | 2 | [] | [] | 2026-04-29T11:44:49Z | 2026-05-03T01:23:13Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | aderius | 78,624,045 | MDQ6VXNlcjc4NjI0MDQ1 | User | false |
openclaw/openclaw | 4,350,080,174 | I_kwDOQb6kR88AAAABA0j0rg | 74,297 | https://github.com/openclaw/openclaw/issues/74297 | https://api.github.com/repos/openclaw/openclaw/issues/74297 | auth-profiles.json: `type: "token"` breaks Anthropic Enterprise accounts (529 errors) | # auth-profiles.json: `type: "token"` breaks Anthropic Enterprise accounts (529 errors)
## Environment
- **OpenClaw Version:** 2026.4.25
- **Provider:** Anthropic (Enterprise API account)
- **OS:** Ubuntu 24.04 LTS
## Summary
When `~/.openclaw/agents/main/agent/auth-profiles.json` contains:
```json
{
"anthropic:... | open | null | false | 1 | [] | [] | 2026-04-29T11:18:22Z | 2026-05-03T04:44:08Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | lolaopenclaw | 259,362,584 | U_kgDOD3WPGA | User | false |
openclaw/openclaw | 4,350,271,844 | I_kwDOQb6kR88AAAABA0vhZA | 74,313 | https://github.com/openclaw/openclaw/issues/74313 | https://api.github.com/repos/openclaw/openclaw/issues/74313 | [Bug]: doctor flags live agents/main store as orphaned when default agent id is not main | ### Bug type
Incorrect behavior without a crash
### Summary
`openclaw doctor` can flag `~/.openclaw/agents/main` as an orphaned agent directory and suggest removing it, even though the same install still uses `agents/main` for live auth/session-related state.
This is dangerous because the warning is framed as clean... | open | null | false | 6 | [] | [] | 2026-04-29T11:46:53Z | 2026-05-03T09:49:35Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | Skeptomenos | 22,348,986 | MDQ6VXNlcjIyMzQ4OTg2 | User | false |
openclaw/openclaw | 4,350,026,783 | I_kwDOQb6kR88AAAABA0gkHw | 74,292 | https://github.com/openclaw/openclaw/issues/74292 | https://api.github.com/repos/openclaw/openclaw/issues/74292 | active-memory embedded sub-agent WebSocket handshake timeout (all models affected) | **Environment:**
- OpenClaw: 2026.4.26 (be8c246)
- OS: macOS 15.2 (arm64, Mac mini M2)
- Node: v24.14.0
- Config: 6 agents, active-memory enabled for all, models tested: GLM-5.1 / GLM-5-Turbo / DeepSeek V4 Flash
**Description:**
The `active-memory` plugin embedded sub-agent consistently times out (~17-45s) regardless... | closed | completed | true | 3 | [] | [] | 2026-04-29T11:10:38Z | 2026-05-02T04:28:05Z | 2026-04-29T14:12:09Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | hfgwq2dgx8-tech | 279,440,970 | U_kgDOEKfuSg | User | false |
openclaw/openclaw | 4,350,303,764 | I_kwDOQb6kR88AAAABA0xeFA | 74,316 | https://github.com/openclaw/openclaw/issues/74316 | https://api.github.com/repos/openclaw/openclaw/issues/74316 | media-understanding: `MediaFetchError` on Discord audio fetch is given up after 1 attempt — no retry | ## Summary
When the `media-understanding` subsystem fails to fetch a Discord audio attachment (`MediaFetchError`), it gives up after **a single attempt** with no retry. The transcription pipeline (Groq Whisper API → MLX-whisper local fallback) never gets a chance — the audio file was never downloaded.
User impact: vo... | open | null | false | 2 | [] | [] | 2026-04-29T11:51:59Z | 2026-05-04T02:38:31Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | gabrielexito-stack | 263,964,070 | U_kgDOD7vFpg | User | false |
openclaw/openclaw | 4,350,274,235 | I_kwDOQb6kR88AAAABA0vquw | 74,314 | https://github.com/openclaw/openclaw/issues/74314 | https://api.github.com/repos/openclaw/openclaw/issues/74314 | [Bug]: google-antigravity provider broken: API deprecated version 1.21.9, internal token refresh corrupts credential cache | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
... | open | null | false | 1 | [
"bug"
] | [] | 2026-04-29T11:47:18Z | 2026-05-03T12:31:59Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | ponsikk | 156,168,613 | U_kgDOCU7xpQ | User | false |
openclaw/openclaw | 4,350,324,239 | I_kwDOQb6kR88AAAABA0yuDw | 74,318 | https://github.com/openclaw/openclaw/issues/74318 | https://api.github.com/repos/openclaw/openclaw/issues/74318 | Integration: Adding prompt injection protection and audit logging to OpenClaw with Ombre[Feature]: | ### Summary
Add Ombre as an official security and audit integration for OpenClaw
### Problem to solve
OpenClaw has documented prompt injection vulnerabilities and
469 open security issues. SECURITY.md lists prompt injection
as out of scope. With 5,700+ community skills, every skill
install is a potential attack s... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-29T11:55:30Z | 2026-05-02T04:28:18Z | 2026-04-29T12:06:38Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | pypl0 | 277,652,227 | U_kgDOEIyjAw | User | false |
openclaw/openclaw | 4,350,445,739 | I_kwDOQb6kR88AAAABA06Iqw | 74,325 | https://github.com/openclaw/openclaw/issues/74325 | https://api.github.com/repos/openclaw/openclaw/issues/74325 | [Bug]: gateway restart blocks main thread for ~75s — `sidecars.channels` takes 45s, then a second ~28s freeze right after "gateway ready" | ### Bug type
Behavior bug
### Summary
After every \`openclaw gateway restart\` on 2026.4.26, the gateway is unresponsive for ~75 seconds total, in two distinct synchronous phases visible from logs and from a plugin-side event-loop-lag monitor:
1. **Pre-ready (~45s)**: between \`starting channels and sidecars...\` a... | open | null | false | 2 | [] | [] | 2026-04-29T12:15:33Z | 2026-05-03T10:19:58Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | xiaohuaxi | 45,277,353 | MDQ6VXNlcjQ1Mjc3MzUz | User | false |
openclaw/openclaw | 4,350,555,355 | I_kwDOQb6kR88AAAABA1A02w | 74,333 | https://github.com/openclaw/openclaw/issues/74333 | https://api.github.com/repos/openclaw/openclaw/issues/74333 | v2026.4.24: Agent's `memory_search` tool never invokes `recordShortTermRecalls`; only the human CLI does, leaving `recallCount=0` for all agent-driven recalls | ## Summary
`recordShortTermRecalls` is the function that increments per-entry recall counters in the short-term recall store (`~/.openclaw/workspace-<agent>/memory/.dreams/short-term-recall.json`). Those counters feed the dreaming-promotion ranker, which surfaces frequently-recalled snippets for write-up into the agen... | closed | completed | true | 2 | [] | [] | 2026-04-29T12:34:03Z | 2026-05-02T04:26:51Z | 2026-04-29T12:38:30Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | haumanto | 72,009,799 | MDQ6VXNlcjcyMDA5Nzk5 | User | false |
openclaw/openclaw | 4,350,553,752 | I_kwDOQb6kR88AAAABA1AumA | 74,332 | https://github.com/openclaw/openclaw/issues/74332 | https://api.github.com/repos/openclaw/openclaw/issues/74332 | v2026.4.24: Plugin synthetic operator client lacks `operator.admin` scope, breaking `subagent.deleteSession` for memory-core dreaming narrative cleanup | ## Summary
The plugin synthetic operator client created by `server-plugin-bootstrap` defaults to the scope set `["operator.write"]`. Several first-party gateway methods that plugins invoke through `dispatchGatewayMethod` — including `sessions.delete`, `sessions.reset`, `sessions.patch`, and `sessions.compact` — requir... | closed | completed | true | 2 | [] | [] | 2026-04-29T12:33:50Z | 2026-05-02T04:26:53Z | 2026-04-29T12:38:57Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | haumanto | 72,009,799 | MDQ6VXNlcjcyMDA5Nzk5 | User | false |
openclaw/openclaw | 4,350,558,136 | I_kwDOQb6kR88AAAABA1A_uA | 74,334 | https://github.com/openclaw/openclaw/issues/74334 | https://api.github.com/repos/openclaw/openclaw/issues/74334 | v2026.4.24: Stored snippet normalization mismatch with `relocateCandidateRange` causes silent rehydration failure for all post-Apr-23 promotion candidates | ## Summary
The dreaming-promotion `--apply` pipeline rehydrates each ranked candidate by re-locating the snippet inside the source file, then re-emitting a promotion section from the freshly-located range. The rehydration is gated by `relocateCandidateRange` → `compareCandidateWindow`, which compares the *stored* snip... | open | null | false | 3 | [] | [] | 2026-04-29T12:34:31Z | 2026-05-04T10:51:45Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | haumanto | 72,009,799 | MDQ6VXNlcjcyMDA5Nzk5 | User | false |
openclaw/openclaw | 4,181,320,028 | I_kwDOQb6kR875OeFc | 58,519 | https://github.com/openclaw/openclaw/issues/58519 | https://api.github.com/repos/openclaw/openclaw/issues/58519 | Slack Socket Mode: event loop starvation causes pong timeouts and silent message loss | ## Problem
Slack Socket Mode WebSocket pong responses are missed when the Node.js event loop is busy processing agent turns. We empirically observe 4+ consecutive pong timeouts (5000ms deadline) during heavy turns, which causes Slack to drop the WebSocket connection. Messages sent by the gateway during or just after a... | open | null | false | 2 | [] | [] | 2026-03-31T18:12:52Z | 2026-04-30T17:39:26Z | null | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | jrex-jooni | 266,595,103 | U_kgDOD-PrHw | User | false |
openclaw/openclaw | 4,350,368,273 | I_kwDOQb6kR88AAAABA01aEQ | 74,321 | https://github.com/openclaw/openclaw/issues/74321 | https://api.github.com/repos/openclaw/openclaw/issues/74321 | [Bug]: Delivery-recovery retries indefinitely on permanent HTTP 400 errors (message too long, auth, not-found) — should classify and halt on non-transient failures | ## Environment
- **OpenClaw:** 2026.4.25 (aa36ee6)
- **Channel:** Telegram
- **Delivery entry:** `8520a19c-4047-4e5c-ae57-561fd8acc78f` (evidence below)
## Summary
The delivery-recovery system retries failed delivery queue entries indefinitely, regardless of whether the failure is transient or permanent. A Telegram ... | closed | not_planned | true | 3 | [] | [] | 2026-04-29T12:02:35Z | 2026-05-04T04:34:52Z | 2026-05-01T05:28:13Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | axonrelaybot | 275,827,808 | U_kgDOEHDMYA | User | false |
openclaw/openclaw | 4,340,554,308 | I_kwDOQb6kR88AAAABAreaRA | 73,254 | https://github.com/openclaw/openclaw/issues/73254 | https://api.github.com/repos/openclaw/openclaw/issues/73254 | [Bug]: Duplicate response: single user message triggers 3 API calls with different responseIds (MiniMax-M2.7) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Bug Description
When sending a single user message, OpenClaw generates THREE separate API calls with different responseIds. Only the last output is shown, but all three calls consume input tokens.
### Steps to reproduce... | closed | not_planned | true | 1 | [
"bug",
"r: support"
] | [
"vincentkoc"
] | 2026-04-28T04:18:55Z | 2026-05-02T04:28:12Z | 2026-04-29T13:10:50Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | lijuan900 | 11,470,270 | MDQ6VXNlcjExNDcwMjcw | User | false |
openclaw/openclaw | 4,350,726,328 | I_kwDOQb6kR88AAAABA1LQuA | 74,343 | https://github.com/openclaw/openclaw/issues/74343 | https://api.github.com/repos/openclaw/openclaw/issues/74343 | Slash command cross-routing when multiple Apps register same commands | ## Bug Description
When multiple Slack Apps in the same workspace register identical slash commands (e.g. `/model`, `/help`, `/config`), the Gateway exhibits cross-routing: slash commands invoked in one agent's channel are handled by a different agent.
## Environment
- OpenClaw version: 2026.4.26 (be8c246)
- Server:... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T13:00:51Z | 2026-05-03T04:34:17Z | 2026-05-01T02:08:35Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Javierpa20 | 172,389,320 | U_kgDOCkZzyA | User | false |
openclaw/openclaw | 4,350,779,627 | I_kwDOQb6kR88AAAABA1Og6w | 74,345 | https://github.com/openclaw/openclaw/issues/74345 | https://api.github.com/repos/openclaw/openclaw/issues/74345 | Event-loop saturation and ACP session leak on 2026.4.27 (9d8de70) | # Gateway event-loop pegs and ACP session lifecycle leaks on `2026.4.27` post-tag drift (commit `9d8de70`)
## Summary
Between `a8b64b7d52` (good) and `9d8de70c20` (bad) — both shipped under the `2026.4.27` tag, ~509 commits of post-tag drift — the gateway becomes unable to close ACP sessions during `task-registry-mai... | open | null | false | 1 | [] | [] | 2026-04-29T13:09:18Z | 2026-05-03T01:38:42Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | solomonneas | 41,877,493 | MDQ6VXNlcjQxODc3NDkz | User | false |
openclaw/openclaw | 4,350,902,001 | I_kwDOQb6kR88AAAABA1V-8Q | 74,353 | https://github.com/openclaw/openclaw/issues/74353 | https://api.github.com/repos/openclaw/openclaw/issues/74353 | [Bug]: idletimeout seconds | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Provider timeoutSeconds does not appear to fully replace retired agents.defaults.llm.idleTimeoutSeconds for LM Studio cold-load models
### Steps to reproduce
A slightly shorter version:
```md
## Steps to Reproduce
1. Con... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-29T13:27:16Z | 2026-05-02T04:28:10Z | 2026-04-29T13:32:37Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | NP2025-13 | 199,014,100 | U_kgDOC9y21A | User | false |
openclaw/openclaw | 4,350,999,233 | I_kwDOQb6kR88AAAABA1b6wQ | 74,363 | https://github.com/openclaw/openclaw/issues/74363 | https://api.github.com/repos/openclaw/openclaw/issues/74363 | Subagent runs can be falsely marked failed/lost after clean gateway close or pending wait | ## Summary
Subagent lifecycle recovery can falsely mark a run as failed/lost when the gateway connection closes cleanly, `agent.wait` returns `pending`, or the in-memory run context is missing while the backing session is still running.
This showed up as repeated Mission Control/Workshop subagent failures such as:
-... | open | null | false | 1 | [] | [] | 2026-04-29T13:41:56Z | 2026-05-03T10:20:00Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | nicolasdmolina | 223,382,489 | U_kgDODVCL2Q | User | false |
openclaw/openclaw | 4,351,017,258 | I_kwDOQb6kR88AAAABA1dBKg | 74,364 | https://github.com/openclaw/openclaw/issues/74364 | https://api.github.com/repos/openclaw/openclaw/issues/74364 | Strip <relevant-memories> blocks from outbound channel messages (Telegram, Discord, etc.) | # Strip `<relevant-memories>` blocks from outbound channel messages (Telegram, Discord, etc.)
## Description
When the memory plugin (`openclaw-mem0` or `memory-lancedb`) injects `<relevant-memories>` context via the `prependContext` hook, and the LLM echoes/reflects this block in its response, the raw `<relevant-memo... | open | null | false | 1 | [] | [] | 2026-04-29T13:44:16Z | 2026-05-03T09:49:36Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | Veda-openclaw | 267,342,076 | U_kgDOD-9Q_A | User | false |
openclaw/openclaw | 4,351,055,074 | I_kwDOQb6kR88AAAABA1fU4g | 74,368 | https://github.com/openclaw/openclaw/issues/74368 | https://api.github.com/repos/openclaw/openclaw/issues/74368 | [Feature]: Ensure dropdown during model selection (in openclaw onboard) | ### Summary
Right now the CLI asks the user to enter the model manually, this requires the user to search the model key name only and enter it, which is too much friction for a non-technical user
### Problem to solve
General input box for model name is confusing and hard to fill for non tech users
### Proposed solu... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-29T13:49:20Z | 2026-05-02T04:28:08Z | 2026-04-29T13:53:23Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | sanjarcode | 48,156,230 | MDQ6VXNlcjQ4MTU2MjMw | User | false |
openclaw/openclaw | 4,351,037,290 | I_kwDOQb6kR88AAAABA1ePag | 74,365 | https://github.com/openclaw/openclaw/issues/74365 | https://api.github.com/repos/openclaw/openclaw/issues/74365 | [Feature]: Handle loading states, because commands are generally slow. | ### Summary
OpenClaw CLI lacks loaders at many points, the CLI just goes blank for seconds on end. A loader can really help uncover if the command is stuck or progressing.
### Problem to solve
UX is bad without loaders
### Proposed solution
Add loaders
### Alternatives considered
_No response_
### Impact
- Bet... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-29T13:46:51Z | 2026-05-03T09:49:37Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | sanjarcode | 48,156,230 | MDQ6VXNlcjQ4MTU2MjMw | User | false |
openclaw/openclaw | 4,351,064,431 | I_kwDOQb6kR88AAAABA1f5bw | 74,370 | https://github.com/openclaw/openclaw/issues/74370 | https://api.github.com/repos/openclaw/openclaw/issues/74370 | [Bug] 消息在 memory flush + compaction 后被重复投递 | ## Bug Description
消息在 memory flush + compaction 后被重复投递。用户只发送了一条消息，但系统在大约 9 分钟后重新投递了相同的消息。
## Environment
- OpenClaw version: 2026.4.26 (be8c246)
- Platform: macOS (Darwin 24.6.0)
- Model: minimax/MiniMax-M2.7
## Steps to Reproduce
1. 用户发送消息：仔细检查一下，整个火炬之光wiki的数据是不是已经收录完全了
2. 系统正常处理消息
3. 大约 9 分钟后（01:35），触发了 memory flu... | open | null | false | 1 | [] | [] | 2026-04-29T13:50:36Z | 2026-05-03T09:13:34Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | allesgutefy | 32,950,536 | MDQ6VXNlcjMyOTUwNTM2 | User | false |
openclaw/openclaw | 4,351,074,878 | I_kwDOQb6kR88AAAABA1giPg | 74,372 | https://github.com/openclaw/openclaw/issues/74372 | https://api.github.com/repos/openclaw/openclaw/issues/74372 | [Bug]: Whatsapp package not found on npm | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
◇ Install WhatsApp plugin?
│ Download from npm (@openclaw/whatsapp)
│
◇ Install failed: WhatsApp
│
◇ Plugin install ──────────────────────────────────────────────────────────────────────────╮
│ ... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-29T13:52:08Z | 2026-05-02T04:28:08Z | 2026-04-29T13:56:49Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | sanjarcode | 48,156,230 | MDQ6VXNlcjQ4MTU2MjMw | User | false |
openclaw/openclaw | 4,351,085,528 | I_kwDOQb6kR88AAAABA1hL2A | 74,374 | https://github.com/openclaw/openclaw/issues/74374 | https://api.github.com/repos/openclaw/openclaw/issues/74374 | [Bug]: DeepSeek v4-pro: 400 "reasoning_content must be passed back" with thinking=disabled | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
When using deepseek-v4-pro with thinking:disabled via extra_body, conversations intermittently fail with:
400 The reasoning_content in the thinking mode must be passed back to the API.
The error occurs because DeepSeek occ... | open | null | false | 2 | [
"bug",
"regression"
] | [] | 2026-04-29T13:53:34Z | 2026-05-03T09:13:36Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | simmssun-hashh | 272,729,586 | U_kgDOEEGF8g | User | false |
openclaw/openclaw | 4,187,420,949 | I_kwDOQb6kR875lvkV | 59,076 | https://github.com/openclaw/openclaw/issues/59076 | https://api.github.com/repos/openclaw/openclaw/issues/59076 | JSON parse error leaked to Discord when streaming tool calls with long CJK text | ## Bug Report
**Version:** `2026.3.31` (213a704)
**Environment:** macOS arm64, Node v25.6.0, Anthropic Claude Opus 4
### Description
After upgrading to `2026.3.31`, raw JSON parse errors are being sent as Discord messages instead of being handled internally. The error appears as a standalone message in the channel:
... | closed | completed | true | 2 | [] | [] | 2026-04-01T14:36:01Z | 2026-05-02T04:28:07Z | 2026-04-29T13:59:48Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | bobfreeman1989 | 49,040,836 | MDQ6VXNlcjQ5MDQwODM2 | User | false |
openclaw/openclaw | 4,351,104,005 | I_kwDOQb6kR88AAAABA1iUBQ | 74,375 | https://github.com/openclaw/openclaw/issues/74375 | https://api.github.com/repos/openclaw/openclaw/issues/74375 | [Bug]: Discord cron `announce` delivery to `user:<id>` is rewritten to `channel:<id>`, causing `Unknown Channel | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Discord cron announce delivery appears to lose the `user:` prefix for DM targets somewhere between cron job storage and outbound queue delivery.
Direct/manual Discord DM send works to the same user, but cron a... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-29T13:56:24Z | 2026-05-02T04:28:06Z | 2026-04-29T14:02:51Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | MichelV69 | 111,482 | MDQ6VXNlcjExMTQ4Mg== | User | false |
openclaw/openclaw | 4,351,112,362 | I_kwDOQb6kR88AAAABA1i0qg | 74,376 | https://github.com/openclaw/openclaw/issues/74376 | https://api.github.com/repos/openclaw/openclaw/issues/74376 | [Bug]: Browser automation not usable in local Windows install due to Chrome sandbox failure | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Managed browser launch fails with Chrome namespace/sandbox error, preventing browser automation from functioning.
### Steps to reproduce
1. Install OpenClaw via npm on Windows.
2. Run `openclaw onboard` and c... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-29T13:57:40Z | 2026-05-03T04:35:49Z | 2026-04-30T17:05:25Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | curt-ctrl | 280,408,001 | U_kgDOELavwQ | User | false |
openclaw/openclaw | 4,351,145,658 | I_kwDOQb6kR88AAAABA1k2ug | 74,379 | https://github.com/openclaw/openclaw/issues/74379 | https://api.github.com/repos/openclaw/openclaw/issues/74379 | Feature: execDenyPathPatterns — config-level denylist for file paths in exec commands | ## Problem
OpenClaw exec policy has `safeBins` (allowlist of binaries) but no path-pattern denylist. This means an agent can run `cat /home/user/.openclaw/secrets/foo.env` and the plaintext token lands in:
1. The session JSONL transcript (on disk, persistent)
2. The next outbound LLM API request body (provider sees it... | open | null | false | 1 | [] | [] | 2026-04-29T14:02:43Z | 2026-05-03T01:23:16Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | xonaman | 45,240,351 | MDQ6VXNlcjQ1MjQwMzUx | User | false |
openclaw/openclaw | 4,351,113,696 | I_kwDOQb6kR88AAAABA1i54A | 74,377 | https://github.com/openclaw/openclaw/issues/74377 | https://api.github.com/repos/openclaw/openclaw/issues/74377 | [Bug]: tools array empty at Anthropic provider despite 17 tools computed in attempt.ts (Telegram channel) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
When using OpenClaw via Telegram channel, the Anthropic API receives tools:[] even though attempt.ts confirms tools post-normalize length=17. The agent generates tool_use as plain text instead of actual API tool_use blocks.
... | open | null | false | 4 | [
"bug",
"regression"
] | [] | 2026-04-29T13:57:52Z | 2026-05-03T04:44:13Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | everafloors-cmyk | 280,387,002 | U_kgDOELZdug | User | false |
openclaw/openclaw | 4,351,146,810 | I_kwDOQb6kR88AAAABA1k7Og | 74,380 | https://github.com/openclaw/openclaw/issues/74380 | https://api.github.com/repos/openclaw/openclaw/issues/74380 | [Bug]: Multiple skill packages fail installation | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
```
◇ Install missing skill dependencies
│ 🧩 clawhub, 🎮 gog, 📍 goplaces, 📦 mcporter, 📊 model-usage, 🎤 openai-whisper, 📜 session-logs, 🧾 summarize, 📱 wacli
│
```
```
◇ Install failed: model-usage (exit 1) — Error... | open | null | false | 1 | [
"bug"
] | [] | 2026-04-29T14:02:55Z | 2026-05-03T05:00:09Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | sanjarcode | 48,156,230 | MDQ6VXNlcjQ4MTU2MjMw | User | false |
openclaw/openclaw | 4,350,782,800 | I_kwDOQb6kR88AAAABA1OtUA | 74,346 | https://github.com/openclaw/openclaw/issues/74346 | https://api.github.com/repos/openclaw/openclaw/issues/74346 | [Bug] plugin-runtime-deps lock staleness check uses PID alone, blocks Docker gateway restarts (PID is always 7) | ## Summary
`shouldRemoveRuntimeDepsLock` (in the bundled `plugin-runtime-deps` installer) decides a stale lock is "fresh" whenever `owner.pid` matches a live PID. **Inside Docker, the gateway's Node process is always PID 1 (or PID 7 with `init: true`) in its container PID namespace.** Two different incarnations of the... | closed | completed | true | 4 | [] | [] | 2026-04-29T13:09:46Z | 2026-05-02T04:27:49Z | 2026-04-29T15:54:07Z | CONTRIBUTOR | null | 20260502T060142Z | 2026-05-02T06:01:42Z | jhsmith409 | 198,252,919 | U_kgDOC9EZdw | User | false |
openclaw/openclaw | 4,351,204,093 | I_kwDOQb6kR88AAAABA1oa_Q | 74,382 | https://github.com/openclaw/openclaw/issues/74382 | https://api.github.com/repos/openclaw/openclaw/issues/74382 | [Bug]: Asks for OpenAI Whisper/ElevenLabs API key even though initial list says KEY not needed | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
```
Install missing skill dependencies
│ 🧩 clawhub, 🎮 gog, 📍 goplaces, 📦 mcporter, 📊 model-usage, 🎤 openai-whisper, 📜 session-logs, 🧾 summarize, 📱 wacli
│
```
Near this text, in openai-whisper, the text in In pa... | open | null | false | 1 | [
"bug",
"regression"
] | [] | 2026-04-29T14:11:18Z | 2026-05-03T08:26:12Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | sanjarcode | 48,156,230 | MDQ6VXNlcjQ4MTU2MjMw | User | false |
openclaw/openclaw | 4,351,251,330 | I_kwDOQb6kR88AAAABA1rTgg | 74,386 | https://github.com/openclaw/openclaw/issues/74386 | https://api.github.com/repos/openclaw/openclaw/issues/74386 | [test] permission check | test - will close | closed | completed | true | 0 | [] | [] | 2026-04-29T14:17:55Z | 2026-05-02T04:28:05Z | 2026-04-29T14:17:56Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | 17329971 | 83,849,762 | MDQ6VXNlcjgzODQ5NzYy | User | false |
openclaw/openclaw | 4,351,233,563 | I_kwDOQb6kR88AAAABA1qOGw | 74,384 | https://github.com/openclaw/openclaw/issues/74384 | https://api.github.com/repos/openclaw/openclaw/issues/74384 | [Bug]: Plugin registerTool() tools not exposed to agent on 4.26 | ### Environment
- **OpenClaw**: 4.26 (stable, commit be8c246)
- **Plugin**: `@memtensor/memos-local-openclaw-plugin` v1.0.10
- **OS**: Windows 10 x64, Node.js v24.14.1
### Issue
The `memos-local-openclaw-plugin` registers multiple tools via `api.registerTool()`: `memory_share`, `memory_write_public`, `memory_timeline`... | open | null | false | 1 | [] | [] | 2026-04-29T14:15:26Z | 2026-05-04T02:38:34Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | 17329971 | 83,849,762 | MDQ6VXNlcjgzODQ5NzYy | User | false |
openclaw/openclaw | 4,350,740,367 | I_kwDOQb6kR88AAAABA1MHjw | 74,344 | https://github.com/openclaw/openclaw/issues/74344 | https://api.github.com/repos/openclaw/openclaw/issues/74344 | Telegram polling watchdog livelock: rebuild loop fires before runner first getUpdates | ### Bug type
Bug + small enhancement request
### Summary
On a fleet of 18 OpenClaw v2026.4.15 agents (3 pods × 6 each, Docker on macOS, Telegram polling enabled), a brief upstream Telegram blip on 2026-04-27 02:50 UTC triggered a self-sustaining "polling stall storm" lasting ~3h 13m and affecting every agent. The po... | closed | completed | true | 3 | [] | [] | 2026-04-29T13:03:12Z | 2026-05-02T04:26:14Z | 2026-04-29T14:21:57Z | CONTRIBUTOR | null | 20260502T060142Z | 2026-05-02T06:01:42Z | koshaji | 243,474,115 | U_kgDODoMeww | User | false |
openclaw/openclaw | 4,351,266,980 | I_kwDOQb6kR88AAAABA1sQpA | 74,388 | https://github.com/openclaw/openclaw/issues/74388 | https://api.github.com/repos/openclaw/openclaw/issues/74388 | Bug: New session fails with anthropic-messages API models due to empty messages array | ## Description
When a new session starts with a model using the `anthropic-messages` API format (e.g., minimax-M2.7 via minimax-portal), the LLM request fails with `"messages must not be empty"`. The fallback model (using `openai-completions` format) then succeeds.
## Root Cause
During new session initialization, Op... | closed | not_planned | true | 1 | [] | [] | 2026-04-29T14:20:15Z | 2026-05-02T04:28:03Z | 2026-04-29T14:24:52Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | ariesy | 956,025 | MDQ6VXNlcjk1NjAyNQ== | User | false |
openclaw/openclaw | 4,351,266,442 | I_kwDOQb6kR88AAAABA1sOig | 74,387 | https://github.com/openclaw/openclaw/issues/74387 | https://api.github.com/repos/openclaw/openclaw/issues/74387 | MCP tools not accessible to main agent — only to embedded Pi/coding sessions | ## Problem
Currently, MCP servers configured under `mcp.servers` in `openclaw.json` are only accessible to embedded Pi / coding agent runtimes, not to the main OpenClaw agent.
The `openclaw mcp list` command only manages the MCP server **registry**, not the actual tool exposure to the main agent. MCP server definitio... | closed | completed | true | 3 | [] | [] | 2026-04-29T14:20:09Z | 2026-05-02T04:26:04Z | 2026-04-29T14:25:12Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | luuq2026 | 268,951,258 | U_kgDOEAfe2g | User | false |
openclaw/openclaw | 4,350,387,057 | I_kwDOQb6kR88AAAABA02jcQ | 74,323 | https://github.com/openclaw/openclaw/issues/74323 | https://api.github.com/repos/openclaw/openclaw/issues/74323 | Gateway WebSocket handshake timeout on loopback affects active-memory, streaming, and embedded runs | ## Gateway WebSocket handshake timeout affects active-memory and streaming output
### Summary
The OpenClaw Gateway\''s local WebSocket connections (loopback 127.0.0.1:18789) experience handshake timeouts that affect multiple subsystems: active-memory plugin, streaming output from @larksuite/openclaw-lark, and potenti... | closed | duplicate | true | 4 | [] | [] | 2026-04-29T12:05:44Z | 2026-05-02T04:26:31Z | 2026-04-29T14:29:07Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | hfgwq2dgx8-tech | 279,440,970 | U_kgDOEKfuSg | User | false |
openclaw/openclaw | 4,351,234,261 | I_kwDOQb6kR88AAAABA1qQ1Q | 74,385 | https://github.com/openclaw/openclaw/issues/74385 | https://api.github.com/repos/openclaw/openclaw/issues/74385 | [Bug]: Stuck and ultra slow during Hatch (terminal mode) - took a minute to come back up | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
```
◇ How do you want to hatch your bot?
│ Hatch in Terminal (recommended)
🦞 OpenClaw 2026.4.26 (be8c246) — Your .zshrc wishes it could do what I do.
openclaw tui - local embedded - agent main - session m... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-29T14:15:31Z | 2026-05-03T09:49:39Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | sanjarcode | 48,156,230 | MDQ6VXNlcjQ4MTU2MjMw | User | false |
openclaw/openclaw | 4,351,321,784 | I_kwDOQb6kR88AAAABA1vmuA | 74,394 | https://github.com/openclaw/openclaw/issues/74394 | https://api.github.com/repos/openclaw/openclaw/issues/74394 | [Bug]: DeepSeek V4-pro 400 error via opencode-go: reasoning_content passthrough regression in v2026.4.26 | ### Environment
- OpenClaw Version: 2026.4.26 (aa36ee6)
- Provider: opencode-go
- Model: deepseek-v4-pro / deepseek-v4-flash
- Environment: Arch Linux Chroot (arm64)
### Description
Multi-turn conversations with DeepSeek V4 models via OpenCode-Go are failing with HTTP 400. This occurs when thinking is enabled (and int... | closed | completed | true | 2 | [] | [] | 2026-04-29T14:28:03Z | 2026-05-02T04:28:01Z | 2026-04-29T14:32:54Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | AmnAnon | 48,866,635 | MDQ6VXNlcjQ4ODY2NjM1 | User | false |
openclaw/openclaw | 4,351,330,559 | I_kwDOQb6kR88AAAABA1wI_w | 74,395 | https://github.com/openclaw/openclaw/issues/74395 | https://api.github.com/repos/openclaw/openclaw/issues/74395 | Auto-migration silently switches default model to a paid `-preview` variant on upgrade, causing unintended billing | ### Summary
On upgrade from 2026.3.2 → 2026.3.22, my ~/.openclaw/openclaw.json was silently mutated such that agents.defaults.model.primary changed from google/gemini-2.5-pro to google/gemini-3.1-pro-preview without any user-facing prompt, changelog notice highlighting price-tier impact, or explicit consent step. The ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-29T14:29:12Z | 2026-05-03T06:50:53Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | ihongsoon | 162,789,618 | U_kgDOCbP48g | User | false |
openclaw/openclaw | 4,351,412,973 | I_kwDOQb6kR88AAAABA11K7Q | 74,405 | https://github.com/openclaw/openclaw/issues/74405 | https://api.github.com/repos/openclaw/openclaw/issues/74405 | [Bug]: prestageGatewayBundledRuntimeDeps ignores channels.X.enabled=false, installs deps for all bundled channels on startup |
## Title
`[Bug]: prestageGatewayBundledRuntimeDeps ignores channels.X.enabled=false, installs deps for all bundled channels on startup`
## Body
### Bug type
Regression (partially fixed in doctor/health path but not startup path)
### Summary
`prestageGatewayBundledRuntimeDeps` during gateway startup installs runtim... | closed | completed | true | 1 | [] | [] | 2026-04-29T14:40:44Z | 2026-05-02T04:27:22Z | 2026-04-29T19:31:35Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | zeeyuu222 | 267,768,244 | U_kgDOD_XRtA | User | false |
openclaw/openclaw | 4,351,376,400 | I_kwDOQb6kR88AAAABA1y8EA | 74,401 | https://github.com/openclaw/openclaw/issues/74401 | https://api.github.com/repos/openclaw/openclaw/issues/74401 | [Bug]: sqlite-vec unavailable on node:sqlite — vector recall degraded (macOS ARM64, v2026.4.26) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
sqlite-vec fails to load on macOS ARM64 with node:sqlite runtime. Vector recall degraded to BM25-only. Root cause: sqliteVec.load() is better-sqlite3 API, incompatible with node:sqlite used in 2026.4.26.
¥
### Steps to repr... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-04-29T14:35:48Z | 2026-05-02T04:28:01Z | 2026-04-29T14:41:29Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | googlerest | 127,843,198 | U_kgDOB567fg | User | false |
openclaw/openclaw | 4,351,351,714 | I_kwDOQb6kR88AAAABA1xbog | 74,399 | https://github.com/openclaw/openclaw/issues/74399 | https://api.github.com/repos/openclaw/openclaw/issues/74399 | [Feature]: Add ETA timeline during `openclaw onboard` | ### Summary
It takes around 40 minutes. Its a huge disservice to not show this upfront to a beginner. Leads to very bad onboarding experience.
### Problem to solve
No ETA is shown when `openclaw onboard starts`
### Proposed solution
Show prompt that setup can take a long time, viz 40 mins
### Alternatives conside... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-29T14:32:15Z | 2026-05-03T07:05:31Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | sanjarcode | 48,156,230 | MDQ6VXNlcjQ4MTU2MjMw | User | false |
openclaw/openclaw | 4,350,854,082 | I_kwDOQb6kR88AAAABA1TDwg | 74,350 | https://github.com/openclaw/openclaw/issues/74350 | https://api.github.com/repos/openclaw/openclaw/issues/74350 | [Security] Windows ACL audit bypass: Anonymous and Guest SIDs are misclassified as "group" instead of "world" | ### Description
In the Windows file permission auditing mechanism (`src/security/windows-acl.ts`), the `classifyPrincipal` function categorizes Windows ACL entries into `trusted`, `world`, or `group`.
Currently, `WORLD_PRINCIPALS` and `WORLD_SIDS` strictly cover `Everyone` (S-1-1-0), `Authenticated Users` (S-1-5-11), ... | open | null | false | 2 | [] | [] | 2026-04-29T13:20:22Z | 2026-05-03T09:13:33Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | Liu-XinYuan | 20,487,589 | MDQ6VXNlcjIwNDg3NTg5 | User | false |
openclaw/openclaw | 4,350,890,451 | I_kwDOQb6kR88AAAABA1VR0w | 74,351 | https://github.com/openclaw/openclaw/issues/74351 | https://api.github.com/repos/openclaw/openclaw/issues/74351 | [Bug]: allowRfc2544BenchmarkRange SSRF exemption does not cover IPv6 ULA addresses from fake-ip proxies | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When fake-ip proxy tools (sing-box, Clash, Surge) are used, DNS returns both IPv4 addresses in the 198.18.0.0/15 (RFC 2544 benchmark) range AND IPv6 addresses in the fc00::/7 (Unique Local Address) range. Setti... | closed | completed | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-04-29T13:25:33Z | 2026-05-02T04:27:23Z | 2026-04-29T19:31:39Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | Joelincn | 144,782,293 | U_kgDOCKEz1Q | User | false |
openclaw/openclaw | 4,350,902,999 | I_kwDOQb6kR88AAAABA1WC1w | 74,354 | https://github.com/openclaw/openclaw/issues/74354 | https://api.github.com/repos/openclaw/openclaw/issues/74354 | [Bug]: 通过 Control UI 发送纯数字时，数字会被自动添加千位分隔符（如 123456789 → 123-456-789） | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Control UI 自动将数字转换为千位分隔符格式
### 问题描述
通过 Control UI 发送纯数字时（如 `123456789`），数字会被自动转换为带连字符的千位分隔符格式（如 `123-456-789`）。
这导致：
- 无法正确传递数字类型的 ID（如 GitHub 用户名）
- 无法传递纯数字参数
- 导致URL无法打开
### 预期行为
数字应该保持原样发送，不应自动格式化。
### 环境信息
- OpenC... | open | null | false | 2 | [
"bug"
] | [] | 2026-04-29T13:27:25Z | 2026-05-03T04:44:10Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | s1sunny | 51,485,377 | MDQ6VXNlcjUxNDg1Mzc3 | User | false |
openclaw/openclaw | 4,350,919,036 | I_kwDOQb6kR88AAAABA1XBfA | 74,356 | https://github.com/openclaw/openclaw/issues/74356 | https://api.github.com/repos/openclaw/openclaw/issues/74356 | Dockerfile: Bun installed via unverified curl | bash while Node base images are SHA256-pinned | ## Summary
`Dockerfile:47` fetches and executes the Bun installer from `https://bun.sh/install` with no version pin, checksum, or signature verification:
\`\`\`dockerfile
if curl --retry 5 --retry-all-errors --retry-delay 2 -fsSL https://bun.sh/install | bash; then
\`\`\`
This is inconsistent with how every other ba... | closed | completed | true | 2 | [] | [] | 2026-04-29T13:29:55Z | 2026-05-05T04:27:02Z | 2026-05-02T14:46:53Z | CONTRIBUTOR | null | 20260505T060034Z | 2026-05-05T06:00:34Z | fede-kamel | 209,537,060 | U_kgDODH1IJA | User | false |
openclaw/openclaw | 4,350,921,258 | I_kwDOQb6kR88AAAABA1XKKg | 74,357 | https://github.com/openclaw/openclaw/issues/74357 | https://api.github.com/repos/openclaw/openclaw/issues/74357 | WEB_TOOLS_TRUSTED_NETWORK_SSRF_POLICY applies dangerouslyAllowPrivateNetwork to all 10 web search providers | ## Summary
`src/agents/tools/web-guarded-fetch.ts:10–13` defines a single SSRF policy used by every web search provider in OpenClaw:
\`\`\`typescript
const WEB_TOOLS_TRUSTED_NETWORK_SSRF_POLICY: SsrFPolicy = {
dangerouslyAllowPrivateNetwork: true,
allowRfc2544BenchmarkRange: true,
};
\`\`\`
This policy is applie... | closed | completed | true | 4 | [] | [] | 2026-04-29T13:30:16Z | 2026-05-05T04:27:34Z | 2026-05-02T05:40:00Z | CONTRIBUTOR | null | 20260505T060034Z | 2026-05-05T06:00:34Z | fede-kamel | 209,537,060 | U_kgDODH1IJA | User | false |
openclaw/openclaw | 4,350,921,664 | I_kwDOQb6kR88AAAABA1XLwA | 74,358 | https://github.com/openclaw/openclaw/issues/74358 | https://api.github.com/repos/openclaw/openclaw/issues/74358 | Slack: streaming preview completely silent when toolProgress: false — verbose mode broken since v2026.4.21 | ## Bug: Slack verbose mode produces no in-flight preview with `streaming.preview.toolProgress: false` (v2026.4.26)
**Relates to:** #71320 (partial fix), #69611 (introduced regression)
**Environment**
- OpenClaw `2026.4.26`
- Channel: Slack socket mode
- Verbose level: `on` and `full`
- Config: `channels.slack.account... | open | null | false | 6 | [] | [] | 2026-04-29T13:30:20Z | 2026-05-03T04:44:12Z | null | CONTRIBUTOR | null | 20260503T060030Z | 2026-05-03T06:00:30Z | rafaelreis-r | 57,492,577 | MDQ6VXNlcjU3NDkyNTc3 | User | false |
openclaw/openclaw | 4,351,453,966 | I_kwDOQb6kR88AAAABA13rDg | 74,415 | https://github.com/openclaw/openclaw/issues/74415 | https://api.github.com/repos/openclaw/openclaw/issues/74415 | test | test | closed | completed | true | 0 | [] | [] | 2026-04-29T14:46:22Z | 2026-05-02T04:27:59Z | 2026-04-29T14:46:24Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | laobang | 35,260,880 | MDQ6VXNlcjM1MjYwODgw | User | false |
openclaw/openclaw | 4,351,435,288 | I_kwDOQb6kR88AAAABA12iGA | 74,410 | https://github.com/openclaw/openclaw/issues/74410 | https://api.github.com/repos/openclaw/openclaw/issues/74410 | meridian/anthropic-messages stream reader drops text content blocks when response is [thinking, text] | ## Summary
When a model accessed via `meridian` (or anthropic direct) returns content in `[thinking, text]` shape via the anthropic-messages streaming API, the openclaw stream reader drops the `text` block entirely. The persisted assistant message ends up with `content: []` and `usage.output: 0`, despite the upstream ... | closed | completed | true | 1 | [] | [] | 2026-04-29T14:43:53Z | 2026-05-02T04:27:24Z | 2026-04-29T19:16:54Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | jesuskannolis | 10,068,263 | MDQ6VXNlcjEwMDY4MjYz | User | false |
openclaw/openclaw | 4,351,397,259 | I_kwDOQb6kR88AAAABA10Niw | 74,404 | https://github.com/openclaw/openclaw/issues/74404 | https://api.github.com/repos/openclaw/openclaw/issues/74404 | [Bug]: Gateway becomes CPU-saturated and agents stop progressing mid-work | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Issues started when upgrading to 2026.4.23 and up, with last version 2026.4.25.
Every version got worse, higher cpu and slower startup and agents slow response:
OpenClaw Gateway becomes CPU-saturated and agents stop progres... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-04-29T14:38:34Z | 2026-05-04T04:36:35Z | 2026-05-01T11:48:04Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | najef1979-code | 233,555,716 | U_kgDODevHBA | User | false |
openclaw/openclaw | 4,350,521,340 | I_kwDOQb6kR88AAAABA0-v_A | 74,328 | https://github.com/openclaw/openclaw/issues/74328 | https://api.github.com/repos/openclaw/openclaw/issues/74328 | [Bug]: Gateway main thread CPU-bound at ~100% on v2026.4.26 / current main; clean on v2026.4.22 (fs.stat storm in microtask queue) | ### Bug type
Regression (worked before, now fails)
### Summary
After upgrading from `v2026.4.22` to `v2026.4.26` (also reproduces on current `main` at `9bb1e59`, which package.json reports as `2026.4.27`), the gateway sits at ~100% CPU on its single main thread and stops responding to local probes. Same host, same `~/... | closed | completed | true | 8 | [] | [] | 2026-04-29T12:28:19Z | 2026-05-04T04:36:35Z | 2026-05-01T11:48:02Z | CONTRIBUTOR | null | 20260504T060031Z | 2026-05-04T06:00:31Z | odrobnik | 333,270 | MDQ6VXNlcjMzMzI3MA== | User | false |
openclaw/openclaw | 4,351,456,731 | I_kwDOQb6kR88AAAABA1312w | 74,416 | https://github.com/openclaw/openclaw/issues/74416 | https://api.github.com/repos/openclaw/openclaw/issues/74416 | DeepSeek V4 Flash streaming with reasoning_content always triggers fallback to fallback model |
## Description
When using `deepseek/deepseek-v4-flash` as the primary model with fallbacks configured (e.g., `minimax-cn/MiniMax-M2.7`), **every** response generation triggers fallback to the fallback model, even though DeepSeek API responds correctly when tested directly.
This happens 100% of the time. The model on... | closed | completed | true | 1 | [] | [] | 2026-04-29T14:46:44Z | 2026-05-02T04:27:59Z | 2026-04-29T14:52:06Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | laobang | 35,260,880 | MDQ6VXNlcjM1MjYwODgw | User | false |
openclaw/openclaw | 4,351,447,677 | I_kwDOQb6kR88AAAABA13SfQ | 74,413 | https://github.com/openclaw/openclaw/issues/74413 | https://api.github.com/repos/openclaw/openclaw/issues/74413 | [Feature]: Do hello-world/trial of WhatsApp channel when that's set up/ also docs show setup but again, no hello world example given | ### Summary
Right now, WhatsApp config is done, and the onboarding ends. But there's no indication as to how to use the WhatsApp in OpenClaw. There should be some hello-world message or hook here that starts helps in discovering how WhatsApp works.
### Problem to solve
WhatsApp gets connected, but there's no hook as... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-29T14:45:32Z | 2026-05-03T09:49:41Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | sanjarcode | 48,156,230 | MDQ6VXNlcjQ4MTU2MjMw | User | false |
openclaw/openclaw | 4,351,490,364 | I_kwDOQb6kR88AAAABA155PA | 74,419 | https://github.com/openclaw/openclaw/issues/74419 | https://api.github.com/repos/openclaw/openclaw/issues/74419 | feat: support Jina Embeddings v5 task parameter for task-specific adapters | ## Feature request
Jina Embeddings v5 (`jina-embeddings-v5-text-nano` / `jina-embeddings-v5-text-small`) introduces **task-specific LoRA adapters** that significantly boost performance per task type. This requires passing a `task` parameter in the embeddings API body.
### Current behavior
OpenClaw only sends `model`... | open | null | false | 1 | [] | [] | 2026-04-29T14:51:02Z | 2026-05-03T10:52:48Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | seoseo-ai | 261,797,377 | U_kgDOD5q2AQ | User | false |
openclaw/openclaw | 4,351,432,589 | I_kwDOQb6kR88AAAABA12XjQ | 74,409 | https://github.com/openclaw/openclaw/issues/74409 | https://api.github.com/repos/openclaw/openclaw/issues/74409 | silent-reply detection gated behind isGroupChat — direct chats can't honor silentReplyPolicy="allow" | ## Summary
In the bundled `get-reply-eY9NJdyX.js` (source `src/auto-reply/.../get-reply.ts`), `allowEmptyAssistantReplyAsSilent` is gated behind `isGroupChat &&`, which makes it impossible to honor `silentReplyPolicy: "allow"` on direct (DM) chats. Empty/silent-token replies on direct chats always trigger the empty-re... | closed | completed | true | 1 | [] | [] | 2026-04-29T14:43:30Z | 2026-05-02T04:27:26Z | 2026-04-29T19:02:39Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | jesuskannolis | 10,068,263 | MDQ6VXNlcjEwMDY4MjYz | User | false |
openclaw/openclaw | 4,351,534,502 | I_kwDOQb6kR88AAAABA18lpg | 74,421 | https://github.com/openclaw/openclaw/issues/74421 | https://api.github.com/repos/openclaw/openclaw/issues/74421 | [Bug]: Telegram voice notes not auto-transcribed despite unified inbound handling fix (#20591) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Problem
Telegram voice notes are not being automatically transcribed to text, even though:
• tools.media.audio is correctly configured with OpenAI Whisper
• The API key works (tested manually with curl)
• PR #20591 (merged... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-29T14:56:43Z | 2026-05-04T04:36:48Z | 2026-05-01T11:11:46Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | jasontoff | 2,187,423 | MDQ6VXNlcjIxODc0MjM= | User | false |
openclaw/openclaw | 4,351,540,119 | I_kwDOQb6kR88AAAABA187lw | 74,423 | https://github.com/openclaw/openclaw/issues/74423 | https://api.github.com/repos/openclaw/openclaw/issues/74423 | [Bug]: `/models` and Web chat model dropdown show full catalog (900+ models) instead of only configured providers | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When using the `/models` command or the model dropdown in the Web chat, the UI displays the entire model catalog (900+ entries) rather than filtering to only the providers and models the user has configured aut... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-04-29T14:57:27Z | 2026-05-02T04:27:40Z | 2026-04-29T17:05:36Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | guarismo | 33,353,023 | MDQ6VXNlcjMzMzUzMDIz | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.