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 | 4,060,272,117 | I_kwDOQb6kR87yAtX1 | 43,432 | https://github.com/openclaw/openclaw/issues/43432 | https://api.github.com/repos/openclaw/openclaw/issues/43432 | [Feature]: Memory durability config — hard flush threshold, priority-aware compaction, cache TTL | ### Summary
Add three config keys to improve memory durability at scale: memoryFlush.hardThresholdTokens, compaction.strategy (preserve-priority), and memorySearch.cache.ttlSeconds.
### Problem to solve
Running 32 agents on a single OpenClaw gateway (v2026.3.8) with the memory system fully enabled. Three gaps in the... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-11T19:40:51Z | 2026-05-04T00:52:15Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | omegabyte-ai | 259,880,558 | U_kgDOD312bg | User | false |
openclaw/openclaw | 4,062,001,104 | I_kwDOQb6kR87yHTfQ | 43,661 | https://github.com/openclaw/openclaw/issues/43661 | https://api.github.com/repos/openclaw/openclaw/issues/43661 | Session hangs indefinitely when compaction times out, causing repeated duplicate message sends | ## Bug Report
### Summary
When a session context becomes large enough to trigger compaction, and that compaction process times out, the agent enters a silent failure loop. Each timeout (~10 min) triggers a delivery retry that resends the same message to the user repeatedly — with no recovery, no fallback, and no way ... | open | null | false | 7 | [] | [
"vincentkoc"
] | 2026-03-12T04:12:38Z | 2026-05-04T00:43:03Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | thedanchez | 4,277,756 | MDQ6VXNlcjQyNzc3NTY= | User | false |
openclaw/openclaw | 4,062,137,074 | I_kwDOQb6kR87yH0ry | 43,681 | https://github.com/openclaw/openclaw/issues/43681 | https://api.github.com/repos/openclaw/openclaw/issues/43681 | [Feature]: OAuth Rotation for same Model Provider (Multi-Team Subscription Support) | ### Summary
I am requesting an enhancement to the Auth Profile system to support multiple team subscriptions under the exact same OAuth account. This would allow users with multiple business tier subscriptions (like OpenAI Codex) to fully utilize OpenClaw's built-in fallback and auth rotation mechanisms.
Current Beh... | closed | not_planned | false | 1 | [
"enhancement"
] | [] | 2026-03-12T04:51:04Z | 2026-05-04T12:09:14Z | 2026-05-04T12:09:14Z | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | maltese53 | 90,551,205 | MDQ6VXNlcjkwNTUxMjA1 | User | false |
openclaw/openclaw | 4,062,695,313 | I_kwDOQb6kR87yJ8-R | 43,786 | https://github.com/openclaw/openclaw/issues/43786 | https://api.github.com/repos/openclaw/openclaw/issues/43786 | [Feature]: auth mode = none is required | ### Summary
Feature request / improvement
openclaw may be deployed inside a trusted internal network where no user authorization is required. At the moment the only viable approach is to enable the trusted proxy option and place nginx in front of the service.
This workaround feels unnecessary and awkward in environm... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-12T07:13:46Z | 2026-05-04T00:30:46Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | jetcar | 5,941,928 | MDQ6VXNlcjU5NDE5Mjg= | User | false |
openclaw/openclaw | 4,062,904,803 | I_kwDOQb6kR87yKwHj | 43,830 | https://github.com/openclaw/openclaw/issues/43830 | https://api.github.com/repos/openclaw/openclaw/issues/43830 | [Feature]: Automatic Output Sanitization for Sensitive Data | ### Summary
Add automatic output sanitization to prevent accidental leakage of sensitive information (API keys, tokens, passwords, PII) in agent responses, regardless of whether the requester is a tester or administrator.
### Problem to solve
Currently, OpenClaw has an `output-sanitizer` skill that provides guidance... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-12T07:58:51Z | 2026-05-04T00:43:08Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | 937089773 | 20,830,184 | MDQ6VXNlcjIwODMwMTg0 | User | false |
openclaw/openclaw | 4,063,586,125 | I_kwDOQb6kR87yNWdN | 43,929 | https://github.com/openclaw/openclaw/issues/43929 | https://api.github.com/repos/openclaw/openclaw/issues/43929 | [Feature]: logical chat history across rolled sessions | ### Summary
Add a logical-history mode to `chat.history` so long-lived chats can retain continuity across daily, idle, or manual session rollovers.
### Problem to solve
A stable session key such as `agent:main:main` can roll over to a new `sessionId` after daily, idle, or manual reset. That behavior is expected, but... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-03-12T10:14:07Z | 2026-05-04T00:43:09Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | brian-chenyang | 212,660,948 | U_kgDODKzy1A | User | false |
openclaw/openclaw | 4,066,041,443 | I_kwDOQb6kR87yWt5j | 44,253 | https://github.com/openclaw/openclaw/issues/44253 | https://api.github.com/repos/openclaw/openclaw/issues/44253 | Per-agent tools.selfDeny — allow tool inheritance to subagents while denying direct use | ### Summary
Add a `tools.selfDeny` key to `agents.list[].tools` so an orchestrator agent can pass tools (like `exec`, `write`, `edit`) to its subagents while being config-blocked from using those tools directly in its own session.
### Problem to solve
In the orchestrator pattern (main → orchestrator → worker subagen... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-12T17:04:59Z | 2026-05-03T23:37:27Z | null | NONE | null | 20260504T000036Z | 2026-05-04T00:00:36Z | Vinaykumarvanam5 | 195,172,718 | U_kgDOC6IZbg | User | false |
openclaw/openclaw | 4,219,519,108 | I_kwDOQb6kR877gMCE | 62,644 | https://github.com/openclaw/openclaw/issues/62644 | https://api.github.com/repos/openclaw/openclaw/issues/62644 | [Bug/UX] Silent breaking change: tools.sessions.visibility default=tree breaks existing isolated cron sessions_send | ## Problem
After upgrading OpenClaw (to v2026.4.2+), **existing isolated cron tasks that use `sessions_send` to report results back to the main session silently fail** with:
```
Session send visibility is restricted. Set tools.sessions.visibility=all to allow cross-agent access.
```
## Root Cause
Security advisory ... | closed | not_planned | true | 1 | [] | [] | 2026-04-07T17:38:05Z | 2026-05-03T04:36:19Z | 2026-04-30T15:29:59Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | zhuyankarl | 7,082,758 | MDQ6VXNlcjcwODI3NTg= | User | false |
openclaw/openclaw | 4,359,349,632 | I_kwDOQb6kR88AAAABA9ZlgA | 75,130 | https://github.com/openclaw/openclaw/issues/75130 | https://api.github.com/repos/openclaw/openclaw/issues/75130 | [Bug]: gateway accepts agent runs before plugin tool registration completes; pre-ready window produces spurious 'No callable tools remain' | ## Summary
Gateway accepts agent runs (via WS) before plugin tool registration completes. Any inbound message that arrives in the ~17s window between `[gateway] starting HTTP server` and `[gateway] ready` is dispatched to an agent whose tool registry is still empty. The `*` wildcard in the agent's tool allowlist resol... | closed | completed | true | 1 | [] | [] | 2026-04-30T15:35:55Z | 2026-05-03T04:36:11Z | 2026-04-30T15:56:04Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | solosage1 | 158,075,162 | U_kgDOCWwJGg | User | false |
openclaw/openclaw | 4,359,468,968 | I_kwDOQb6kR88AAAABA9g3qA | 75,147 | https://github.com/openclaw/openclaw/issues/75147 | https://api.github.com/repos/openclaw/openclaw/issues/75147 | Agent drops tool execution responses when new user message arrives mid-task | When the agent is executing a long-running tool (e.g. `gateway restart`) and a new user message arrives before the tool completes, the execution appears to be interrupted and the tool result is never delivered to the user. The agent does not follow up or retry.
**Steps to reproduce:**
1. Send a command that triggers a... | closed | completed | true | 1 | [] | [] | 2026-04-30T15:55:08Z | 2026-05-03T04:36:10Z | 2026-04-30T16:00:53Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | avatasia | 198,357 | MDQ6VXNlcjE5ODM1Nw== | User | false |
openclaw/openclaw | 4,359,478,790 | I_kwDOQb6kR88AAAABA9heBg | 75,150 | https://github.com/openclaw/openclaw/issues/75150 | https://api.github.com/repos/openclaw/openclaw/issues/75150 | [Feature]: Is there any reason to disable openai-codex/gpt5.4mini? | ### Summary
I'm using Openai-codex/gpt-5.4-mini for daily cron like heartbeat or sending some stuff, summerize email etc.
Why did you blocked openai-codex/gpt-5.4-mini? It works well with my codex.
I'm having this problem
lane task error: lane=session:agent:news:main durationMs=5130 error="FailoverError: Unknown model... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-30T15:56:33Z | 2026-05-03T04:36:10Z | 2026-04-30T16:01:12Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | boeingchoco | 194,511,351 | U_kgDOC5gB9w | User | false |
openclaw/openclaw | 4,359,477,026 | I_kwDOQb6kR88AAAABA9hXIg | 75,149 | https://github.com/openclaw/openclaw/issues/75149 | https://api.github.com/repos/openclaw/openclaw/issues/75149 | Bundled runtime deps cache reports 1 missing for chokidar on every restart | On every gateway restart, the following log appears:
[plugins] staging bundled runtime deps before gateway startup (1 missing, 30 install specs): chokidar@^5.0.0
[plugins] installed bundled runtime deps before gateway startup in ~4s: chokidar@^5.0.0
**Root cause:** chokidar@^5.0.0 is declared in openclaw's own packag... | closed | completed | true | 1 | [] | [] | 2026-04-30T15:56:17Z | 2026-05-03T04:36:09Z | 2026-04-30T16:01:57Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | avatasia | 198,357 | MDQ6VXNlcjE5ODM1Nw== | User | false |
openclaw/openclaw | 4,217,980,465 | I_kwDOQb6kR877aUYx | 62,516 | https://github.com/openclaw/openclaw/issues/62516 | https://api.github.com/repos/openclaw/openclaw/issues/62516 | [Bug]: tools.profile: "coding" | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
With top-level "tools.profile": "coding" configured, openclaw agent --agent main --json --message "List every tool available to you right now." produced systemPromptReport.tools.entries: [] in the JSON output.
... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-07T13:27:17Z | 2026-05-03T04:36:05Z | 2026-04-30T16:16:52Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | TehCl4w | 270,170,462 | U_kgDOEBp5Xg | User | false |
openclaw/openclaw | 4,218,075,505 | I_kwDOQb6kR877arlx | 62,525 | https://github.com/openclaw/openclaw/issues/62525 | https://api.github.com/repos/openclaw/openclaw/issues/62525 | Managed hooks should receive plugin lifecycle events (before_dispatch, agent_end, etc.) | # Feature Request: Managed Hooks Should Receive Plugin Events
## Summary
Managed hooks (in `~/.openclaw/hooks/`) cannot receive important events like `before_dispatch`, `agent_end`, or `message_sent` because they register via `registerInternalHook()` but these events only trigger `hookRunner.hasHooks()` (plugin hook s... | closed | not_planned | true | 1 | [] | [] | 2026-04-07T13:42:48Z | 2026-05-03T04:36:04Z | 2026-04-30T16:16:58Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | kyle20026 | 274,152,656 | U_kgDOEFc80A | User | false |
openclaw/openclaw | 4,359,685,590 | I_kwDOQb6kR88AAAABA9uF1g | 75,161 | https://github.com/openclaw/openclaw/issues/75161 | https://api.github.com/repos/openclaw/openclaw/issues/75161 | [Bug] sqlite-vec package not found after v2026.4.27 upgrade — memory vector search broken | ## Bug Description
After upgrading to v2026.4.27, the memory system logs repeated errors:
```
[memory] sqlite-vec unavailable: Cannot find package 'sqlite-vec' imported from /path/to/openclaw/plugin-runtime-deps/openclaw-2026.4.27-*/dist/engine-storage-*.js
Did you mean to import "sqlite-vec/index.cjs"?
[memory] chun... | closed | completed | true | 1 | [] | [] | 2026-04-30T16:28:54Z | 2026-05-03T04:35:58Z | 2026-04-30T16:34:05Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | xdengli | 149,284,229 | U_kgDOCOXlhQ | User | false |
openclaw/openclaw | 4,220,153,155 | I_kwDOQb6kR877im1D | 62,681 | https://github.com/openclaw/openclaw/issues/62681 | https://api.github.com/repos/openclaw/openclaw/issues/62681 | [Feature Request] Support for internationalization (i18n) of Telegram bot commands | ## Feature Request
Add support for translating Telegram bot commands (and Web UI commands) into different languages via configuration.
### Current Behavior
All Telegram commands (`/new`, `/reset`, `/model`, `/status`, etc.) are hardcoded in English.
### Desired Behavior
Allow users to configure a locale (e.g., `fr`,... | closed | not_planned | true | 1 | [] | [] | 2026-04-07T19:37:23Z | 2026-05-03T04:35:58Z | 2026-04-30T16:42:45Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | yotisama | 209,351,030 | U_kgDODHpxdg | User | false |
openclaw/openclaw | 4,359,623,511 | I_kwDOQb6kR88AAAABA9qTVw | 75,158 | https://github.com/openclaw/openclaw/issues/75158 | https://api.github.com/repos/openclaw/openclaw/issues/75158 | before_prompt_build hook timeout hardcoded to 15s in 2026.4.27 — active-memory plugin's timeoutMs setting silently ineffective | ## Summary
In OpenClaw 2026.4.27, the `before_prompt_build` modifying-hook timeout is hardcoded to 15 seconds and not exposed through any `openclaw.json` field. As a result, the `active-memory` plugin's documented `timeoutMs` config (which we have set to 20000) is silently ineffective — the outer hook layer aborts the... | closed | completed | true | 1 | [] | [] | 2026-04-30T16:18:49Z | 2026-05-03T04:35:56Z | 2026-04-30T16:43:00Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Chris2Eng | 30,375,456 | MDQ6VXNlcjMwMzc1NDU2 | User | false |
openclaw/openclaw | 4,064,275,894 | I_kwDOQb6kR87yP-22 | 44,025 | https://github.com/openclaw/openclaw/issues/44025 | https://api.github.com/repos/openclaw/openclaw/issues/44025 | [Feature]: Intelligent Model Fallback with Wildcard/Auto Selection | ### Summary
Auto-fallback model selection for provider-specific model pools (e.g. openai-codex/auto or openai-codex/*).
### Problem to solve
When a provider silently renames or rotates their default model (e.g. OpenAI changed their Codex default to gpt-5.4), the configured default_model string breaks. The system fal... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-12T12:20:52Z | 2026-05-04T12:45:50Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | seggl | 10,183,569 | MDQ6VXNlcjEwMTgzNTY5 | User | false |
openclaw/openclaw | 4,071,777,291 | I_kwDOQb6kR87ysmQL | 45,233 | https://github.com/openclaw/openclaw/issues/45233 | https://api.github.com/repos/openclaw/openclaw/issues/45233 | [Feature]: FreeBSD | ### Summary
Please add FreeBSD to the list of the OS supported.
### Problem to solve
release for FreeBSD
### Proposed solution
build a pkg and push to the FreeBSD ports and packages system
### Alternatives considered
_No response_
### Impact
Tier 1 OS support
### Evidence/examples
_No response_
### Additio... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-13T15:22:44Z | 2026-05-04T00:11:57Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | lucarinaorg | 84,498,936 | MDQ6VXNlcjg0NDk4OTM2 | User | false |
openclaw/openclaw | 4,075,083,780 | I_kwDOQb6kR87y5NgE | 45,865 | https://github.com/openclaw/openclaw/issues/45865 | https://api.github.com/repos/openclaw/openclaw/issues/45865 | [Feature]: api.runtime.llm() — Plugin SDK inference method | ### Summary
Add api.runtime.llm() to the Plugin SDK so plugins can make LLM calls through OpenClaw's existing inference stack instead of calling provider APIs directly.
### Problem to solve
Plugins that need LLM calls (extraction, classification, summarisation, transformation) currently have no SDK method for infere... | open | reopened | false | 3 | [
"enhancement"
] | [] | 2026-03-14T08:06:10Z | 2026-05-04T00:08:06Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | afroboyX | 8,142,609 | MDQ6VXNlcjgxNDI2MDk= | User | false |
openclaw/openclaw | 4,075,854,114 | I_kwDOQb6kR87y8Jki | 46,184 | https://github.com/openclaw/openclaw/issues/46184 | https://api.github.com/repos/openclaw/openclaw/issues/46184 | [Bug]: google-gemini-cli OAuth fails with TypeError: fetch failed behind HTTP proxy on macOS (even with pinDns: false & 60s timeout) | ### Bug type
Regression (worked before, now fails)
### Summary
On macOS behind an HTTP(S) proxy, google-gemini-cli OAuth consistently fails with:
TypeError: fetch failed
This happens even though:
• The proxy works for all relevant Google endpoints (verified via curl)
• The oauth.ts in @google/gemini-cli has already... | open | null | false | 2 | [
"bug",
"regression"
] | [
"vincentkoc"
] | 2026-03-14T13:45:27Z | 2026-05-04T00:43:40Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | EvetteYoung | 51,390,883 | MDQ6VXNlcjUxMzkwODgz | User | false |
openclaw/openclaw | 4,075,896,485 | I_kwDOQb6kR87y8T6l | 46,217 | https://github.com/openclaw/openclaw/issues/46217 | https://api.github.com/repos/openclaw/openclaw/issues/46217 | [Feature]: 更完善的汉化，"Settings" section，Chinese | ### Summary
Excuse me, the configuration items in the "Settings" section of the Chinese backend are still in English. Are there any plans to translate this part into Chinese?
请问，中文时候的后台，”设置“模块的配置还都是英文，这部分有计划翻译成中文吗？
### Problem to solve
Right now, everything is still in English. For most people with poor English ski... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-14T14:09:48Z | 2026-05-04T00:31:45Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | jujunchen | 29,937,421 | MDQ6VXNlcjI5OTM3NDIx | User | false |
openclaw/openclaw | 4,077,985,060 | I_kwDOQb6kR87zER0k | 47,120 | https://github.com/openclaw/openclaw/issues/47120 | https://api.github.com/repos/openclaw/openclaw/issues/47120 | [Feature Request] Session List: Show timestamp + last message preview | ### Summary
Add timestamp and last message preview to session list in Control UI for easier identification
### Problem to solve
Currently the session list in Control UI shows long technical IDs like `openai-user:dingtalk-connector:default:333446632538210751:1773...`
This makes it very hard to:
- Distinguish between... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-15T09:07:38Z | 2026-05-04T02:38:28Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | SiriuSM00N | 265,539,072 | U_kgDOD9POAA | User | false |
openclaw/openclaw | 4,265,726,159 | I_kwDOQb6kR87-QdDP | 66,923 | https://github.com/openclaw/openclaw/issues/66923 | https://api.github.com/repos/openclaw/openclaw/issues/66923 | OpenClaw Model Registry: GPT-4.1 family passes through unmanaged | ## Issue
OpenClaw 2026.4.7 uses the GPT-5.4 model family internally for OpenAI workloads. When a user configures a model ID outside the registry — such as `openai/gpt-4.1-nano`, `openai/gpt-4.1-mini`, or `openai/gpt-4.1` — the ID passes through to OpenAI's API raw, without OpenClaw's exec profile, tool-use optimizatio... | closed | completed | true | 1 | [] | [] | 2026-04-15T01:43:59Z | 2026-05-03T04:35:52Z | 2026-04-30T17:04:43Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | perlowjanv | 201,165,108 | U_kgDOC_2JNA | User | false |
openclaw/openclaw | 4,086,005,657 | I_kwDOQb6kR87zi3-Z | 48,700 | https://github.com/openclaw/openclaw/issues/48700 | https://api.github.com/repos/openclaw/openclaw/issues/48700 | [Feature]: | ### Summary
## Feature Request: WebChat file upload support
### What
Add file upload capability to WebChat channel (Control UI chat dialog) to support:
- Excel (.xlsx, .xls)
- Word (.docx, .doc)
- PDF (.pdf)
- PowerPoint (.pptx, .ppt)
### Why
Currently WebChat only supports text and basic emojis. Users need to upl... | closed | completed | false | 1 | [
"enhancement"
] | [] | 2026-03-17T03:35:20Z | 2026-05-04T14:56:34Z | 2026-05-04T14:56:34Z | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | chenbagain | 189,839,208 | U_kgDOC1C3aA | User | false |
openclaw/openclaw | 4,359,950,600 | I_kwDOQb6kR88AAAABA9-RCA | 75,176 | https://github.com/openclaw/openclaw/issues/75176 | https://api.github.com/repos/openclaw/openclaw/issues/75176 | [Bug]: Active-memory on multi-agent Telegram gateway is inconsistent (ok / empty / timeout) and not safe to enable globally | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Summary
On a multi-agent OpenClaw gateway using Telegram direct chats, the active-memory plugin appears partially functional but unreliable when enabled: across real agent sessions it produces a mix of status=... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-04-30T17:12:39Z | 2026-05-03T04:35:48Z | 2026-04-30T17:20:03Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | MIHHHMIH | 158,168,620 | U_kgDOCW12LA | User | false |
openclaw/openclaw | 4,359,986,231 | I_kwDOQb6kR88AAAABA-AcNw | 75,178 | https://github.com/openclaw/openclaw/issues/75178 | https://api.github.com/repos/openclaw/openclaw/issues/75178 | Bug: WhatsApp auto-reply delivery silently fails in group chats (typing indicator only, message never sent) | ## Description
WhatsApp auto-reply delivery silently fails in group chats. The bot shows a "typing" indicator but the message is never actually delivered to the group. This happens **only for auto-reply responses** — manual sends via the `message` tool work fine.
## Environment
- OpenClaw: 2026.4.27
- Node: 22.22.1
-... | closed | not_planned | true | 1 | [] | [] | 2026-04-30T17:18:54Z | 2026-05-03T04:35:47Z | 2026-04-30T17:22:33Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | sizoune | 13,961,066 | MDQ6VXNlcjEzOTYxMDY2 | User | false |
openclaw/openclaw | 4,066,411,247 | I_kwDOQb6kR87yYILv | 44,302 | https://github.com/openclaw/openclaw/issues/44302 | https://api.github.com/repos/openclaw/openclaw/issues/44302 | [Feature]: Optional webchat-to-external-channel mirroring | ## Summary
Add an opt-in config flag that mirrors webchat (control UI) replies to the session's last active external channel (Discord, Telegram, etc.).
## Problem to solve
Between v3.2 and v3.11, replies sent from the control UI were incidentally mirrored to external channels as a side effect of the routing layer no... | open | null | false | 1 | [
"enhancement"
] | [
"vincentkoc"
] | 2026-03-12T18:16:28Z | 2026-05-04T12:46:55Z | null | CONTRIBUTOR | null | 20260504T180039Z | 2026-05-04T18:00:39Z | rbutera | 6,047,293 | MDQ6VXNlcjYwNDcyOTM= | User | false |
openclaw/openclaw | 4,068,586,491 | I_kwDOQb6kR87ygbP7 | 44,603 | https://github.com/openclaw/openclaw/issues/44603 | https://api.github.com/repos/openclaw/openclaw/issues/44603 | [Feature]: pty setting | ### Summary
{
"tools": {
"exec": {
"timeoutSec": 900,
"pty": true // 全局默认启用 pty
}
}
}
See the intermediate results of loss output， i cannot use pty
### Problem to solve
I have to wait for a long time to see the results, and users are unaware of what's happening during this process.
### Pr... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-13T03:26:34Z | 2026-05-04T13:18:44Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | 520jefferson | 5,691,554 | MDQ6VXNlcjU2OTE1NTQ= | User | false |
openclaw/openclaw | 4,068,767,188 | I_kwDOQb6kR87yhHXU | 44,655 | https://github.com/openclaw/openclaw/issues/44655 | https://api.github.com/repos/openclaw/openclaw/issues/44655 | [Feature]: Implement Dynamic Model Discovery for Zhipu (Z.ai) Provider via /models Endpoint | ### Summary
I am currently using OpenClaw with the Zhipu (Z.ai) International Coding Plan. I noticed that the current implementation seems to use a hardcoded list of supported models for the Zhipu provider, which often lags behind the official updates (e.g., missing glm-4.7-flash or the latest versions).
However, the... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-13T04:29:51Z | 2026-05-04T13:18:45Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | yorha59 | 24,535,089 | MDQ6VXNlcjI0NTM1MDg5 | User | false |
openclaw/openclaw | 4,071,947,110 | I_kwDOQb6kR87ytPtm | 45,258 | https://github.com/openclaw/openclaw/issues/45258 | https://api.github.com/repos/openclaw/openclaw/issues/45258 | [Feature]: | ### Summary
Subject: Feature Request: Per-agent cron job isolation in tool responses
Hi,
I'm building a multi-agent desktop application (OpenChat) on top of openclaw and have run into a data isolation issue with the cron tool that I believe would benefit from a server-side fix.
Problem:
When an agent calls cron({ ... | closed | not_planned | false | 1 | [
"enhancement"
] | [] | 2026-03-13T15:54:19Z | 2026-05-04T01:08:33Z | 2026-05-04T01:08:33Z | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | jjx0369-crypto | 256,642,330 | U_kgDOD0wNGg | User | false |
openclaw/openclaw | 4,073,730,282 | I_kwDOQb6kR87y0DDq | 45,508 | https://github.com/openclaw/openclaw/issues/45508 | https://api.github.com/repos/openclaw/openclaw/issues/45508 | [Feature]: Self-hosted STT/TTS provider support in webchat (Route webchat TTS through the gateway instead of browser Speech API) | ### Summary
The webchat "Read aloud" button and "Voice input" mic both use the browser's built-in Web Speech APIs (SpeechSynthesis and Web Speech API). This completely ignores any TTS/STT configuration in openclaw.json, making self-hosted voice setups invisible to the webchat frontend.
### Problem to solve
Users who... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-03-13T22:43:30Z | 2026-05-04T00:07:00Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | mcfex | 112,344,317 | U_kgDOBrI8_Q | User | false |
openclaw/openclaw | 4,073,815,950 | I_kwDOQb6kR87y0X-O | 45,519 | https://github.com/openclaw/openclaw/issues/45519 | https://api.github.com/repos/openclaw/openclaw/issues/45519 | [Feature]: Collapsible messages in Control UI chat | ### Summary
Collapsible messages in Control UI chat
### Problem to solve
When chatting with the agent through the Control UI, long conversations require excessive scrolling. The agent's responses (especially code blocks, tables, or detailed explanations) take up a lot of vertical space, making it hard to review earl... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-03-13T23:07:38Z | 2026-05-04T00:12:30Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | lazyom | 266,662,874 | U_kgDOD-Tz2g | User | false |
openclaw/openclaw | 4,074,991,853 | I_kwDOQb6kR87y43Dt | 45,839 | https://github.com/openclaw/openclaw/issues/45839 | https://api.github.com/repos/openclaw/openclaw/issues/45839 | [Feature Request] Telegram inline buttons for pairing approve/deny | ### Summary
When a WhatsApp pairing request comes in, the agent notifies the operator via Telegram with the sender's name, phone number, and pairing code. Currently, approve/deny can only be done via CLI commands pasted as text.
... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-03-14T07:25:30Z | 2026-05-04T00:08:01Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | lpg-tech | 54,527,242 | MDQ6VXNlcjU0NTI3MjQy | User | false |
openclaw/openclaw | 4,074,994,545 | I_kwDOQb6kR87y43tx | 45,841 | https://github.com/openclaw/openclaw/issues/45841 | https://api.github.com/repos/openclaw/openclaw/issues/45841 | [Feature]: Sandboxing + ACP | ### Summary
Sandboxed OpenClaw sessions should be able to spawn ACP sessions.
It's currently a [documented limitation](https://docs.openclaw.ai/tools/acp-agents#sandbox-compatibility) that breaks the use of 2 very important openclaw features - sandboxing and ACP.
ACP is great for better interop with coding harnesses,... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-03-14T07:26:47Z | 2026-05-04T00:13:06Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | Thinkscape | 270,528 | MDQ6VXNlcjI3MDUyOA== | User | false |
openclaw/openclaw | 4,075,031,584 | I_kwDOQb6kR87y5Awg | 45,854 | https://github.com/openclaw/openclaw/issues/45854 | https://api.github.com/repos/openclaw/openclaw/issues/45854 | [Feature]: Android Node Tools & Session Enhancement | ### Summary
## 1. Android Node Tool Enhancement
I plan to enhance the tools for Android nodes to support more high-value scenarios
For example,
```
1. Generate a weekly summary, view the schedule, call history, and text messages of the past week
2. Where is my phone? Let it vibrate and make a sound
3. Help me find th... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-03-14T07:43:26Z | 2026-05-04T00:08:05Z | null | CONTRIBUTOR | null | 20260504T030031Z | 2026-05-04T03:00:31Z | lixuankai | 8,060,486 | MDQ6VXNlcjgwNjA0ODY= | User | false |
openclaw/openclaw | 4,076,240,755 | I_kwDOQb6kR87y9n9z | 46,398 | https://github.com/openclaw/openclaw/issues/46398 | https://api.github.com/repos/openclaw/openclaw/issues/46398 | [Feature]: Add persistent token usage indicator to Control UI / WebChat | ### Summary
## Problem
When using Ollama models with limited context windows, users have no way to monitor token usage in real-time from the chat UI. The context can suddenly exceed the max limit, causing errors or unexpected behavior.
### Problem to solve
## Expected Behavior
A persistent token counter displa... | open | null | false | 1 | [
"enhancement"
] | [
"BunsDev"
] | 2026-03-14T16:57:24Z | 2026-05-05T03:49:06Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | walterwkchoy | 180,582,537 | U_kgDOCsN4iQ | User | false |
openclaw/openclaw | 4,076,693,298 | I_kwDOQb6kR87y_Wcy | 46,565 | https://github.com/openclaw/openclaw/issues/46565 | https://api.github.com/repos/openclaw/openclaw/issues/46565 | [Feature]: deterministic context injection for bootstrap files and session-type filtering | **Summary**
Three platform-level gaps affect any user running autonomous agents with heartbeats, subagents, or scheduled work: static bootstrap files, no session-type filtering, and missing time context.
**Problem**
1. Bootstrap files don't support dynamic injection
Skills support !command`` for dynamic context, but... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-03-14T20:44:21Z | 2026-05-04T00:44:17Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | Joncik91 | 190,874,150 | U_kgDOC2CCJg | User | false |
openclaw/openclaw | 4,076,978,857 | I_kwDOQb6kR87zAcKp | 46,630 | https://github.com/openclaw/openclaw/issues/46630 | https://api.github.com/repos/openclaw/openclaw/issues/46630 | [Feature]: Shortcut/Keyword Sidebar Panel | ### Summary
Add a scrollable sidebar panel to the webchat dashboard displaying user-defined shortcut keywords with click-to-insert and hover tooltips.
### Problem to solve
Users build up trigger keywords over time (e.g., "Onboarding", "Status", "Security Audit") that their AI assistant recognizes and acts on. These ... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-14T22:59:07Z | 2026-05-04T00:44:20Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | gjoham | 181,919,056 | U_kgDOCtfdUA | User | false |
openclaw/openclaw | 4,077,100,643 | I_kwDOQb6kR87zA55j | 46,661 | https://github.com/openclaw/openclaw/issues/46661 | https://api.github.com/repos/openclaw/openclaw/issues/46661 | [Feature]: Support Custom ASR (Speech-to-Text) Server Configuration | ### Summary
Allow users to configure a custom ASR (Automatic Speech Recognition) server endpoint instead of relying on the built-in macOS speech recognition.
### Problem to solve
The current macOS built-in ASR (via the Apple Speech Recognition framework) has several limitations:
1. **Poor accuracy** for technical v... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-03-14T23:52:15Z | 2026-05-04T00:44:27Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | sheldon123z | 42,455,304 | MDQ6VXNlcjQyNDU1MzA0 | User | false |
openclaw/openclaw | 4,077,116,705 | I_kwDOQb6kR87zA90h | 46,664 | https://github.com/openclaw/openclaw/issues/46664 | https://api.github.com/repos/openclaw/openclaw/issues/46664 | [Feature]: Offline Command Queue for iOS/Android Node App | ### Summary
Add an offline command queue to the OpenClaw mobile node app (iOS/Android) that captures user intents while disconnected and automatically syncs them to the gateway for execution when connectivity is restored.
### Problem to solve
Many OpenClaw users operate in environments with intermittent or no cellul... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-03-15T00:00:29Z | 2026-05-04T00:44:29Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | sbmilburn | 11,755,869 | MDQ6VXNlcjExNzU1ODY5 | User | false |
openclaw/openclaw | 4,077,327,493 | I_kwDOQb6kR87zBxSF | 46,724 | https://github.com/openclaw/openclaw/issues/46724 | https://api.github.com/repos/openclaw/openclaw/issues/46724 | [Feature]: Gateway-to-Gateway Interconnection / Federation | ### Summary
I'm wondering if it would be valuable to support **Gateway-to-Gateway interconnection** (or "federation").
### Problem to solve
Currently, each OpenClaw Gateway operates as an isolated instance. It accepts connections from clients (CLI, desktop/mobile apps) and agents, but there is no direct communicatio... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-15T01:47:57Z | 2026-05-04T00:32:54Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | mcocdaa | 144,591,624 | U_kgDOCJ5LCA | User | false |
openclaw/openclaw | 4,078,468,995 | I_kwDOQb6kR87zGH-D | 47,386 | https://github.com/openclaw/openclaw/issues/47386 | https://api.github.com/repos/openclaw/openclaw/issues/47386 | [Feature] Webchat UI: Add collapsible tool output summary mode | ### Summary
Add collapsible summary mode for tool executions in webchat UI to reduce visual noise
### Problem to solve
Currently the webchat UI displays each tool execution as separate messages (e.g., "1 tool exec", "tool output"), creating excessive visual noise. When an agent runs multiple tools in one turn, users... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-15T14:18:01Z | 2026-05-04T01:32:42Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | owen-cjw | 92,360,573 | U_kgDOBYFPfQ | User | false |
openclaw/openclaw | 4,078,670,204 | I_kwDOQb6kR87zG5F8 | 47,461 | https://github.com/openclaw/openclaw/issues/47461 | https://api.github.com/repos/openclaw/openclaw/issues/47461 | [Telegram] 机器人头像更改后无法同步到 Telegram 客户端 | ### Summary
实现 Telegram 机器人头像与 Web UI 的双向同步
### Problem to solve
当前头像同步是单向的（Telegram → Web UI），用户在 Web UI 中更改头像后，Telegram 客户端仍然显示旧头像，导致品牌标识不一致。用户必须手动通过 @BotFather 设置头像，体验割裂。
### Proposed solution
在 Gateway 或 Control UI 中添加监听逻辑：当检测到 `avatars/avatar.jpg` 文件变更时，自动调用 Telegram Bot API (`setChatPhoto`) 更新机器人头像。实现双向同步：We... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-15T16:12:43Z | 2026-05-04T02:26:42Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | wangjianjq | 259,905,141 | U_kgDOD33WdQ | User | false |
openclaw/openclaw | 4,079,417,381 | I_kwDOQb6kR87zJvgl | 47,684 | https://github.com/openclaw/openclaw/issues/47684 | https://api.github.com/repos/openclaw/openclaw/issues/47684 | Add auth sync command or auto-sync for API key rotation across agents | ### Summary
When updating env.ANTHROPIC_API_KEY via openclaw config set, the change doesn't propagate to agent auth-profiles.json files, causing silent 401 errors.
### Problem to solve
Currently, when a user updates their Anthropic API key via `openclaw config set env.ANTHROPIC_API_KEY "new-key"`, the change is writ... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-15T23:12:07Z | 2026-05-04T01:33:23Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | AnthonyFrancis | 3,010,099 | MDQ6VXNlcjMwMTAwOTk= | User | false |
openclaw/openclaw | 4,079,797,272 | I_kwDOQb6kR87zLMQY | 47,750 | https://github.com/openclaw/openclaw/issues/47750 | https://api.github.com/repos/openclaw/openclaw/issues/47750 | [Feature]: Non-interactive command to use Codex OAuth for model provider | ### Summary
Add non-interactive way to setup model provider using OpenAI Codex OAuth (e.g. just by supplying something like OPENAI_CODEX_OAUTH_CALLBACK_URL in the env)
### Problem to solve
I have seen the docs: https://docs.openclaw.ai/providers/openai
Currently there's a non-interactive way to set API key, which i... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-16T02:02:45Z | 2026-05-04T14:02:13Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | terryds | 11,571,395 | MDQ6VXNlcjExNTcxMzk1 | User | false |
openclaw/openclaw | 4,086,973,787 | I_kwDOQb6kR87zmkVb | 48,855 | https://github.com/openclaw/openclaw/issues/48855 | https://api.github.com/repos/openclaw/openclaw/issues/48855 | [Feature]: Android Gateway Support | ### Summary
## Summary
Requesting gateway service support for Android devices (via Termux) to enable dashboard access and remote connections without requiring a separate PC or VPS.
### Problem to solve
## Possible Solutions
1. Native Android gateway support (if technically feasible)
2. Alternative lightweight gatewa... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-03-17T08:26:50Z | 2026-05-04T14:55:41Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | Vasanthdev2004 | 148,849,890 | U_kgDOCN9E4g | User | false |
openclaw/openclaw | 4,099,249,374 | I_kwDOQb6kR870VZTe | 50,205 | https://github.com/openclaw/openclaw/issues/50205 | https://api.github.com/repos/openclaw/openclaw/issues/50205 | [Feature]: Support configurable request labels for Gemini API calls (GCP billing tracking) | ### Summary
## Summary
Add support for attaching custom labels to outbound Gemini API (generateContent) requests, as recommended by Google Cloud for billing cost attribution.
## Motivation
GCP billing reports do not support grouping costs by API key directly. The recommended approach from Google Cloud Support is to... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-03-19T04:40:11Z | 2026-05-04T20:22:58Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | ikujyh | 3,983,958 | MDQ6VXNlcjM5ODM5NTg= | User | false |
openclaw/openclaw | 4,099,940,988 | I_kwDOQb6kR870YCJ8 | 50,310 | https://github.com/openclaw/openclaw/issues/50310 | https://api.github.com/repos/openclaw/openclaw/issues/50310 | [Feature]: Auto-cleanup stale/orphaned Discord thread sessions | ### Summary
When a Discord thread is archived or deleted, OpenClaw resets or marks the associated session (appending `.reset` or `.deleted` suffix to the `.jsonl` file), but the session entry in `sessions.json` is never removed. Over time this results in orphaned keys in `sessions.json` that point to non-existent or i... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-03-19T07:46:18Z | 2026-05-04T20:58:59Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | alainlx2024 | 163,292,621 | U_kgDOCbulzQ | User | false |
openclaw/openclaw | 4,360,090,149 | I_kwDOQb6kR88AAAABA-GyJQ | 75,188 | https://github.com/openclaw/openclaw/issues/75188 | https://api.github.com/repos/openclaw/openclaw/issues/75188 | Bug: openai/gpt-image-2 via Codex OAuth — Tool choice 'image_generation' not found in 'tools' parameter | ## Bug: openai/gpt-image-2 via Codex OAuth — "Tool choice 'image_generation' not found in 'tools' parameter"
### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Environment
- **OS:** macOS
- **OpenClaw version:** 2026.4.27
- **Node:** v25.8.1
### Summary
When using `open... | closed | completed | true | 1 | [] | [] | 2026-04-30T17:37:56Z | 2026-05-03T04:35:46Z | 2026-04-30T17:42:50Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | martinning-lab | 267,717,844 | U_kgDOD_UM1A | User | false |
openclaw/openclaw | 4,360,106,570 | I_kwDOQb6kR88AAAABA-HySg | 75,190 | https://github.com/openclaw/openclaw/issues/75190 | https://api.github.com/repos/openclaw/openclaw/issues/75190 | [Bug]: Telegram messages from bot disappear after delivery — regression in 2026.4.27 (upgrade from 4.15) | ### Bug type
Regression (worked before, now fails)
### Summary
After upgrading from OpenClaw **2026.4.15** to **2026.4.27**, messages delivered by the bot in Telegram DM intermittently disappear after being shown. The user sees the message briefly, then it vanishes from the chat. This was not observed on 4.15.
### ... | closed | completed | true | 1 | [] | [] | 2026-04-30T17:40:54Z | 2026-05-03T04:35:45Z | 2026-04-30T17:48:39Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | jabisu-sudo | 262,010,920 | U_kgDOD534KA | User | false |
openclaw/openclaw | 4,077,546,476 | I_kwDOQb6kR87zCmvs | 46,851 | https://github.com/openclaw/openclaw/issues/46851 | https://api.github.com/repos/openclaw/openclaw/issues/46851 | [Feature]: Feature Request: Configurable fixed-interval progress heartbeat for long-running tasks (e.g. every 60s) | ### Summary
When the assistant runs a long task in chat surfaces (e.g., QQ, Telegram), users often cannot tell whether it is still running or unresponsive.
I’d like OpenClaw to support a built-in fixed-interval progress heartbeat during a single long-running turn, so users can reliably see that
execut... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-03-15T04:15:03Z | 2026-05-03T04:35:43Z | 2026-04-30T17:59:19Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | tangkk | 1,814,604 | MDQ6VXNlcjE4MTQ2MDQ= | User | false |
openclaw/openclaw | 4,360,127,479 | I_kwDOQb6kR88AAAABA-JD9w | 75,193 | https://github.com/openclaw/openclaw/issues/75193 | https://api.github.com/repos/openclaw/openclaw/issues/75193 | [Bug]: Trailing assistant turn causes silent provider failover on every first Anthropic request in persisted sessions | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Gateway heartbeat carries the prior session's trailing assistant turn into the next provider call without appending a synthetic user event.
Anthropic claude-opus-4-* enforces strict role ordering and 400s with:
"This model ... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-30T17:44:22Z | 2026-05-03T04:35:40Z | 2026-04-30T17:48:45Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | andydysna | 151,802,995 | U_kgDOCQxUcw | User | false |
openclaw/openclaw | 4,099,857,078 | I_kwDOQb6kR870Xtq2 | 50,298 | https://github.com/openclaw/openclaw/issues/50298 | https://api.github.com/repos/openclaw/openclaw/issues/50298 | [Feature]: When messages are merged and processed, the agent event is expected to carry the original runId(s) | ### Summary
It is hoped that after the conversation message is queued, if a new runId is created when the agent is actually run (for example, multiple queued messages are merged and processed), a set of original runIds can be carried in the websocket event.
### Problem to solve
My scenario is to establish a dialogue... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-03-19T07:23:29Z | 2026-05-04T18:22:29Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | yourse007 | 38,979,056 | MDQ6VXNlcjM4OTc5MDU2 | User | false |
openclaw/openclaw | 4,104,265,792 | I_kwDOQb6kR870oiBA | 50,646 | https://github.com/openclaw/openclaw/issues/50646 | https://api.github.com/repos/openclaw/openclaw/issues/50646 | [Feature]: sessions.list returns misleading results when visibility is restricted (no indication of scope) | ### Summary
sessions.list (and related tools) can return incomplete results due to visibility restrictions (e.g. default "tree") without indicating that the output is scoped or filtered.
### Problem to solve
When tools.sessions.visibility is set to its default ("tree"), session-related tools (e.g. sessions.list) onl... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-03-19T21:17:00Z | 2026-05-04T18:23:27Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | playaopindo-dev | 257,603,176 | U_kgDOD1q2aA | User | false |
openclaw/openclaw | 4,104,811,364 | I_kwDOQb6kR870qnNk | 50,701 | https://github.com/openclaw/openclaw/issues/50701 | https://api.github.com/repos/openclaw/openclaw/issues/50701 | [Feature]: WebUI usage history appears to reset after openclaw update; needs transcript-backed historical lineage rollups | ### Summary
## Summary
After running openclaw update, the WebUI usage area appears to lose prior usage history. In practice, the underlying transcript data and usage records still exist on disk, but the WebUI does not present them as a continuous historical view.
This appears to be a backend + UX gap rather than tru... | open | null | false | 3 | [
"enhancement"
] | [
"BunsDev"
] | 2026-03-19T23:38:38Z | 2026-05-04T20:23:02Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | dev-gideon-llc | 261,372,791 | U_kgDOD5Q7dw | User | false |
openclaw/openclaw | 4,106,753,989 | I_kwDOQb6kR870yBfF | 50,963 | https://github.com/openclaw/openclaw/issues/50963 | https://api.github.com/repos/openclaw/openclaw/issues/50963 | [Feature]: Subagent session reuse for expert agents (avoid new session per spawn) | ### Summary
Allow expert subagents to reuse an existing session instead of creating a new one every sessions_spawn(mode="run")
### Problem to solve
Every sessions_spawn(mode="run") generates a new random UUID (subagent-spawn.ts:374), so expert agents lose all conversation context between spawns. The existing mode="s... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-03-20T09:13:20Z | 2026-05-04T17:48:36Z | null | NONE | null | 20260504T180039Z | 2026-05-04T18:00:39Z | TomsChen | 29,362,516 | MDQ6VXNlcjI5MzYyNTE2 | User | false |
openclaw/openclaw | 4,360,452,638 | I_kwDOQb6kR88AAAABA-c6Hg | 75,207 | https://github.com/openclaw/openclaw/issues/75207 | https://api.github.com/repos/openclaw/openclaw/issues/75207 | [Bug]: Discord replies generated locally but not delivered/rendered in Discord | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Summary
Discord inbound works and the model generates replies, but final assistant replies do not appear in D... | closed | not_planned | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-30T18:37:51Z | 2026-05-03T04:35:36Z | 2026-04-30T18:42:02Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | aa-on-ai | 94,061,050 | U_kgDOBZtB-g | User | false |
openclaw/openclaw | 4,097,209,753 | I_kwDOQb6kR870NnWZ | 49,982 | https://github.com/openclaw/openclaw/issues/49982 | https://api.github.com/repos/openclaw/openclaw/issues/49982 | [Feature]: Auto-compress image attachments | ### Summary
Currently, images exceeding 5 MB are rejected with an error. Other AI assistants automatically compress images client-side before processing. It would improve UX if OpenClaw did the same. Greatly appreciate it.
### Problem to solve
Example error:
Error: attachment image: exceeds size limit (8885860 ... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-03-18T19:11:11Z | 2026-05-03T04:35:34Z | 2026-04-30T18:44:50Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | aabbasnia-cpu | 269,288,258 | U_kgDOEA0DQg | User | false |
openclaw/openclaw | 4,105,885,829 | I_kwDOQb6kR870utiF | 50,865 | https://github.com/openclaw/openclaw/issues/50865 | https://api.github.com/repos/openclaw/openclaw/issues/50865 | [Feature]: Old Sessions | ### Summary
Add a "Restore Deleted Session" to the webui to easily change to an old chat.
### Problem to solve
It is so annoying if you accidentally hit the plus button and your session is gone.
### Proposed solution
Make old sessions accessible in gui
### Alternatives considered
_No response_
### Impact
... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-03-20T05:30:07Z | 2026-05-03T04:35:33Z | 2026-04-30T18:45:04Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | averageaidude | 215,313,238 | U_kgDODNVrVg | User | false |
openclaw/openclaw | 4,360,610,793 | I_kwDOQb6kR88AAAABA-mj6Q | 75,224 | https://github.com/openclaw/openclaw/issues/75224 | https://api.github.com/repos/openclaw/openclaw/issues/75224 | [Bug]: 2026.4.26 hangs in futex_wait_queue on docker restart of an existing container; fresh boot works | ## TL;DR
`docker restart` of an existing OpenClaw 2026.4.26 container deadlocks at `[gateway] starting…`. All Node threads sit idle in `futex_wait_queue`, the event loop is empty, and the gateway never reaches `http server listening`. A fresh image pull boots fine; the hang only manifests on restart of a previously-ru... | closed | completed | true | 1 | [] | [] | 2026-04-30T19:05:26Z | 2026-05-03T04:35:26Z | 2026-04-30T19:10:05Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Jason-Vaughan | 95,194,903 | U_kgDOBayPFw | User | false |
openclaw/openclaw | 4,360,130,920 | I_kwDOQb6kR88AAAABA-JRaA | 75,194 | https://github.com/openclaw/openclaw/issues/75194 | https://api.github.com/repos/openclaw/openclaw/issues/75194 | [Bug]: Telegram reply/quote context not passed to assistant — regression in 2026.4.27 (upgrade from 4.15) | ### Bug type
Regression (worked before, now fails)
### Summary
After upgrading from OpenClaw **2026.4.15** to **2026.4.27**, when a user replies to (quotes) a specific bot message in Telegram, the reply context does not reach the assistant correctly. The assistant cannot tell which message the user is referring to, ... | closed | completed | true | 3 | [] | [] | 2026-04-30T17:44:59Z | 2026-05-03T04:33:27Z | 2026-04-30T19:19:40Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | jabisu-sudo | 262,010,920 | U_kgDOD534KA | User | false |
openclaw/openclaw | 4,360,564,079 | I_kwDOQb6kR88AAAABA-jtbw | 75,216 | https://github.com/openclaw/openclaw/issues/75216 | https://api.github.com/repos/openclaw/openclaw/issues/75216 | [Feature]: Rolling context OPTION instead of compaction | ### Summary
Create an Option for a Rolling Stack session context as an alternative to the current compaction implementation
### Problem to solve
Dealing with the rather small session context window (in comparison to most cloud model context windows) is a bit troubling if cloud models are doing heavy lifting. Tool c... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-30T18:56:17Z | 2026-05-03T04:35:19Z | 2026-04-30T19:19:47Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | vesperlightre | 253,157,875 | U_kgDODxbh8w | User | false |
openclaw/openclaw | 4,360,694,044 | I_kwDOQb6kR88AAAABA-rpHA | 75,229 | https://github.com/openclaw/openclaw/issues/75229 | https://api.github.com/repos/openclaw/openclaw/issues/75229 | feat: make TUI streaming watchdog timeout configurable | ## Problem
The TUI streaming watchdog in the TUI bundle has a hardcoded timeout of 30 seconds:
```js
const DEFAULT_STREAMING_WATCHDOG_MS = 3e4; // 30000ms
```
When no stream updates arrive within 30s, the watchdog fires:
```
streaming watchdog: no stream updates for 30s; resetting status.
The backend may have dropp... | closed | not_planned | true | 1 | [] | [] | 2026-04-30T19:20:40Z | 2026-05-03T04:35:18Z | 2026-04-30T19:24:06Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | PingboJiang | 2,818,917 | MDQ6VXNlcjI4MTg5MTc= | User | false |
openclaw/openclaw | 4,360,711,854 | I_kwDOQb6kR88AAAABA-surg | 75,232 | https://github.com/openclaw/openclaw/issues/75232 | https://api.github.com/repos/openclaw/openclaw/issues/75232 | [Bug]: Control UI leaks internal async completion events into chat transcript | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On OpenClaw 2026.4.24 (cbcfdf6), I’m seeing internal async exec completion events appear in the visible Control UI chat transcript even when user delivery is disabled.
This happens both:
when using the Contro... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-04-30T19:24:20Z | 2026-05-03T04:35:18Z | 2026-04-30T19:27:45Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | cscarfie | 5,605,963 | MDQ6VXNlcjU2MDU5NjM= | User | false |
openclaw/openclaw | 4,360,730,400 | I_kwDOQb6kR88AAAABA-t3IA | 75,234 | https://github.com/openclaw/openclaw/issues/75234 | https://api.github.com/repos/openclaw/openclaw/issues/75234 | [ignore] test - accidental scope check, closing | test | closed | not_planned | true | 0 | [] | [] | 2026-04-30T19:28:01Z | 2026-05-03T04:35:17Z | 2026-04-30T19:29:08Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | EVEHetzner | 260,284,917 | U_kgDOD4Oh9Q | User | false |
openclaw/openclaw | 4,360,711,049 | I_kwDOQb6kR88AAAABA-sriQ | 75,231 | https://github.com/openclaw/openclaw/issues/75231 | https://api.github.com/repos/openclaw/openclaw/issues/75231 | [Bug]: Plugin runtime-deps tree staged in waves during openclaw update; gateway restart fires before tree is complete, causing channel-loader ENOENT spam (~60+ min unhealthy gateway) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
On macOS, `openclaw update` from 2026.4.26 to 2026.4.27 fires the LaunchAgent supervisor restart before the plugin runtime-deps `dist/` tree finishes staging. The gateway comes up against an incomplete tree, eight bundled ch... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-30T19:24:11Z | 2026-05-03T04:35:16Z | 2026-04-30T19:29:39Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | nphyde2 | 269,757,124 | U_kgDOEBQqxA | User | false |
openclaw/openclaw | 4,360,763,612 | I_kwDOQb6kR88AAAABA-v43A | 75,237 | https://github.com/openclaw/openclaw/issues/75237 | https://api.github.com/repos/openclaw/openclaw/issues/75237 | 2026.4.27: Missing plugins/loader.js causes 'Could not load plugin runtime loader for channel setup fallback' on startup | ## Summary
After upgrading to `2026.4.27`, `openclaw status` (and any CLI command) fails immediately with:
```
[openclaw] Failed to start CLI: Error: Could not load plugin runtime loader for channel setup fallback.
at loadPluginLoaderModule (file:///...openclaw/dist/read-only-B4EkEtUx.js:32:8)
at resolveReadO... | closed | completed | true | 1 | [] | [] | 2026-04-30T19:33:35Z | 2026-05-03T04:35:16Z | 2026-04-30T19:37:50Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | CRamsan | 695,667 | MDQ6VXNlcjY5NTY2Nw== | User | false |
openclaw/openclaw | 4,360,669,508 | I_kwDOQb6kR88AAAABA-qJRA | 75,227 | https://github.com/openclaw/openclaw/issues/75227 | https://api.github.com/repos/openclaw/openclaw/issues/75227 | /new and /reset slash commands rotate openclaw session but do not archive the LCM conversation | ## Summary
`/new` (and `/reset`) on Discord and Telegram channels rotates the openclaw session JSONL but does **not** archive the underlying LCM conversation. The same `conversation_id` keeps absorbing across "fresh" sessions, accumulating thousands of messages over hours/days. From a user perspective, "reset" doesn't... | closed | not_planned | true | 1 | [] | [] | 2026-04-30T19:16:14Z | 2026-05-03T04:35:14Z | 2026-04-30T19:40:57Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | jesuskannolis | 10,068,263 | MDQ6VXNlcjEwMDY4MjYz | User | false |
openclaw/openclaw | 4,360,784,088 | I_kwDOQb6kR88AAAABA-xI2A | 75,238 | https://github.com/openclaw/openclaw/issues/75238 | https://api.github.com/repos/openclaw/openclaw/issues/75238 | Telegram group topic delivery broken in 2026.4.27 - replies sent to DM instead of topic | ## Describe the bug
After upgrading to **2026.4.27**, assistant replies in Telegram group topics are not delivered to the correct chat destination. Instead of being sent to the group topic (with correct `chat_id` and `message_thread_id`), replies silently fail — they are neither delivered as DM nor as topic message.
... | closed | not_planned | true | 1 | [] | [] | 2026-04-30T19:37:35Z | 2026-05-03T04:35:13Z | 2026-04-30T19:45:43Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Felox63 | 249,820,014 | U_kgDODuPzbg | User | false |
openclaw/openclaw | 4,360,880,888 | I_kwDOQb6kR88AAAABA-3C-A | 75,243 | https://github.com/openclaw/openclaw/issues/75243 | https://api.github.com/repos/openclaw/openclaw/issues/75243 | Opened in wrong repository | Opened in the wrong repository by mistake. No actionable content here. | closed | completed | true | 0 | [] | [] | 2026-04-30T19:56:24Z | 2026-05-03T04:35:10Z | 2026-04-30T19:57:53Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Verlbot | 259,695,255 | U_kgDOD3qilw | User | false |
openclaw/openclaw | 4,360,845,136 | I_kwDOQb6kR88AAAABA-03UA | 75,240 | https://github.com/openclaw/openclaw/issues/75240 | https://api.github.com/repos/openclaw/openclaw/issues/75240 | [Bug]: OpenAI Codex attribution transport causes ChatGPT auth 401 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
OpenAI-Codex requests fail with `401 We got your request, but your ChatGPT login did not make it to this service.` after the OpenClaw attribution wrapper routes Codex traffic through the generic OpenAI Responses SDK transport.
... | closed | completed | true | 3 | [] | [] | 2026-04-30T19:50:17Z | 2026-05-03T04:35:09Z | 2026-04-30T20:06:11Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | bcdonadio | 1,458,878 | MDQ6VXNlcjE0NTg4Nzg= | User | false |
openclaw/openclaw | 4,361,232,549 | I_kwDOQb6kR88AAAABA_MgpQ | 75,263 | https://github.com/openclaw/openclaw/issues/75263 | https://api.github.com/repos/openclaw/openclaw/issues/75263 | [Bug] 2026.4.27 regression: Telegram provider logs "starting" but never begins long-poll; inbound silently dropped, outbound still works (regression of #45714 / #69064) | ## Summary
On openclaw 2026.4.27 (cbc2ba0), macOS 15 (Darwin 25.4.0), the Telegram provider logs `[default] starting provider (@ACP_openclaw_bot)` at startup and then never produces any further telegram-subsystem activity. No getUpdates, no inbound message events, no error. Outbound `sendMessage` still works for any se... | closed | not_planned | true | 3 | [] | [] | 2026-04-30T21:00:17Z | 2026-05-03T04:35:02Z | 2026-04-30T21:07:37Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | apato1978 | 196,211,503 | U_kgDOC7HzLw | User | false |
openclaw/openclaw | 4,359,942,778 | I_kwDOQb6kR88AAAABA99yeg | 75,174 | https://github.com/openclaw/openclaw/issues/75174 | https://api.github.com/repos/openclaw/openclaw/issues/75174 | [Bug]: diagnostics-otel spans don't share evt.trace.traceId — orphan root traces, broken Cloud Logging trace links | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The bundled `diagnostics-otel` plugin creates OTel spans via `tracer.startSpan(name, opts, undefined)`, falling back to `context.active()` for parent context. For non-HTTP-driven runs, and for any event dispatc... | closed | completed | true | 3 | [] | [] | 2026-04-30T17:11:21Z | 2026-05-03T04:34:59Z | 2026-04-30T20:39:07Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | tseller | 467,755 | MDQ6VXNlcjQ2Nzc1NQ== | User | false |
openclaw/openclaw | 4,361,351,294 | I_kwDOQb6kR88AAAABA_Twfg | 75,269 | https://github.com/openclaw/openclaw/issues/75269 | https://api.github.com/repos/openclaw/openclaw/issues/75269 | [Bug]: MiniMax fetch blocks Node.js event loop for ~135s → watchdog SIGTERM (regression in v4.27) | ## Environment
- OpenClaw version: v2026.4.27 (regression; v2026.4.26 unaffected)
- OS: Ubuntu 24.04 LTS, Linux kernel 6.17
- Node.js: default bundled
- Primary model: minimax/MiniMax-M2.7
- Deployment: systemd service, loopback-only, Telegram channel
## Symptom
After upgrading from v4.26 to v4.27, the gateway enter... | closed | not_planned | true | 1 | [] | [] | 2026-04-30T21:25:39Z | 2026-05-03T04:34:53Z | 2026-04-30T21:33:10Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | xiaobai-khhome | 262,472,254 | U_kgDOD6UCPg | User | false |
openclaw/openclaw | 4,361,131,042 | I_kwDOQb6kR88AAAABA_GUIg | 75,259 | https://github.com/openclaw/openclaw/issues/75259 | https://api.github.com/repos/openclaw/openclaw/issues/75259 | Beta blocker: Telegram - 2026.4.29-beta.4 runtime-deps reinstall loop and staged deps cannot resolve json5 | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
Yes
### Summary
While testing `2026.4.29-beta.3` / `2026.4.29-beta.4` to validate the recent gateway CPU/runtime-deps fixes, bundled plugin runtime dependency staging repeatedly reports success, but Telegram still exits because the mirrored b... | closed | completed | true | 4 | [
"beta-blocker"
] | [] | 2026-04-30T20:40:26Z | 2026-05-03T04:34:52Z | 2026-04-30T21:46:02Z | CONTRIBUTOR | null | 20260503T060030Z | 2026-05-03T06:00:30Z | brokemac79 | 255,583,030 | U_kgDODzvjNg | User | false |
openclaw/openclaw | 4,361,448,618 | I_kwDOQb6kR88AAAABA_Zsqg | 75,276 | https://github.com/openclaw/openclaw/issues/75276 | https://api.github.com/repos/openclaw/openclaw/issues/75276 | [Bug] 2026.4.27: Telegram-inbound-triggered agent replies silently dropped when output contains [[reply_to_current]] directive (groups affected) | ## Summary
On openclaw 2026.4.27, agent replies triggered by an inbound Telegram message in a group are silently dropped if the agent's output begins with the `[[reply_to_current]]` directive. The same agent, same chat, same session, when invoked via `openclaw agent --deliver`, delivers successfully because the CLI-tri... | closed | completed | true | 1 | [] | [] | 2026-04-30T21:46:25Z | 2026-05-03T04:34:51Z | 2026-04-30T21:51:31Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | apato1978 | 196,211,503 | U_kgDOC7HzLw | User | false |
openclaw/openclaw | 4,361,498,479 | I_kwDOQb6kR88AAAABA_cvbw | 75,278 | https://github.com/openclaw/openclaw/issues/75278 | https://api.github.com/repos/openclaw/openclaw/issues/75278 | Upgrade can leave old OpenClaw install in place and create split-path versions | Upgrading OpenClaw can leave the previous install on disk instead of replacing it. I have now seen this on two separate machines.
Observed pattern:
- old system/global install remained under `/usr/lib/node_modules/openclaw`
- new install was placed under a user/NVM path like `~/.nvm/.../lib/node_modules/openclaw`
- `/... | closed | not_planned | true | 1 | [] | [] | 2026-04-30T21:55:36Z | 2026-05-03T04:34:51Z | 2026-04-30T21:59:17Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Kru442 | 187,809,917 | U_kgDOCzHAfQ | User | false |
openclaw/openclaw | 4,361,444,379 | I_kwDOQb6kR88AAAABA_ZcGw | 75,275 | https://github.com/openclaw/openclaw/issues/75275 | https://api.github.com/repos/openclaw/openclaw/issues/75275 | 2026.4.29: Gateway repeatedly spawns runtime-deps pnpm installs and starves event loop | ### Bug type
Performance / hang / runtime-deps reinstall loop / gateway health regression
### Summary
After upgrading the same Ubuntu VPS that reproduced #75259 to the stable `2026.4.29` package, the original beta runtime-deps import failure appears fixed: the gateway reaches ready, Telegram starts, and there are no... | closed | not_planned | true | 7 | [] | [] | 2026-04-30T21:45:43Z | 2026-05-03T04:34:50Z | 2026-04-30T22:07:35Z | CONTRIBUTOR | null | 20260503T060030Z | 2026-05-03T06:00:30Z | brokemac79 | 255,583,030 | U_kgDODzvjNg | User | false |
openclaw/openclaw | 4,361,626,091 | I_kwDOQb6kR88AAAABA_kh6w | 75,286 | https://github.com/openclaw/openclaw/issues/75286 | https://api.github.com/repos/openclaw/openclaw/issues/75286 | [Bug]: bare /reset triggers OpenAI Responses API HTTP 400 (empty input/previous_response_id/prompt/conversation_id) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On openclaw v2026.4.26, sending a bare /reset to a Telegram-bridged agent on the openai-codex provider triggers an automatic LLM call before any user input arrives; the runtime constructs a Responses API reques... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-30T22:25:14Z | 2026-05-03T04:34:47Z | 2026-04-30T22:29:00Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | scottdominguez | 41,765,233 | MDQ6VXNlcjQxNzY1MjMz | User | false |
openclaw/openclaw | 4,361,671,291 | I_kwDOQb6kR88AAAABA_nSew | 75,291 | https://github.com/openclaw/openclaw/issues/75291 | https://api.github.com/repos/openclaw/openclaw/issues/75291 | 🐛 TUI freezes on startup — Node.js event loop blocking (100%+ CPU) | ## 🐛 TUI freezes on startup — Node.js event loop blocking
### Issue Description
The `openclaw tui` command freezes immediately upon launch, both over SSH and directly on the local machine (Ghostty terminal). The process consumes 100%+ CPU indefinitely, the terminal becomes unresponsive, and the only resolution is to... | closed | completed | true | 1 | [] | [] | 2026-04-30T22:36:51Z | 2026-05-03T04:34:46Z | 2026-04-30T22:40:29Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | iddar | 199,103 | MDQ6VXNlcjE5OTEwMw== | User | false |
openclaw/openclaw | 4,361,740,220 | I_kwDOQb6kR88AAAABA_rfvA | 75,296 | https://github.com/openclaw/openclaw/issues/75296 | https://api.github.com/repos/openclaw/openclaw/issues/75296 | [Bug] bundled-runtime-deps installer reports success after partial install on ETIMEDOUT, leaving channel modules unable to resolve json5 | ## Summary
After running `openclaw update` from 2026.4.26 → 2026.4.29, the bundled-runtime-deps installer reported a successful install in `~1s` but actually installed only **36 of 50 specs** into `~/.openclaw/plugin-runtime-deps/openclaw-2026.4.29-<hash>/node_modules/`. The discord and telegram channels then crash-lo... | closed | completed | true | 2 | [] | [] | 2026-04-30T22:52:07Z | 2026-05-03T04:33:41Z | 2026-04-30T22:56:22Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | mogglemoss | 10,966,195 | MDQ6VXNlcjEwOTY2MTk1 | User | false |
openclaw/openclaw | 4,360,358,670 | I_kwDOQb6kR88AAAABA-XLDg | 75,205 | https://github.com/openclaw/openclaw/issues/75205 | https://api.github.com/repos/openclaw/openclaw/issues/75205 | Track Codex app-server terminal notification hardening | ## Summary
OpenClaw hit a stuck Discord reply lane after Codex app-server accepted a `turn/start` request and then failed to deliver the terminal `turn/completed` or abort notification OpenClaw was waiting for. The active embedded-run handle stayed registered, so diagnostics correctly treated the session as an active ... | closed | completed | true | 5 | [
"maintainer"
] | [
"pash-openai"
] | 2026-04-30T18:22:51Z | 2026-05-04T04:35:52Z | 2026-05-01T17:55:38Z | CONTRIBUTOR | null | 20260504T060031Z | 2026-05-04T06:00:31Z | steipete | 58,493 | MDQ6VXNlcjU4NDkz | User | false |
openclaw/openclaw | 4,361,883,653 | I_kwDOQb6kR88AAAABA_0QBQ | 75,304 | https://github.com/openclaw/openclaw/issues/75304 | https://api.github.com/repos/openclaw/openclaw/issues/75304 | [Bug]: OpenClaw 2026.4.29 — Telegram/Discord channels crash-loop due to missing ESM exports in bundled runtime deps | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Environment
Node v22.22.0
OpenClaw 2026.4.29 (a448042)
Installed via npm globally: /opt/homebrew/lib/node_modules/openclaw
Symptoms
Telegram and Discord channels crash immediately on every gateway start with:
Cannot find ... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-30T23:26:15Z | 2026-05-03T04:34:39Z | 2026-04-30T23:30:30Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Simonwintle | 262,380,096 | U_kgDOD6OaQA | User | false |
openclaw/openclaw | 4,361,834,498 | I_kwDOQb6kR88AAAABA_xQAg | 75,300 | https://github.com/openclaw/openclaw/issues/75300 | https://api.github.com/repos/openclaw/openclaw/issues/75300 | [Bug]: Anthropic prompt cache busted on every turn: per-message metadata injected inside system block with cache_control | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
## Summary
Every message causes a full ~9,500-token cache write on the system block, making
the system prompt cache ineffective despite `cache_control: {"type":"ephemeral","ttl":"1h"}`.
Cost: ~$0.035/message (C... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-30T23:13:13Z | 2026-05-03T04:34:38Z | 2026-04-30T23:18:45Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | obiter | 41,659,171 | MDQ6VXNlcjQxNjU5MTcx | User | false |
openclaw/openclaw | 4,361,926,132 | I_kwDOQb6kR88AAAABA_219A | 75,307 | https://github.com/openclaw/openclaw/issues/75307 | https://api.github.com/repos/openclaw/openclaw/issues/75307 | Background tasks stay 'running' forever after gateway crash/restart, blocking channel reload + saturating CPU | ## Summary
When the gateway process crashes or is restarted while background task runs (cron, subagent, cli) are in-flight, the task records remain in 'running' status indefinitely. They block channel reload, accumulate across restarts, and eventually saturate the event loop.
## Repro
1. Trigger an isolated agentTur... | closed | completed | true | 1 | [] | [] | 2026-04-30T23:39:06Z | 2026-05-03T04:34:38Z | 2026-04-30T23:43:32Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | xonaman | 45,240,351 | MDQ6VXNlcjQ1MjQwMzUx | User | false |
openclaw/openclaw | 4,362,122,533 | I_kwDOQb6kR88AAAABBAC1JQ | 75,318 | https://github.com/openclaw/openclaw/issues/75318 | https://api.github.com/repos/openclaw/openclaw/issues/75318 | [Bug]: 4.29 openclaw status (fast) shows empty Channels table — channel visible only in --deep | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
`openclaw status` (fast) shows an empty Channels table. The same channel appears correctly in `openclaw status --deep`. Channel functionality (Telegram send/receive) works normally.
### Steps to reproduce
1. Configure Tel... | closed | not_planned | true | 1 | [
"bug",
"regression"
] | [] | 2026-05-01T00:32:52Z | 2026-05-03T04:34:33Z | 2026-05-01T00:37:24Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | fenglanhua | 257,674,093 | U_kgDOD1vLbQ | User | false |
openclaw/openclaw | 4,184,922,487 | I_kwDOQb6kR875cNl3 | 58,838 | https://github.com/openclaw/openclaw/issues/58838 | https://api.github.com/repos/openclaw/openclaw/issues/58838 | fix(compaction): retry loop can burn tokens when summarizer model is unavailable | ## Bug Description
`summarizeChunks()` in `src/agents/compaction.ts` uses `retryAsync()` with 3 attempts per chunk to handle transient errors during compaction. However, when the summarizer model is persistently unavailable (overloaded, rate-limited, or misconfigured), the retry loop burns tokens across multiple chunk... | open | null | false | 1 | [] | [] | 2026-04-01T07:51:29Z | 2026-05-01T00:44:50Z | null | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | aaron-he-zhu | 139,607,425 | U_kgDOCFI9gQ | User | false |
openclaw/openclaw | 4,186,789,127 | I_kwDOQb6kR875jVUH | 59,022 | https://github.com/openclaw/openclaw/issues/59022 | https://api.github.com/repos/openclaw/openclaw/issues/59022 | [Gateway] A2UI host not reachable | ## 问题描述
Gateway 的 A2UI 服务不可用，导致 `canvas.a2ui.push` 命令失败。
## 环境
- Gateway: OpenClaw 2026.3.31
- macOS 节点: 曾蔚的MacBook Pro, macOS 26.3.0
## 复现步骤
1. 执行 `canvas a2ui_push --node "曾蔚的MacBook Pro" --jsonl '{"beginRendering": {"surfaceId": "main"}}'`
2. 返回错误: `UNAVAILABLE: A2UI_HOST_UNAVAILABLE: A2UI host not reachable`
... | open | null | false | 1 | [] | [] | 2026-04-01T13:05:56Z | 2026-05-01T00:44:57Z | null | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | nightq | 3,429,433 | MDQ6VXNlcjM0Mjk0MzM= | User | false |
openclaw/openclaw | 4,201,503,843 | I_kwDOQb6kR876bdxj | 60,450 | https://github.com/openclaw/openclaw/issues/60450 | https://api.github.com/repos/openclaw/openclaw/issues/60450 | v2026.3.28: Runaway API calls burned entire monthly budget across all providers without user action | ## Summary
OpenClaw release **2026.3.28** (March 31, 2026) caused uncontrolled runaway API calls across **all configured LLM providers simultaneously** on March 31 and April 1, 2026. No user interaction triggered this — it happened autonomously. The entire monthly budget across Anthropic, OpenAI, Google (Gemini), and ... | open | null | false | 1 | [] | [] | 2026-04-03T17:34:19Z | 2026-05-02T07:49:42Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | jarvisaibowen | 273,410,257 | U_kgDOEEvo0Q | User | false |
openclaw/openclaw | 4,203,795,758 | I_kwDOQb6kR876kNUu | 60,713 | https://github.com/openclaw/openclaw/issues/60713 | https://api.github.com/repos/openclaw/openclaw/issues/60713 | Windows: resolvePreferredOpenClawTmpDir uses C:\tmp instead of proper temp directory | ## Description
On Windows, `resolvePreferredOpenClawTmpDir()` resolves to `C:\tmp\openclaw` instead of the proper Windows temp directory (`C:\Users\<user>\AppData\Local\Temp\openclaw`).
The gateway log line confirms this:
```
[gateway] log file: \tmp\openclaw\openclaw-2026-04-04.log
```
## Root Cause
In `src/infra/... | closed | completed | false | 1 | [] | [] | 2026-04-04T06:14:29Z | 2026-05-05T01:32:38Z | 2026-05-05T01:32:38Z | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | KurtVogel | 22,785,743 | MDQ6VXNlcjIyNzg1NzQz | User | false |
openclaw/openclaw | 4,204,612,602 | I_kwDOQb6kR876nUv6 | 60,827 | https://github.com/openclaw/openclaw/issues/60827 | https://api.github.com/repos/openclaw/openclaw/issues/60827 | Add default container resource limits for sandbox | ## Description
Sandbox containers created by OpenClaw have no CPU or memory limits by default. On a resource-constrained host (e.g., 8GB Mac mini running 6+ sandboxed agents), a single runaway agent can consume all host RAM and crash the gateway, other agents, and host services.
## Current Behavior
New sandbox conta... | open | null | false | 1 | [] | [] | 2026-04-04T11:09:00Z | 2026-05-02T09:00:05Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | Belthezar46 | 85,139,005 | MDQ6VXNlcjg1MTM5MDA1 | User | false |
openclaw/openclaw | 4,213,199,482 | I_kwDOQb6kR877IFJ6 | 62,055 | https://github.com/openclaw/openclaw/issues/62055 | https://api.github.com/repos/openclaw/openclaw/issues/62055 | Windows: CLI crashes with stack overflow / heap OOM on v2026.4.5 (large ESM module graph exceeds V8 default stack) | ## Bug: Windows crashes on CLI startup (stack overflow + heap OOM)
### Summary
On Windows, `openclaw dashboard` (and likely other heavy commands) crashes with either:
1. **`RangeError: Maximum call stack size exceeded`** during ESM module evaluation, or
2. **`FATAL ERROR: Reached heap limit Allocation failed ... | open | null | false | 3 | [] | [] | 2026-04-06T17:58:50Z | 2026-05-02T09:22:24Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | LouisGameDev | 1,916,460 | MDQ6VXNlcjE5MTY0NjA= | User | false |
openclaw/openclaw | 4,216,740,824 | I_kwDOQb6kR877VlvY | 62,413 | https://github.com/openclaw/openclaw/issues/62413 | https://api.github.com/repos/openclaw/openclaw/issues/62413 | [Bug]: Dream Diary tab markdown rendering broken | ## Bug Description
The Dream Diary tab in the Control UI has broken markdown rendering. Markdown formatting is displayed literally instead of being rendered properly.
### Screenshot
<img width="1600" height="902" alt="Image" src="https://github.com/user-attachments/assets/9332d102-e3b7-42ee-b8b4-168df30846ce" />
##... | closed | completed | false | 3 | [] | [] | 2026-04-07T09:27:49Z | 2026-05-04T03:26:44Z | 2026-05-04T03:26:44Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | AnthonyFrancis | 3,010,099 | MDQ6VXNlcjMwMTAwOTk= | User | false |
openclaw/openclaw | 4,217,423,565 | I_kwDOQb6kR877YMbN | 62,455 | https://github.com/openclaw/openclaw/issues/62455 | https://api.github.com/repos/openclaw/openclaw/issues/62455 | [Bug]: Discord channel config rejects channel-level agentId and can abort gateway reload | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On OpenClaw `2026.4.5`, if `agentId` is added under a Discord channel entry at `channels.discord.guilds.<guildId>.channels.<channelId>`, config reload becomes invalid and the gateway can abort with `Unrecognize... | closed | completed | true | 2 | [] | [] | 2026-04-07T11:45:49Z | 2026-05-04T04:35:20Z | 2026-05-01T21:43:34Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | foundationfivepro-gif | 271,533,003 | U_kgDOEC9Dyw | User | false |
openclaw/openclaw | 4,217,557,085 | I_kwDOQb6kR877YtBd | 62,468 | https://github.com/openclaw/openclaw/issues/62468 | https://api.github.com/repos/openclaw/openclaw/issues/62468 | [Bug] Feishu message delivery fails with 'Unknown target jaydenli' | ## Bug Description
When the Feishu channel plugin tries to send a message to target 'jaydenli' (a username), it fails with 'Unknown target jaydenli'. The Feishu plugin requires open_id or chat_id format, not display names.
## Log Evidence
```
[tools] message failed: Unknown target "jaydenli" for Feishu. Hint: <chatId|... | open | null | false | 1 | [] | [] | 2026-04-07T12:12:18Z | 2026-05-02T09:00:11Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | jaydenlichong29-create | 268,866,858 | U_kgDOEAaVKg | User | false |
openclaw/openclaw | 4,217,690,055 | I_kwDOQb6kR877ZNfH | 62,495 | https://github.com/openclaw/openclaw/issues/62495 | https://api.github.com/repos/openclaw/openclaw/issues/62495 | Avoid noisy skill-path escape warnings for sibling skill references | OpenClaw currently logs `Skipping skill path that resolves outside its configured root.` for a normal sibling reference in `workspace/skills/openclaw-local/SKILL.md`:
- `openclaw-local` links to `../openclaw-optimisation/SKILL.md`
- the loader treats that Markdown link as an escaped skill path
- this produces repeated... | open | null | false | 1 | [] | [] | 2026-04-07T12:36:47Z | 2026-05-02T09:22:25Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | apoapostolov | 4,083,812 | MDQ6VXNlcjQwODM4MTI= | User | false |
openclaw/openclaw | 4,219,271,267 | I_kwDOQb6kR877fPhj | 62,604 | https://github.com/openclaw/openclaw/issues/62604 | https://api.github.com/repos/openclaw/openclaw/issues/62604 | [Bug]: Control UI image paste flow can inject massive base64 text into the composer; send/draft handling is hard to patch safely in installed dist bundles | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
## Summary
I hit a Control UI issue where pasting an image could inject a very large `data:image/...;base64,...` string directly into the chat composer as text instead of being handled cleanly as an attachment... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-07T16:53:25Z | 2026-05-02T07:49:58Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | cpwilhelmi | 64,565,210 | MDQ6VXNlcjY0NTY1MjEw | User | false |
openclaw/openclaw | 4,220,480,963 | I_kwDOQb6kR877j23D | 62,697 | https://github.com/openclaw/openclaw/issues/62697 | https://api.github.com/repos/openclaw/openclaw/issues/62697 | feat(whatsapp): Support WhatsApp Channel (Newsletter) messages | ## Feature Request
### Description
Add support for receiving and processing messages from **WhatsApp Channels (Newsletters)**.
### Current Behavior
- Newsletter messages arrive via Baileys `messages.upsert` with JIDs ending in `@newsletter`
- They are treated as DMs and blocked by `dmPolicy: allowlist`
### Use Case
... | closed | completed | true | 1 | [] | [] | 2026-04-07T20:41:07Z | 2026-05-05T04:26:25Z | 2026-05-02T20:09:33Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | nadav2cohen | 58,531,605 | MDQ6VXNlcjU4NTMxNjA1 | User | false |
openclaw/openclaw | 4,362,147,629 | I_kwDOQb6kR88AAAABBAEXLQ | 75,320 | https://github.com/openclaw/openclaw/issues/75320 | https://api.github.com/repos/openclaw/openclaw/issues/75320 | [Bug]: 4.29 runtime-deps missing grammy, @grammyjs/runner, @grammyjs/transformer-throttler — Telegram channel fails to load | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to 4.29, the Telegram channel fails to load due to missing grammy runtime dependencies. Manually installing into the runtime-deps directory restores Telegram functionality.
Error
[channels] failed to load bun... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-05-01T00:41:16Z | 2026-05-03T04:34:33Z | 2026-05-01T00:47:21Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | fenglanhua | 257,674,093 | U_kgDOD1vLbQ | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.