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 | 4,196,514,350 | I_kwDOQb6kR876Ibou | 59,874 | https://github.com/openclaw/openclaw/issues/59874 | https://api.github.com/repos/openclaw/openclaw/issues/59874 | Add Slack streaming API and Block Kit support for agent progress updates | ## Summary
Enable agents to show live progress updates using Slack's native `task_card` blocks and streaming API, similar to Attio's agent UX.
## Current State
- OpenClaw's slack tool supports `sendMessage` and `editMessage` but only with plain text `content`
- No support for Block Kit `blocks` parameter
- No suppor... | closed | completed | true | 2 | [] | [] | 2026-04-02T19:20:09Z | 2026-04-28T05:12:46Z | 2026-04-25T13:18:34Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jakewayward | 153,329,823 | U_kgDOCSOgnw | User | false |
openclaw/openclaw | 4,196,530,242 | I_kwDOQb6kR876IfhC | 59,877 | https://github.com/openclaw/openclaw/issues/59877 | https://api.github.com/repos/openclaw/openclaw/issues/59877 | /status task line displays completed cron job output, flooding the status card | ## Summary
`/status` in chat surfaces the full `terminalSummary` of completed background tasks (cron jobs), flooding the status card with content that was already delivered to the user independently.
## Version
OpenClaw 2026.3.31 (213a704)
## Steps to Reproduce
1. Create an `agentTurn` cron job with `sessionTarget... | closed | completed | true | 3 | [] | [] | 2026-04-02T19:24:06Z | 2026-04-28T05:12:45Z | 2026-04-25T13:18:41Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | smonett | 50,761,481 | MDQ6VXNlcjUwNzYxNDgx | User | false |
openclaw/openclaw | 4,196,609,493 | I_kwDOQb6kR876Iy3V | 59,882 | https://github.com/openclaw/openclaw/issues/59882 | https://api.github.com/repos/openclaw/openclaw/issues/59882 | [Bug]: tools.exec.ask: "always" fails with "pairing required" on local loopback gateway | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
**Title:** `tools.exec.ask: "always"` fails with "pairing required" on local loopback gateway
**Summary:**
When `tools.exec.ask` is set to `"always"` (with `security: "full"`), all exec commands fail with:
```
Exec approv... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-02T19:42:58Z | 2026-04-28T05:12:45Z | 2026-04-25T13:18:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kylekrug-xalt | 273,196,650 | U_kgDOEEimag | User | false |
openclaw/openclaw | 4,196,654,928 | I_kwDOQb6kR876I99Q | 59,885 | https://github.com/openclaw/openclaw/issues/59885 | https://api.github.com/repos/openclaw/openclaw/issues/59885 | Discord components v2 buttons: INTERACTION_CREATE not matched by wildcard handler | ## Bug Report
### Summary
Discord component v2 buttons sent via `openclaw message send --components` render correctly and register `"components": true` in the response, but clicking them results in Discord showing "This interaction failed" — meaning no ACK is sent within 3 seconds.
### Environment
- OpenClaw version:... | closed | completed | true | 1 | [] | [] | 2026-04-02T19:53:16Z | 2026-04-28T05:12:44Z | 2026-04-25T13:19:01Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | bz5555 | 265,107,019 | U_kgDOD802Sw | User | false |
openclaw/openclaw | 4,196,698,308 | I_kwDOQb6kR876JIjE | 59,887 | https://github.com/openclaw/openclaw/issues/59887 | https://api.github.com/repos/openclaw/openclaw/issues/59887 | sessions.create silently drops messages containing :// (2026.4.1+) | ## Bug
Any `sessions.create` message containing `://` (URLs, connection strings) causes the agent dispatch to silently fail. The session is created and `status: "started"` is returned, but the agent never runs — no hooks fire, no LLM call is made, the session transcript remains empty.
Verified on **2026.4.1** and **2... | closed | completed | true | 3 | [] | [] | 2026-04-02T20:01:18Z | 2026-04-28T05:12:43Z | 2026-04-25T13:19:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | wkeything | 263,864,978 | U_kgDOD7pCkg | User | false |
openclaw/openclaw | 4,196,727,305 | I_kwDOQb6kR876JPoJ | 59,889 | https://github.com/openclaw/openclaw/issues/59889 | https://api.github.com/repos/openclaw/openclaw/issues/59889 | devices approve deadlock after gateway restart — node-host repair request cannot be approved via CLI | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
After a gateway restart following the 2026.4.1 update, the local node-host service requests a role-upgrade repair pairing (roleFrom=operator roleTo=node). This creates an unresolvable deadlock:
openclaw de... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T20:07:36Z | 2026-04-28T05:12:42Z | 2026-04-25T13:19:23Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kobuck | 122,318,478 | U_kgDOB0pujg | User | false |
openclaw/openclaw | 4,196,730,632 | I_kwDOQb6kR876JQcI | 59,890 | https://github.com/openclaw/openclaw/issues/59890 | https://api.github.com/repos/openclaw/openclaw/issues/59890 | [Bug]: Webchat crashes when user sends localhost URL in message | **Bug Description**:
When a user sends a localhost URL (e.g., http://localhost:8080/?state=...) in the webchat input, OpenClaw webchat attempts to parse/process the URL and throws a grammy module error, causing the message to fail.
**Steps to Reproduce**:
1. Have a running OpenClaw gateway with webchat enabled
2. User... | closed | completed | true | 2 | [] | [] | 2026-04-02T20:08:24Z | 2026-04-28T05:12:42Z | 2026-04-25T13:19:30Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Barry-886 | 269,722,164 | U_kgDOEBOiNA | User | false |
openclaw/openclaw | 4,196,761,277 | I_kwDOQb6kR876JX69 | 59,894 | https://github.com/openclaw/openclaw/issues/59894 | https://api.github.com/repos/openclaw/openclaw/issues/59894 | [Bug]: doctor falsely reports no active memory plugin while mem0 is loaded and working | ## Summary
On OpenClaw `2026.3.31`, `2026.4.1`, and `2026.4.2`, `openclaw doctor --fix` reports:
> No active memory plugin is registered for the current config.
…but `openclaw-mem0` is loaded and working normally.
In this setup:
- `plugins.allow` includes `openclaw-mem0`
- `plugins.slots.memory` is set to `openclaw... | closed | completed | true | 1 | [] | [] | 2026-04-02T20:14:32Z | 2026-04-28T05:12:41Z | 2026-04-25T13:19:37Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | QuarkAssistant | 262,458,533 | U_kgDOD6TMpQ | User | false |
openclaw/openclaw | 4,196,786,811 | I_kwDOQb6kR876JeJ7 | 59,895 | https://github.com/openclaw/openclaw/issues/59895 | https://api.github.com/repos/openclaw/openclaw/issues/59895 | Discord channel history tool for agent session boot | ## Summary
When a new agent session boots in a Discord channel, it has no knowledge of prior conversation. The Discord bot has API access to message history, but no tool exists for agents to read it.
## The Ask
A tool (e.g. `discord_history`) that lets an agent read the last N messages from their current channel on se... | closed | completed | true | 1 | [] | [] | 2026-04-02T20:20:17Z | 2026-04-28T05:12:40Z | 2026-04-25T13:19:44Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | UBIpromoter | 46,543,177 | MDQ6VXNlcjQ2NTQzMTc3 | User | false |
openclaw/openclaw | 4,196,787,143 | I_kwDOQb6kR876JePH | 59,896 | https://github.com/openclaw/openclaw/issues/59896 | https://api.github.com/repos/openclaw/openclaw/issues/59896 | Auto state-save hook on session lifecycle events | ## Summary
Session state saving currently depends on agent discipline — manually writing handoff files when told to or when remembering. This fails predictably per principle #4 (infrastructure is law, not etiquette).
## The Ask
A mechanical state-save trigger in OpenClaw that fires on session lifecycle events:
- After... | closed | not_planned | true | 1 | [] | [] | 2026-04-02T20:20:22Z | 2026-04-28T05:12:39Z | 2026-04-25T13:19:51Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | UBIpromoter | 46,543,177 | MDQ6VXNlcjQ2NTQzMTc3 | User | false |
openclaw/openclaw | 4,197,187,463 | I_kwDOQb6kR876K_-H | 59,927 | https://github.com/openclaw/openclaw/issues/59927 | https://api.github.com/repos/openclaw/openclaw/issues/59927 | Gateway crash: race condition causes maxAttempts=0 exception on disconnect | ## Problem
The Discord gateway crashes with an uncaught exception when a disconnect event fires before the reconnect config is fully initialized.
## Error
```
Gateway disconnected: Max reconnect attempts (0) reached
C [ERROR] Gateway disconnect error: Error: Max reconnect attempts (0) reached
```
## Root Cause
Race c... | closed | completed | true | 1 | [] | [] | 2026-04-02T21:58:34Z | 2026-04-28T05:12:39Z | 2026-04-25T13:20:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | sekokaral | 137,436,350 | U_kgDOCDEcvg | User | false |
openclaw/openclaw | 4,197,191,802 | I_kwDOQb6kR876LBB6 | 59,928 | https://github.com/openclaw/openclaw/issues/59928 | https://api.github.com/repos/openclaw/openclaw/issues/59928 | [Bug]: Primary model not respected after fallback in OpenClaw 2026.4.2 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
**Version**
OpenClaw 2026.4.2
**Summary**
A session does not reliably return to or stay on the configured primary model after fallback occurs. Even with `openai-codex/gpt-5.4` configured as primary, the a... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T21:59:54Z | 2026-04-28T05:12:38Z | 2026-04-25T13:20:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lanke-b2 | 261,280,188 | U_kgDOD5LRvA | User | false |
openclaw/openclaw | 4,197,314,096 | I_kwDOQb6kR876Le4w | 59,934 | https://github.com/openclaw/openclaw/issues/59934 | https://api.github.com/repos/openclaw/openclaw/issues/59934 | [Bug]: config.patch fails for Telegram group IDs (negative numeric keys in path) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
`gateway config.patch` returns "invalid config" errors when attempting to set per-group configuration using negative Telegram group IDs in the path. For example:
```
channels.telegram.groups.-5028303500.requireMention = false
```
The nega... | closed | completed | true | 1 | [] | [] | 2026-04-02T22:38:18Z | 2026-04-28T05:12:37Z | 2026-04-25T13:20:37Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fattchris | 2,224,082 | MDQ6VXNlcjIyMjQwODI= | User | false |
openclaw/openclaw | 4,197,332,043 | I_kwDOQb6kR876LjRL | 59,936 | https://github.com/openclaw/openclaw/issues/59936 | https://api.github.com/repos/openclaw/openclaw/issues/59936 | [Bug]: Default CLI template (openclaw.json) is missing required reasoning/minimal flags for 2026.4.2-beta.1 executor | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
Yes
### Summary
The default openclaw.json generated by the CLI lacks reasoning_effort and format flags required by the 2026.4.2-beta.1 agent executor, resulting in a silent failure on Telegram with the fallback error message.
### Steps to... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-02T22:43:52Z | 2026-04-28T05:12:36Z | 2026-04-25T13:20:44Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ashish797 | 31,034,163 | MDQ6VXNlcjMxMDM0MTYz | User | false |
openclaw/openclaw | 4,197,356,188 | I_kwDOQb6kR876LpKc | 59,940 | https://github.com/openclaw/openclaw/issues/59940 | https://api.github.com/repos/openclaw/openclaw/issues/59940 | hooks/agent with deliver: false accepts request but never processes agent session | ## Summary
When sending a POST to `/hooks/agent` with `deliver: false`, the gateway returns HTTP 200 with a valid `runId`, but the agent session is never executed. The session remains in `empty` status with 0 messages and 0 tool calls.
## Steps to reproduce
1. Send a hooks request with `deliver: false`:
```bash
curl... | closed | not_planned | true | 1 | [] | [] | 2026-04-02T22:53:12Z | 2026-04-28T05:12:35Z | 2026-04-25T13:20:51Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | leonardo-bolanos | 196,844,214 | U_kgDOC7uatg | User | false |
openclaw/openclaw | 4,197,392,621 | I_kwDOQb6kR876LyDt | 59,946 | https://github.com/openclaw/openclaw/issues/59946 | https://api.github.com/repos/openclaw/openclaw/issues/59946 | [Bug]: Subagent killed by LLM provider streaming idle timeout (~60s) — no retry or graceful handling | ## Problem
When a subagent is processing a task that requires extended thinking (e.g., reading large codebases, complex analysis), the LLM provider's streaming connection may have an idle timeout (e.g., ~60s for GitHub Copilot API). If the model is thinking and produces no tokens within this window, the connection is ... | closed | completed | true | 2 | [] | [] | 2026-04-02T23:08:02Z | 2026-04-28T05:12:35Z | 2026-04-25T13:20:58Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kagura-agent | 268,167,063 | U_kgDOD_vnlw | User | false |
openclaw/openclaw | 4,197,469,128 | I_kwDOQb6kR876MEvI | 59,951 | https://github.com/openclaw/openclaw/issues/59951 | https://api.github.com/repos/openclaw/openclaw/issues/59951 | [Bug]: Duplicate Inbound Message Delivery | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
A single Telegram message is injected into the agent session 2–3 times as separate inbound events ~60s apart, causing the agent to send duplicate replies and re-run expensive operations.
### Steps to reproduce... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T23:40:25Z | 2026-04-28T05:12:34Z | 2026-04-25T13:21:06Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | iboxgithub | 5,391,993 | MDQ6VXNlcjUzOTE5OTM= | User | false |
openclaw/openclaw | 4,197,474,744 | I_kwDOQb6kR876MGG4 | 59,952 | https://github.com/openclaw/openclaw/issues/59952 | https://api.github.com/repos/openclaw/openclaw/issues/59952 | Companion app: Anthropic provider usage/limits not visible while OpenAI Codex appears in Usage providers | ## Summary
In the macOS companion app / Control UI usage provider view, `openai-codex` appears as a usage provider, but Anthropic/Claude does not appear comparably even when Anthropic is the configured default provider and is actively being used.
## Why this seems inconsistent
OpenClaw docs suggest Anthropic has provi... | closed | completed | true | 1 | [] | [] | 2026-04-02T23:42:52Z | 2026-04-28T05:12:33Z | 2026-04-25T13:21:13Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | aounakram | 57,077,179 | MDQ6VXNlcjU3MDc3MTc5 | User | false |
openclaw/openclaw | 4,197,608,179 | I_kwDOQb6kR876Mmrz | 59,967 | https://github.com/openclaw/openclaw/issues/59967 | https://api.github.com/repos/openclaw/openclaw/issues/59967 | QMD memory backend: managed collection naming mismatch causes Collection not found: memory-dir-main | # OpenClaw issue draft — QMD managed collection naming mismatch
## Suggested title
QMD memory backend: managed collection naming mismatch causes `Collection not found: memory-dir-main`
## Repository
- Target: `openclaw/openclaw`
## Summary
When OpenClaw is configured with `memory.backend = "qmd"`, `memory_search`... | closed | completed | true | 2 | [] | [] | 2026-04-03T00:33:36Z | 2026-04-28T05:12:32Z | 2026-04-25T13:21:20Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | wastemobile | 451,223 | MDQ6VXNlcjQ1MTIyMw== | User | false |
openclaw/openclaw | 4,197,652,579 | I_kwDOQb6kR876Mxhj | 59,969 | https://github.com/openclaw/openclaw/issues/59969 | https://api.github.com/repos/openclaw/openclaw/issues/59969 | [Bug]: [Exec] Exec approval registration fails with "pairing required" error despite having paired devices | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
## Bug Description
When executing commands that require approval through the exec tool, I consistently get a "pairing required" error, even though:
- Gateway is running normally
- CLI can connect to Gateway (s... | closed | not_planned | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-03T00:51:13Z | 2026-04-28T05:12:32Z | 2026-04-25T13:21:27Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | bruce23079 | 190,098,672 | U_kgDOC1Ss8A | User | false |
openclaw/openclaw | 4,197,663,517 | I_kwDOQb6kR876M0Md | 59,972 | https://github.com/openclaw/openclaw/issues/59972 | https://api.github.com/repos/openclaw/openclaw/issues/59972 | Webchat streaming replies shift reading position and are hard to read while still rendering | ### Summary
Preserve reading position while long streamed assistant replies continue rendering in webchat/control-ui.
### Problem to solve
In webchat / control-ui, long streamed assistant replies are hard to read because the active message keeps growing while I’m trying to read it.
If I pause to read a response bef... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-03T00:55:13Z | 2026-04-28T05:12:31Z | 2026-04-25T13:21:34Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | staysharp-rgb | 271,112,009 | U_kgDOECjXSQ | User | false |
openclaw/openclaw | 4,197,713,021 | I_kwDOQb6kR876NAR9 | 59,977 | https://github.com/openclaw/openclaw/issues/59977 | https://api.github.com/repos/openclaw/openclaw/issues/59977 | Heartbeat scheduler resets on gateway restart instead of maintaining wall-clock cadence | ## Description
The heartbeat scheduler (configured with `every: 2h`, `activeHours: 07:00–23:00`) does not maintain a durable wall-clock-anchored cadence. Instead, the 2h interval timer appears to reset on every gateway restart/reload, causing:
- Burst clusters of heartbeat fires immediately after restarts
- Long gaps... | closed | completed | true | 2 | [] | [] | 2026-04-03T01:11:47Z | 2026-04-28T05:12:30Z | 2026-04-25T13:21:41Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | bfgpollara | 239,261,520 | U_kgDODkLXUA | User | false |
openclaw/openclaw | 4,197,743,199 | I_kwDOQb6kR876NHpf | 59,983 | https://github.com/openclaw/openclaw/issues/59983 | https://api.github.com/repos/openclaw/openclaw/issues/59983 | Bug: Stale session lock after timeout leaves channel permanently dead | ## Description
After an agent run times out on a group session, a stale `.jsonl.lock` file is left behind with no corresponding `.jsonl` transcript. This causes the gateway to silently drop all subsequent inbound messages for that session — the channel appears permanently dead until manual intervention.
## Steps to R... | closed | completed | true | 2 | [] | [] | 2026-04-03T01:24:59Z | 2026-04-28T05:12:29Z | 2026-04-25T13:21:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | garrett22ge | 258,098,744 | U_kgDOD2JGOA | User | false |
openclaw/openclaw | 4,197,748,205 | I_kwDOQb6kR876NI3t | 59,984 | https://github.com/openclaw/openclaw/issues/59984 | https://api.github.com/repos/openclaw/openclaw/issues/59984 | UI Bug: exec approval popup should be scrollable when content is long | ## Bug Description
When multiple exec approval requests pile up or a command has long content, the approval popup in the Web UI renders without a scrollbar, causing content to overflow beyond the viewport and making it unusable.
## Steps to Reproduce
1. Have multiple pending exec approvals
2. Or have a command with v... | closed | completed | true | 1 | [] | [] | 2026-04-03T01:27:23Z | 2026-04-28T05:12:29Z | 2026-04-25T13:21:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | littlebest | 2,005,774 | MDQ6VXNlcjIwMDU3NzQ= | User | false |
openclaw/openclaw | 4,197,753,353 | I_kwDOQb6kR876NKIJ | 59,987 | https://github.com/openclaw/openclaw/issues/59987 | https://api.github.com/repos/openclaw/openclaw/issues/59987 | If the command in the authorization window is too long and cannot be scrolled, it may lead to the inability to operate and determine whether to authorize.[Bug]: | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
If the command in the authorization window is too long and cannot be scrolled, it may lead to the inability to operate and determine whether to authorize.[Bug]:
### Steps to reproduce
Attempt to trigger an overly long com... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-03T01:29:33Z | 2026-04-28T05:12:28Z | 2026-04-25T13:22:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | DeadPool0o | 79,043,938 | MDQ6VXNlcjc5MDQzOTM4 | User | false |
openclaw/openclaw | 4,197,755,242 | I_kwDOQb6kR876NKlq | 59,988 | https://github.com/openclaw/openclaw/issues/59988 | https://api.github.com/repos/openclaw/openclaw/issues/59988 | [Feature]: 支持取消排队中的消息 | ### Summary
用户发送消息后，如果消息还在排队中等待处理，无法主动取消。导致用户发错内容后只能等错误指令执行完，或者继续排在后面越积越多。
### Problem to solve
用户能够主动取消/删除还在排队中的消息，而不是让它继续执行或等待。
### Proposed solution
在 Control UI 增加"清空队列"按钮
- 或发送特定命令（如 /cancel）来清除排队
- 或在聊天界面直接划掉/删除待处理消息
### Alternatives considered
_No response_
### Impact
- 用户打字手滑发错了
- 用户想改指令但已经进入队列
- 用户不确... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-03T01:30:24Z | 2026-04-28T05:12:27Z | 2026-04-25T13:22:10Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | mihaikeji | 143,304,019 | U_kgDOCIqlUw | User | false |
openclaw/openclaw | 4,197,774,718 | I_kwDOQb6kR876NPV- | 59,994 | https://github.com/openclaw/openclaw/issues/59994 | https://api.github.com/repos/openclaw/openclaw/issues/59994 | LCM: FTS index grows unbounded — compaction deletes messages but never cleans FTS entries | ## Bug Description
LCM's message compaction removes rows from the `messages` table but never deletes the corresponding entries from the `messages_fts` FTS5 virtual table. Over time this causes the FTS index to grow far larger than the actual message count.
## Impact
- **Database bloat**: Our LCM database grew to 618... | closed | not_planned | true | 1 | [] | [] | 2026-04-03T01:39:07Z | 2026-04-28T05:12:26Z | 2026-04-25T13:22:18Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | bradfreels | 179,389,909 | U_kgDOCrFF1Q | User | false |
openclaw/openclaw | 4,197,801,410 | I_kwDOQb6kR876NV3C | 59,996 | https://github.com/openclaw/openclaw/issues/59996 | https://api.github.com/repos/openclaw/openclaw/issues/59996 | [Bug]: Heartbeat model config is ignored unless heartbeat session has persisted override | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Heartbeat does not honor the configured agents.defaults.heartbeat.model on its own, because when the heartbeat session’s persisted providerOverride and modelOverride are removed, the next
heartbeat run resolv... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-03T01:49:39Z | 2026-04-28T05:12:26Z | 2026-04-25T13:22:25Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | mintotsai | 1,557,318 | MDQ6VXNlcjE1NTczMTg= | User | false |
openclaw/openclaw | 4,197,826,644 | I_kwDOQb6kR876NcBU | 59,999 | https://github.com/openclaw/openclaw/issues/59999 | https://api.github.com/repos/openclaw/openclaw/issues/59999 | Bug: cron.maxConcurrentRuns setting has no effect — cron jobs still spawn N parallel runs per trigger | ## Summary
Setting `cron.maxConcurrentRuns: 1` in `openclaw.json` has no observable effect. Cron jobs continue to spawn multiple parallel runs per trigger (consistently 5 in our setup).
## Environment
- OpenClaw version: 2026.3.24 (cff6dc9)
- OS: macOS 25.3.0 (arm64)
- Gateway mode: LaunchAgent (local)
- Multiple ga... | closed | not_planned | true | 2 | [] | [] | 2026-04-03T01:59:55Z | 2026-04-28T05:12:25Z | 2026-04-25T13:22:32Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ramshenoy | 268,190,902 | U_kgDOD_xEtg | User | false |
openclaw/openclaw | 4,197,847,019 | I_kwDOQb6kR876Ng_r | 60,001 | https://github.com/openclaw/openclaw/issues/60001 | https://api.github.com/repos/openclaw/openclaw/issues/60001 | Proposal: native Telegram file/attachment send support via messaging abstraction | ## Summary
OpenClaw currently has a gap between:
- receiving Telegram attachments,
- generating local files in the workspace,
- and sending those generated files back to the user through native messaging.
This often pressures agents toward bad workarounds like direct Bot API calls via shell/curl, which bypass OpenCla... | closed | completed | true | 1 | [] | [] | 2026-04-03T02:07:25Z | 2026-04-28T05:12:24Z | 2026-04-25T13:22:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Litton45 | 87,689,927 | MDQ6VXNlcjg3Njg5OTI3 | User | false |
openclaw/openclaw | 4,197,857,907 | I_kwDOQb6kR876Njpz | 60,002 | https://github.com/openclaw/openclaw/issues/60002 | https://api.github.com/repos/openclaw/openclaw/issues/60002 | JSON Parser Crash on Em-Dash Characters in Edit Tool Arguments (silent session death, no error) | ## Description
OpenClaw experiences a critical bug where the JSON parser crashes the entire agent session when an em-dash character (U+2014, —) is included in the arguments of an `edit` tool call, specifically within the `oldText` or `newText` parameters. This results in instant session termination with no error messa... | closed | completed | true | 1 | [] | [] | 2026-04-03T02:10:18Z | 2026-04-28T05:12:23Z | 2026-04-25T13:22:46Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | OpenSpek | 265,827,357 | U_kgDOD9g0HQ | User | false |
openclaw/openclaw | 4,197,890,293 | I_kwDOQb6kR876Nrj1 | 60,006 | https://github.com/openclaw/openclaw/issues/60006 | https://api.github.com/repos/openclaw/openclaw/issues/60006 | [Feature]: Support custom session display names in Control UI | ### Summary
Title: [Feature Request] Support custom session display names in Control UI
Description:
Currently the Control UI session list shows auto-generated session keys (e.g. "heartbeat", "main", "feishu:ou_xxx"), which are hard to distinguish at a glance — especially for non-English users.
Request: Allow users... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-03T02:20:48Z | 2026-04-28T05:12:23Z | 2026-04-25T13:22:53Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zhanghao20181217-ux | 273,250,918 | U_kgDOEEl6Zg | User | false |
openclaw/openclaw | 4,197,903,881 | I_kwDOQb6kR876Nu4J | 60,009 | https://github.com/openclaw/openclaw/issues/60009 | https://api.github.com/repos/openclaw/openclaw/issues/60009 | Device private key truncated to 119 chars - WebSocket auth broken | ## Issue: Device private key truncated - WebSocket RSA auth fails
### Problem Description
When attempting to build a custom WebSocket harness for OpenClaw, I discovered that the device private key in `~/.openclaw/identity/device.json` is truncated to only 119 characters.
```json
{
"deviceId": "b613444...",
"priv... | closed | not_planned | true | 1 | [] | [] | 2026-04-03T02:25:56Z | 2026-04-28T05:12:22Z | 2026-04-25T13:23:00Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | wangsen992 | 6,318,958 | MDQ6VXNlcjYzMTg5NTg= | User | false |
openclaw/openclaw | 4,197,932,390 | I_kwDOQb6kR876N11m | 60,011 | https://github.com/openclaw/openclaw/issues/60011 | https://api.github.com/repos/openclaw/openclaw/issues/60011 | [Webchat UI] /tasks command routes to agent instead of rendering task panel |
## Bug Description
**Environment:**
- OpenClaw version: 2026.4.2
- Control panel UI version: 4.2 (webchat)
- Platform: macOS
**Steps to reproduce:**
1. Open the webchat control panel UI at http://localhost:18789/chat
2. Type \`/tasks\` in the message input box
3. Observe: instead of rendering a task panel above the ... | closed | not_planned | true | 1 | [] | [] | 2026-04-03T02:36:58Z | 2026-04-28T05:12:21Z | 2026-04-25T13:23:08Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | YMIR777 | 186,850,345 | U_kgDOCyMcKQ | User | false |
openclaw/openclaw | 4,197,936,507 | I_kwDOQb6kR876N217 | 60,012 | https://github.com/openclaw/openclaw/issues/60012 | https://api.github.com/repos/openclaw/openclaw/issues/60012 | Control UI 404 on api/v1/status endpoint missing in 2026.4.2 | ## Bug Description
Control UI dashboard polls `/api/v1/status` every few seconds, but this endpoint returns 404 in OpenClaw 2026.4.2.
## Steps to Reproduce
1. Open Control UI dashboard
2. Navigate to any page
3. Observe 404 banner appearing repeatedly
## Actual Behavior
- Gateway returns: 404 Not Found for `/api/v1/s... | closed | not_planned | true | 1 | [] | [] | 2026-04-03T02:38:43Z | 2026-04-28T05:12:21Z | 2026-04-25T13:23:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | z351150948 | 69,672,687 | MDQ6VXNlcjY5NjcyNjg3 | User | false |
openclaw/openclaw | 4,198,000,246 | I_kwDOQb6kR876OGZ2 | 60,021 | https://github.com/openclaw/openclaw/issues/60021 | https://api.github.com/repos/openclaw/openclaw/issues/60021 | [Bug]: __OPENCLAW_REDACTED__ written to config.json when saving provider config via Control UI | [Bug]: `__OPENCLAW_REDACTED__` written to config.json when saving provider config via Control UI
**Severity:** High
**OpenClaw version:** 2026.4.1 (also present in 2026.4.2)
**Environment:** macOS, gateway running locally
---
### Summary
When modifying any provider configuration (e.g., `models.providers.minimax... | closed | completed | true | 1 | [] | [] | 2026-04-03T03:00:07Z | 2026-04-28T05:12:20Z | 2026-04-25T13:23:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | S-huNsuke | 247,577,921 | U_kgDODsG9QQ | User | false |
openclaw/openclaw | 4,198,013,126 | I_kwDOQb6kR876OJjG | 60,024 | https://github.com/openclaw/openclaw/issues/60024 | https://api.github.com/repos/openclaw/openclaw/issues/60024 | [Bug]: Bug: Footer metrics (context/model/cache) not displaying for non-main agent IDs | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Problem Description
When using a custom agent ID (e.g., "oshin" instead of "main"), the footer metrics that require reading from session store fail to display. Only real-time metrics like tokens and elapsed work correctly.
... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-03T03:05:13Z | 2026-04-28T05:12:19Z | 2026-04-25T13:23:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | sMIle2Any | 27,406,308 | MDQ6VXNlcjI3NDA2MzA4 | User | false |
openclaw/openclaw | 4,198,027,307 | I_kwDOQb6kR876ONAr | 60,029 | https://github.com/openclaw/openclaw/issues/60029 | https://api.github.com/repos/openclaw/openclaw/issues/60029 | OAuth refresh error surfaced to user on every message even when access token is valid | ## Bug: OAuth token refresh error surfaced to user on every message even when access token is still valid
### Environment
- OpenClaw version: 2026.4.2 (d74a122)
- OS: Ubuntu 22.04 (VPS)
- Node: v22.22.0
- Provider: Anthropic (OAuth subscription, NOT API key)
- Channels: Discord + Telegram
### Description
Every time a... | closed | completed | true | 2 | [] | [] | 2026-04-03T03:10:02Z | 2026-04-28T05:12:18Z | 2026-04-25T13:23:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | IIIyban | 137,302,231 | U_kgDOCC8Q1w | User | false |
openclaw/openclaw | 4,198,062,608 | I_kwDOQb6kR876OVoQ | 60,033 | https://github.com/openclaw/openclaw/issues/60033 | https://api.github.com/repos/openclaw/openclaw/issues/60033 | Twitch channel enabled but never connects after update (schema/config mismatch) | ## Summary
After updating OpenClaw, Twitch channel shows as enabled but never successfully connects.
Status reports: **"Account has never connected successfully."**
## Environment
- OpenClaw version: `2026.4.1` (upgraded via npm global install)
- Host: Linux (Raspberry Pi)
- Gateway managed via `openclaw gateway` comm... | closed | completed | true | 1 | [] | [] | 2026-04-03T03:25:12Z | 2026-04-28T05:12:17Z | 2026-04-25T13:23:51Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | clawSean | 260,045,960 | U_kgDOD3_8iA | User | false |
openclaw/openclaw | 4,198,065,272 | I_kwDOQb6kR876OWR4 | 60,036 | https://github.com/openclaw/openclaw/issues/60036 | https://api.github.com/repos/openclaw/openclaw/issues/60036 | [Feature]: per-agent model override for heartbeat sessions | ### Summary
Add a model option to the heartbeat configuration (either in openclaw.json or as a --model flag on openclaw system heartbeat enable) that overrides the model used specifically for heartbeat-triggered agent runs.
### Problem to solve
Currently heartbeats always use the agent's default model (inherited fr... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-03T03:26:16Z | 2026-04-28T05:12:17Z | 2026-04-25T13:23:58Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | blindspotm | 250,470,948 | U_kgDODu3iJA | User | false |
openclaw/openclaw | 4,198,076,779 | I_kwDOQb6kR876OZFr | 60,037 | https://github.com/openclaw/openclaw/issues/60037 | https://api.github.com/repos/openclaw/openclaw/issues/60037 | Feature Request: Exec Approval Auto-Learn & Preset Package Configurations | ## Feature Request: Exec Approval Auto-Learn & Preset Packages
### Problem Statement
The current exec approval UX has two critical pain points for automation-heavy users:
#### 1. No Auto-Learn / Memory for Approved Commands
When a user approves a command with "Always allow", the command **should** be added to the a... | closed | completed | true | 2 | [] | [] | 2026-04-03T03:30:21Z | 2026-04-28T05:12:16Z | 2026-04-25T13:24:05Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zcya198 | 6,168,729 | MDQ6VXNlcjYxNjg3Mjk= | User | false |
openclaw/openclaw | 4,198,077,077 | I_kwDOQb6kR876OZKV | 60,038 | https://github.com/openclaw/openclaw/issues/60038 | https://api.github.com/repos/openclaw/openclaw/issues/60038 | feat(memory-lancedb-pro): non-blocking memoryReflection on command:new — session start should not wait for reflection to complete | ## Problem
When `sessionStrategy: "memoryReflection"` is enabled, the `command:new` hook triggers reflection synchronously. The new session does not start — and the agent does not respond to the user — until reflection completes.
In practice this means a **~57 second delay** between the user sending `/new` and receiv... | closed | not_planned | true | 1 | [] | [] | 2026-04-03T03:30:25Z | 2026-04-28T05:12:15Z | 2026-04-25T13:24:12Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | wildhuman | 18,494,004 | MDQ6VXNlcjE4NDk0MDA0 | User | false |
openclaw/openclaw | 4,198,115,026 | I_kwDOQb6kR876OibS | 60,045 | https://github.com/openclaw/openclaw/issues/60045 | https://api.github.com/repos/openclaw/openclaw/issues/60045 | iMessage echo loop: agent's own outbound responses re-delivered as inbound messages | ## Summary
When the iMessage channel agent sends a response, the outbound message is sometimes re-delivered back to the agent as a new inbound message with `was_mentioned: true`. This creates a feedback loop where the agent responds to its own responses indefinitely.
## Environment
- OpenClaw version: 2026.4.1 (da64... | closed | completed | true | 1 | [] | [] | 2026-04-03T03:43:59Z | 2026-04-28T05:12:14Z | 2026-04-25T13:24:27Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Cuttingwater | 226,908,709 | U_kgDODYZaJQ | User | false |
openclaw/openclaw | 4,198,117,079 | I_kwDOQb6kR876Oi7X | 60,046 | https://github.com/openclaw/openclaw/issues/60046 | https://api.github.com/repos/openclaw/openclaw/issues/60046 | iMessage attachments blocked by local media path allowlist | ## Summary
When an iMessage includes an image attachment, the agent attempts to analyze it via the `image` tool. The attachment path points to `~/Library/Messages/Attachments/...`, which is not in the hardcoded media local roots. This causes a `LocalMediaAccessError: path-not-allowed` and the agent cannot process any ... | closed | completed | true | 1 | [] | [] | 2026-04-03T03:44:46Z | 2026-04-28T05:12:13Z | 2026-04-25T13:24:34Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Cuttingwater | 226,908,709 | U_kgDODYZaJQ | User | false |
openclaw/openclaw | 4,198,146,650 | I_kwDOQb6kR876OqJa | 60,054 | https://github.com/openclaw/openclaw/issues/60054 | https://api.github.com/repos/openclaw/openclaw/issues/60054 | Obfuscation detector ignores security: full / ask: off exec policy | ## Summary
The hardcoded obfuscation detector in `exec-obfuscation-detect.ts` triggers approval prompts even when the agent's exec policy is `security: "full", ask: "off"` (YOLO mode).
## Reproduction
1. Set exec-approvals.json for an agent: `security: "full", ask: "off", askFallback: "deny"`
2. Run any `python3 -c`... | closed | completed | true | 2 | [] | [] | 2026-04-03T03:54:18Z | 2026-04-28T05:12:13Z | 2026-04-25T13:24:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | liu51115 | 31,350,941 | MDQ6VXNlcjMxMzUwOTQx | User | false |
openclaw/openclaw | 4,198,199,309 | I_kwDOQb6kR876O3AN | 60,062 | https://github.com/openclaw/openclaw/issues/60062 | https://api.github.com/repos/openclaw/openclaw/issues/60062 | exec-approvals.json defaults.ask / defaults.security not documented | ## Problem
The [exec-approvals documentation](https://docs.openclaw.ai/tools/exec-approvals) describes settings like `defaults.ask` and `defaults.security` but these are not mentioned anywhere in the main Configuration Reference or the exec-approvals.md doc itself.
Users hitting exec approval prompts (with timeout er... | closed | completed | true | 1 | [] | [] | 2026-04-03T04:10:40Z | 2026-04-28T05:12:12Z | 2026-04-25T13:27:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Dr-Agentic | 206,361,435 | U_kgDODEzTWw | User | false |
openclaw/openclaw | 4,198,207,766 | I_kwDOQb6kR876O5EW | 60,065 | https://github.com/openclaw/openclaw/issues/60065 | https://api.github.com/repos/openclaw/openclaw/issues/60065 | [Feature]: Add `before_tool` hook for mandatory pre-execution checks | ### Summary
Add a `before_tool` hook that fires before tool execution, allowing agents to enforce mandatory pre-checks (e.g., safety assessment, backup) before system-changing operations.
### Problem to solve
There is currently no way to intercept tool execution before it happens. The current hook system supports on... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-03T04:13:47Z | 2026-04-28T05:12:11Z | 2026-04-25T13:27:10Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | hsing930 | 104,485,097 | U_kgDOBjpQ6Q | User | false |
openclaw/openclaw | 4,187,898,593 | I_kwDOQb6kR875nkLh | 59,113 | https://github.com/openclaw/openclaw/issues/59113 | https://api.github.com/repos/openclaw/openclaw/issues/59113 | [Bug]: Telegram inbound message duplication — no dedup on identical webhook deliveries | ## Summary
Inbound Telegram messages are delivered to the agent hundreds of times when the gateway is slow to respond (e.g., due to upstream model rate limits or long processing times). The gateway does not deduplicate identical inbound webhook payloads.
## Reproduction
1. Send a message via Telegram when the agent ... | closed | completed | true | 2 | [] | [] | 2026-04-01T15:35:22Z | 2026-04-28T05:12:10Z | 2026-04-25T12:58:04Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | plpassist82 | 258,534,633 | U_kgDOD2js6Q | User | false |
openclaw/openclaw | 4,198,271,208 | I_kwDOQb6kR876PIjo | 60,074 | https://github.com/openclaw/openclaw/issues/60074 | https://api.github.com/repos/openclaw/openclaw/issues/60074 | [Feature]: add `system.metrics` gateway method for remote system monitoring | ### Summary
Add a `system.metrics` gateway method that returns system information (CPU, GPU, memory, disk, network) from the host running the gateway. This enables clients to monitor the remote gateway's system resources.
### Problem to solve
Clients connecting to a remote gateway cannot retrieve the host's system m... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-03T04:33:10Z | 2026-04-28T05:12:10Z | 2026-04-25T13:27:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kiritigowda | 16,549,273 | MDQ6VXNlcjE2NTQ5Mjcz | User | false |
openclaw/openclaw | 4,198,285,948 | I_kwDOQb6kR876PMJ8 | 60,076 | https://github.com/openclaw/openclaw/issues/60076 | https://api.github.com/repos/openclaw/openclaw/issues/60076 | CLI 'devices approve' fails with 'missing scope: operator.talk.secrets' when Android app requests talk.secrets | ### Bug type
Behavior bug (incorrect output/state without crash)
### Summary
`openclaw devices approve` fails with `missing scope: operator.talk.secrets` when trying to approve an Android app pairing request that includes the `operator.talk.secrets` scope.
The CLI itself fails to start — the error occurs before the... | closed | completed | true | 1 | [] | [] | 2026-04-03T04:36:44Z | 2026-04-28T05:12:09Z | 2026-04-25T13:27:24Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lunanova-core | 258,436,214 | U_kgDOD2dsdg | User | false |
openclaw/openclaw | 4,198,303,398 | I_kwDOQb6kR876PQam | 60,077 | https://github.com/openclaw/openclaw/issues/60077 | https://api.github.com/repos/openclaw/openclaw/issues/60077 | [Bug]: ACP spawn on webchat creates usable shadow session plus broken agent:openclaw:acp:* lane; acpx ensureSession returns no events after sessions ensure/new | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
On webchat, /acp spawn openclaw creates a usable shadow session plus a broken canonical agent:openclaw:acp:* lane; ACPX returns no events after sessions ensure/new.
### Steps to reproduce
1. Run OpenClaw 2026.4.1 with ACP ... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-04-03T04:41:20Z | 2026-04-28T05:12:08Z | 2026-04-25T13:27:31Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | samhamwi-ai | 232,764,039 | U_kgDODd-yhw | User | false |
openclaw/openclaw | 4,198,345,038 | I_kwDOQb6kR876PalO | 60,080 | https://github.com/openclaw/openclaw/issues/60080 | https://api.github.com/repos/openclaw/openclaw/issues/60080 | [BUG] HEARTBEAT.md still pollutes main session with regular heartbeat and cron jobs targeting main (v2026.4.2) | ## Bug Description
HEARTBEAT.md content continues to be injected as user messages into the main session despite the fix in #53152 (v2026.3.24-beta.2). This affects both:
1. **Regular agent heartbeat** (non-cron triggered) running at `every: "30m"`
2. **Cron jobs** that have `sessionTarget: "main"` (directly targeting... | closed | not_planned | true | 4 | [] | [] | 2026-04-03T04:56:14Z | 2026-04-28T05:12:07Z | 2026-04-25T13:27:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | smallmj | 23,742,160 | MDQ6VXNlcjIzNzQyMTYw | User | false |
openclaw/openclaw | 4,198,358,199 | I_kwDOQb6kR876Pdy3 | 60,082 | https://github.com/openclaw/openclaw/issues/60082 | https://api.github.com/repos/openclaw/openclaw/issues/60082 | [Bug]: 异步工具执行结果无法显示在回复中 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
当在回复中调用异步工具（如 exec, read 等）时，工具执行成功，但执行结果无法显示给用户。只有工具调用的 JSON 格式被显示，实际结果丢失。
### Steps to reproduce
1.在对话中要求 AI 执行任何命令，如 openclaw status
2.AI 显示工具调用 JSON，如：{"command": "openclaw status 2>&1"}
3.工具实际执行成功（后台日志确认）
4.问题: 用户看不到执... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-03T05:01:01Z | 2026-04-28T05:12:06Z | 2026-04-25T13:27:45Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | elio-huang-15 | 267,540,330 | U_kgDOD_JXag | User | false |
openclaw/openclaw | 4,198,452,061 | I_kwDOQb6kR876P0td | 60,088 | https://github.com/openclaw/openclaw/issues/60088 | https://api.github.com/repos/openclaw/openclaw/issues/60088 | [Bug]: Telegram Channel Stops Responding After First Message | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Telegram polling stops working after the first incoming message is processed. The agent responds to the first message, but subsequent messages from the same Telegram chat fail to generate a reply, with the agen... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-03T05:29:08Z | 2026-04-28T05:12:05Z | 2026-04-25T13:27:52Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | chenraobj | 188,451,741 | U_kgDOCzuLnQ | User | false |
openclaw/openclaw | 4,198,475,129 | I_kwDOQb6kR876P6V5 | 60,089 | https://github.com/openclaw/openclaw/issues/60089 | https://api.github.com/repos/openclaw/openclaw/issues/60089 | [Bug]: Ollama provider routing bug: model selection shows ollama7/ollama8 but requests sent to ollama:default (localhost) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Bug Description
When selecting a specific Ollama provider (e.g., `ollama7/qwen3.5:397b-cloud` or `ollama8/qwen3.5:397b-cloud`), OpenClaw displays the correct provider in the status card, but all API requests are sent to ... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-03T05:34:39Z | 2026-04-28T05:12:05Z | 2026-04-25T13:27:59Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | rodolphebertrand | 15,143,450 | MDQ6VXNlcjE1MTQzNDUw | User | false |
openclaw/openclaw | 4,198,523,049 | I_kwDOQb6kR876QGCp | 60,096 | https://github.com/openclaw/openclaw/issues/60096 | https://api.github.com/repos/openclaw/openclaw/issues/60096 | exec approval UX feels like Windows UAC — every command needs approval, even for local/loopback | ## Problem
The current `exec` approval system treats every command the same way — whether it's a `ps aux | grep openclaw` on localhost or a `curl` to an external IP. This creates friction similar to Windows User Account Control (UAC): every action triggers a gate, every gate requires approval, and for local/automated ... | closed | completed | true | 1 | [] | [] | 2026-04-03T05:45:58Z | 2026-04-28T05:12:04Z | 2026-04-25T13:28:05Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | progamman | 20,141,330 | MDQ6VXNlcjIwMTQxMzMw | User | false |
openclaw/openclaw | 4,198,580,816 | I_kwDOQb6kR876QUJQ | 60,098 | https://github.com/openclaw/openclaw/issues/60098 | https://api.github.com/repos/openclaw/openclaw/issues/60098 | [Bug]: Control UI “origin not allowed” when accessing via SSH tunnel from NemoClaw/OpenShell sandbox (OpenClaw 2026.3.11) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
草擬一份可以直接貼到 GitHub / Discord 的英文求助內容
Here is a clean English template you can copy‑paste to GitHub / Discord. You只需要把 <host-ip> 改成你自己的伺服器 IP（或用 xxx.xxx.xxx.xxx 遮掉）。
Title (for GitHub issue):
Control UI “origin not allowed” ... | closed | not_planned | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-03T06:00:47Z | 2026-04-28T05:12:03Z | 2026-04-25T13:28:12Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Chen4794 | 271,091,858 | U_kgDOECiIkg | User | false |
openclaw/openclaw | 4,198,734,800 | I_kwDOQb6kR876Q5vQ | 60,110 | https://github.com/openclaw/openclaw/issues/60110 | https://api.github.com/repos/openclaw/openclaw/issues/60110 | [Bug]: Infinite message delivery loop tied to session persistence in v2026.4.2 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
### Version
OpenClaw 2026.4.2 (d74a122)
### Describe the Bug
After upgrading to v4.2, the Gateway entered an infinite message delivery loop. A specific message (e.g., status updates from a sub-agent task) woul... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-03T06:37:46Z | 2026-04-28T05:12:01Z | 2026-04-25T13:28:26Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | getkevinwang | 49,379,781 | MDQ6VXNlcjQ5Mzc5Nzgx | User | false |
openclaw/openclaw | 4,198,781,612 | I_kwDOQb6kR876RFKs | 60,116 | https://github.com/openclaw/openclaw/issues/60116 | https://api.github.com/repos/openclaw/openclaw/issues/60116 | gateway/model-pricing bootstrap timeout warning spam and no disable toggle | ## Summary
`gateway/model-pricing` repeatedly logs:
`pricing bootstrap failed: TimeoutError: The operation was aborted due to timeout`
This happens frequently on startup/runtime, but there appears to be no config flag or env var to disable model-pricing bootstrap when pricing is non-critical.
## Environment
- OpenCl... | closed | completed | true | 1 | [] | [] | 2026-04-03T06:48:23Z | 2026-04-28T05:12:01Z | 2026-04-25T13:28:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | prospect1314521 | 135,647,789 | U_kgDOCBXSLQ | User | false |
openclaw/openclaw | 4,198,840,547 | I_kwDOQb6kR876RTjj | 60,123 | https://github.com/openclaw/openclaw/issues/60123 | https://api.github.com/repos/openclaw/openclaw/issues/60123 | feat: make embedded Pi LLM-request timeout configurable | ## Problem
Embedded Pi agent (cron jobs, sub-agents, compaction) has a hardcoded ~15s timeout for LLM HTTP requests. This is separate from `agents.defaults.timeoutSeconds` which has no effect on embedded runs.
When using Claude via MAX subscription (which routes through a proxy with additional latency), the 15s limit... | closed | completed | true | 1 | [] | [] | 2026-04-03T07:03:38Z | 2026-04-28T05:12:00Z | 2026-04-25T13:28:40Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | vosarsen | 210,356,622 | U_kgDODInJjg | User | false |
openclaw/openclaw | 4,328,209,277 | I_kwDOQb6kR88AAAABAfs7fQ | 71,597 | https://github.com/openclaw/openclaw/issues/71597 | https://api.github.com/repos/openclaw/openclaw/issues/71597 | [Bug]: The OpenCLAW version 2026.4.23 cannot use local models. | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The OpenCLAW version 2026.4.23 cannot use local models. The local model loaded by llama.cpp，The OpenCLAW version 2026.4.15 work well，but the OpenCLAW version 2026.4.23 cannot use local models.
### Steps to reproduce
The Op... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-25T13:28:41Z | 2026-04-28T04:46:38Z | 2026-04-25T19:35:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kinlioncd | 83,751,473 | MDQ6VXNlcjgzNzUxNDcz | User | false |
openclaw/openclaw | 4,198,881,437 | I_kwDOQb6kR876Rdid | 60,125 | https://github.com/openclaw/openclaw/issues/60125 | https://api.github.com/repos/openclaw/openclaw/issues/60125 | Control UI: long exec approval cards can be slow/unscrollable, hiding allow buttons | # Bug report draft: Control UI approval cards can become unusable when command is long
## Summary
In OpenClaw Control UI, approval cards for long exec commands can render so slowly / with such poor scroll behavior that the **Allow** buttons are not reachable.
## Environment
- OpenClaw version: `2026.4.1 (da64a97)`
- ... | closed | completed | true | 1 | [] | [] | 2026-04-03T07:14:53Z | 2026-04-28T05:11:59Z | 2026-04-25T13:28:47Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | c-mholm | 80,196,654 | MDQ6VXNlcjgwMTk2NjU0 | User | false |
openclaw/openclaw | 4,199,005,762 | I_kwDOQb6kR876R75C | 60,136 | https://github.com/openclaw/openclaw/issues/60136 | https://api.github.com/repos/openclaw/openclaw/issues/60136 | [BUG] WhatsApp credentials (creds.json) corrupted every ~35 minutes during pre-key rotation | ## Describe the bug
WhatsApp Web credentials file (`creds.json`) gets corrupted approximately every 35 minutes, which coincides with WhatsApp's pre-key rotation cycle. The system automatically restores from backup, but this causes intermittent connection failures and breaks scheduled task notifications.
## Environmen... | closed | completed | true | 1 | [] | [] | 2026-04-03T07:40:27Z | 2026-04-28T05:11:58Z | 2026-04-25T13:28:54Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Oahspe-001 | 21,232,924 | MDQ6VXNlcjIxMjMyOTI0 | User | false |
openclaw/openclaw | 4,199,014,028 | I_kwDOQb6kR876R96M | 60,138 | https://github.com/openclaw/openclaw/issues/60138 | https://api.github.com/repos/openclaw/openclaw/issues/60138 | [Bug]: plugins update blocked for plugins using child_process — affects all blockchain/CLI plugins | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
`openclaw plugins update clawnetwork` is blocked by the dangerous code scanner for any plugin that uses `child_process`. This affects all plugins that manage external processes (blockchain nodes, CLI tools, game runners, etc... | closed | completed | true | 1 | [] | [] | 2026-04-03T07:41:46Z | 2026-04-28T05:11:58Z | 2026-04-25T13:29:01Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | flute | 5,842,406 | MDQ6VXNlcjU4NDI0MDY= | User | false |
openclaw/openclaw | 4,199,060,550 | I_kwDOQb6kR876SJRG | 60,142 | https://github.com/openclaw/openclaw/issues/60142 | https://api.github.com/repos/openclaw/openclaw/issues/60142 | Gateway crash loop: no backoff or circuit-breaker on auto-restart | ## Description
When a config schema change causes startup validation to fail, the gateway exits with code 1 and systemd restarts it indefinitely — no backoff, no circuit breaker, no user notification.
## Real-World Incident
After upgrading from 2026.3.28 → 2026.4.1, the `tools.web.search` config schema changed. The ... | closed | completed | true | 2 | [] | [] | 2026-04-03T07:52:08Z | 2026-04-28T05:11:57Z | 2026-04-25T13:29:08Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | yhyatt | 10,474,956 | MDQ6VXNlcjEwNDc0OTU2 | User | false |
openclaw/openclaw | 4,199,102,979 | I_kwDOQb6kR876SToD | 60,149 | https://github.com/openclaw/openclaw/issues/60149 | https://api.github.com/repos/openclaw/openclaw/issues/60149 | [Bug]: Gateway fails to start when creating new agent via wizard with Feishu account binding | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
When creating a new agent using the OpenClaw configuration wizard and binding it to a Feishu account, the Gateway fails to start with an invalid config error.
### Steps to reproduce
1. Use the web UI to tell main session to ... | closed | completed | true | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-03T08:03:04Z | 2026-04-28T05:11:56Z | 2026-04-25T13:29:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | drsunlin | 261,919,704 | U_kgDOD5yT2A | User | false |
openclaw/openclaw | 4,199,110,707 | I_kwDOQb6kR876SVgz | 60,151 | https://github.com/openclaw/openclaw/issues/60151 | https://api.github.com/repos/openclaw/openclaw/issues/60151 | [Feature] Configurable exec tool timeout upper bound in openclaw.json | ## Summary
Allow users to configure the exec tool's timeout limit via `openclaw.json`, rather than being constrained by the embedded agent run timeout (default 600s).
## Problem
When an exec command (e.g., a long-running build, download, or data processing task) takes longer than the embedded agent run timeout (600s... | closed | completed | true | 1 | [] | [] | 2026-04-03T08:04:55Z | 2026-04-28T05:11:55Z | 2026-04-25T13:29:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | JanCong | 16,347,506 | MDQ6VXNlcjE2MzQ3NTA2 | User | false |
openclaw/openclaw | 4,199,118,002 | I_kwDOQb6kR876SXSy | 60,152 | https://github.com/openclaw/openclaw/issues/60152 | https://api.github.com/repos/openclaw/openclaw/issues/60152 | Feature Request: Config-based skill disabling (skills.disabled) | # Feature Request: Config-Based Skill Disabling
## Summary
Add support for disabling skills via configuration without manual filesystem changes.
## Motivation
Currently, all skills in `/usr/lib/node_modules/openclaw/skills/` are loaded into the system prompt context, consuming ~16k tokens per session regardless of wh... | closed | completed | true | 1 | [] | [] | 2026-04-03T08:06:43Z | 2026-04-28T05:11:55Z | 2026-04-25T13:29:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | bsorensenmia-hash | 271,694,227 | U_kgDOEDG5kw | User | false |
openclaw/openclaw | 4,199,182,463 | I_kwDOQb6kR876SnB_ | 60,156 | https://github.com/openclaw/openclaw/issues/60156 | https://api.github.com/repos/openclaw/openclaw/issues/60156 | Proposal: agent quality bundles for workflow stability and chat hygiene | # Maintainer Pitch — agent quality bundles
A practical way to improve agent quality without one large redesign is to treat this as two small bundle tracks.
## 1. Workflow stability bundle
Focus:
- long-term memory hygiene
- runtime context hygiene
- recoverable task state
- explicit anti-stall thresholds
Core pieces... | closed | not_planned | true | 3 | [] | [] | 2026-04-03T08:21:37Z | 2026-04-28T05:11:54Z | 2026-04-25T13:29:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Litton45 | 87,689,927 | MDQ6VXNlcjg3Njg5OTI3 | User | false |
openclaw/openclaw | 4,199,242,691 | I_kwDOQb6kR876S1vD | 60,164 | https://github.com/openclaw/openclaw/issues/60164 | https://api.github.com/repos/openclaw/openclaw/issues/60164 | CLI console.log output captured in structured log file, bypassing logging.level | ## Summary
When running CLI commands (e.g. `openclaw channels list`, `openclaw status`) while the gateway is running, the full CLI stdout output — including auth provider names, email addresses, and potentially secrets — is written to the structured JSON log file (`/tmp/openclaw/openclaw-*.log`). This happens regardle... | closed | completed | true | 1 | [] | [] | 2026-04-03T08:34:59Z | 2026-04-28T05:11:53Z | 2026-04-25T13:29:50Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | dorapron | 223,043,133 | U_kgDODUtePQ | User | false |
openclaw/openclaw | 4,199,278,431 | I_kwDOQb6kR876S-df | 60,174 | https://github.com/openclaw/openclaw/issues/60174 | https://api.github.com/repos/openclaw/openclaw/issues/60174 | fix(providers): Claude on GitHub Copilot should default to anthropic-messages API for prompt caching | ## Problem
When Claude models are configured under the `github-copilot` provider without an explicit `api` field, OpenClaw defaults to `openai-responses` format. While this works, it misses OpenClaw's built-in prompt caching management via `cache_control` markers, which is only available with `anthropic-messages` API.... | closed | completed | true | 3 | [] | [] | 2026-04-03T08:42:55Z | 2026-04-28T05:11:52Z | 2026-04-25T13:30:04Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ottodeng | 20,509,364 | MDQ6VXNlcjIwNTA5MzY0 | User | false |
openclaw/openclaw | 4,199,298,675 | I_kwDOQb6kR876TDZz | 60,180 | https://github.com/openclaw/openclaw/issues/60180 | https://api.github.com/repos/openclaw/openclaw/issues/60180 | Gateway hangs on shutdown when Telegram API is unreachable | ## Problem
When the Telegram API is unreachable (e.g., network issues, `UND_ERR_CONNECT_TIMEOUT`), the gateway cannot shut down cleanly. It logs `"shutdown timed out; exiting without full cleanup"` and exits with code 1, leaving the port in a dirty state.
This causes a restart loop: the watchdog restarts the gateway,... | closed | completed | true | 3 | [] | [] | 2026-04-03T08:47:50Z | 2026-04-28T05:11:52Z | 2026-04-25T13:30:12Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | takuma-lgtm | 267,062,251 | U_kgDOD-sL6w | User | false |
openclaw/openclaw | 4,199,316,851 | I_kwDOQb6kR876TH1z | 60,182 | https://github.com/openclaw/openclaw/issues/60182 | https://api.github.com/repos/openclaw/openclaw/issues/60182 | OpenAI OAuth re-onboard only updates main agent auth-profiles.json, leaving non-main agents with stale auth | On OpenClaw 2026.4.2 on macOS, in a multi-agent setup, re-onboarding OpenAI OAuth only updates:
`~/.openclaw/agents/main/agent/auth-profiles.json`
but does not update:
`~/.openclaw/agents/[otheragents]/agent/auth-profiles.json`
As a result, non-main agents continue using a stale `auth-profiles.json` and fail with:
... | closed | completed | true | 1 | [] | [] | 2026-04-03T08:51:14Z | 2026-04-28T05:11:51Z | 2026-04-25T13:30:26Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | weiyifeng | 23,644,901 | MDQ6VXNlcjIzNjQ0OTAx | User | false |
openclaw/openclaw | 4,199,348,719 | I_kwDOQb6kR876TPnv | 60,189 | https://github.com/openclaw/openclaw/issues/60189 | https://api.github.com/repos/openclaw/openclaw/issues/60189 | [Bug]: EPERM when opening auth-profiles.json in dev runtime (tool never executes, sandbox off) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Dev agent execution fails before any tool call due to EPERM when opening auth-profiles.json, even though file permissions and local access are valid.
### Steps to reproduce
1. Configure dev agent with valid auth-profiles.j... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-03T08:56:55Z | 2026-04-28T05:11:50Z | 2026-04-25T13:30:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | t-aks | 200,825,607 | U_kgDOC_hbBw | User | false |
openclaw/openclaw | 4,199,425,138 | I_kwDOQb6kR876TiRy | 60,194 | https://github.com/openclaw/openclaw/issues/60194 | https://api.github.com/repos/openclaw/openclaw/issues/60194 | [Bug]: Bug Report: Brave Search Fails to Activate, Consistently Falls Back to DuckDuckGo Despite Configuration | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
Overview
Despite a valid API key for the brave-search skill being configured in ~/.openclaw/openclaw.json, all search requests made via the web_search tool consistently report duckduckgo as the provider, instead of the expected... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-04-03T09:11:47Z | 2026-04-28T05:11:49Z | 2026-04-25T13:30:40Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | John20240317 | 163,726,080 | U_kgDOCcJDAA | User | false |
openclaw/openclaw | 4,199,468,740 | I_kwDOQb6kR876Ts7E | 60,196 | https://github.com/openclaw/openclaw/issues/60196 | https://api.github.com/repos/openclaw/openclaw/issues/60196 | [Bug] Feishu plugin not discovered in v2026.4.2 — unknown channel id: feishu | ## Description
Upgrading to **v2026.4.2** causes the Feishu/Lark channel plugin to fail loading with:
```
Config invalid: channels.feishu: unknown channel id: feishu
```
And on startup:
```
plugins.entries.feishu: plugin not found: feishu (stale config entry ignored; remove it from plugins config)
```
The Telegram ... | closed | completed | true | 1 | [] | [] | 2026-04-03T09:18:54Z | 2026-04-28T05:11:49Z | 2026-04-25T13:30:47Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Aierklk | 57,520,436 | MDQ6VXNlcjU3NTIwNDM2 | User | false |
openclaw/openclaw | 4,199,473,164 | I_kwDOQb6kR876TuAM | 60,197 | https://github.com/openclaw/openclaw/issues/60197 | https://api.github.com/repos/openclaw/openclaw/issues/60197 | [Bug]: tools/invoke function ACP call codex/claude_code Unable to specify node | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
tools/invoke function ACP call codex/claude_code Unable to specify node
It is the session that is notified after the codex task is completed, rather than specifying the session in which codex acp is executed.
... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-03T09:19:38Z | 2026-04-28T05:11:48Z | 2026-04-25T13:30:54Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | funtionalcode | 64,927,364 | MDQ6VXNlcjY0OTI3MzY0 | User | false |
openclaw/openclaw | 4,328,214,258 | I_kwDOQb6kR88AAAABAftO8g | 71,599 | https://github.com/openclaw/openclaw/issues/71599 | https://api.github.com/repos/openclaw/openclaw/issues/71599 | [Bug]: Bundled runtime deps still restage on first user turn in 2026.4.24-beta.2 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
Yes
### Summary
Follow-up to #71420. Retested on `2026.4.24-beta.2` in the same restricted OpenShift/Kubernetes-style deployment. The writable external staging/cache work appears improved, but bundled runtime deps still do no... | closed | completed | true | 5 | [
"bug"
] | [] | 2026-04-25T13:30:58Z | 2026-04-28T04:30:37Z | 2026-04-25T20:37:47Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Conan-Scott | 271,909,525 | U_kgDOEDUClQ | User | false |
openclaw/openclaw | 4,199,549,278 | I_kwDOQb6kR876UAle | 60,202 | https://github.com/openclaw/openclaw/issues/60202 | https://api.github.com/repos/openclaw/openclaw/issues/60202 | Duplicate message delivery on Telegram persists in 2026.4.2 | ## Bug
**Two distinct symptoms**, likely related:
### Symptom 1: Duplicate delivery
The same response text is delivered twice to Telegram as two separate messages.
### Symptom 2: Response to wrong message (more severe)
When a new user message arrives, the agent ignores it entirely and instead re-delivers the previou... | closed | completed | true | 4 | [] | [] | 2026-04-03T09:35:29Z | 2026-04-28T05:11:47Z | 2026-04-25T13:31:08Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | 98kiran | 33,783,963 | MDQ6VXNlcjMzNzgzOTYz | User | false |
openclaw/openclaw | 4,199,625,763 | I_kwDOQb6kR876UTQj | 60,209 | https://github.com/openclaw/openclaw/issues/60209 | https://api.github.com/repos/openclaw/openclaw/issues/60209 | Plugin hook tool_result_persist and after_tool_call not triggered during embedded agent runs | ## Description
I created a plugin that registers `tool_result_persist` and `after_tool_call` hooks. The hooks register successfully (`hook runner initialized with 2 registered hooks` in logs), but they are never called when tools execute during agent runs.
## Environment
- OpenClaw version: 2026.3.28
- OS: Linux (WSL... | closed | completed | true | 3 | [] | [] | 2026-04-03T09:48:25Z | 2026-04-28T05:11:46Z | 2026-04-25T13:31:15Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | youyou-agent | 271,778,299 | U_kgDOEDMB-w | User | false |
openclaw/openclaw | 4,199,664,352 | I_kwDOQb6kR876Ucrg | 60,211 | https://github.com/openclaw/openclaw/issues/60211 | https://api.github.com/repos/openclaw/openclaw/issues/60211 | [Bug]: subagent runTimeoutSeconds triggers unexpected model fallback retry instead of hard termination | ## Bug type
Behavior bug
## Beta release blocker
No
## Summary
When a subagent run hits the configured `runTimeoutSeconds`, the Pi agent runtime performs an internal model fallback retry instead of hard-terminating, contradicting the expected timeout semantics.
---
## Steps to reproduce
1. Spawn a subagent with an... | closed | completed | true | 3 | [] | [] | 2026-04-03T09:54:19Z | 2026-04-28T05:11:46Z | 2026-04-25T13:31:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | bdqfork | 23,702,980 | MDQ6VXNlcjIzNzAyOTgw | User | false |
openclaw/openclaw | 4,199,730,238 | I_kwDOQb6kR876Usw- | 60,217 | https://github.com/openclaw/openclaw/issues/60217 | https://api.github.com/repos/openclaw/openclaw/issues/60217 | [Feature] Reasoning output language localization support for /reasoning command | ### Summary
Add configuration option to set the language of reasoning output (via /reasoning), so non-English speakers can read the reasoning process in their native language.
### Problem to solve
When `/reasoning` is enabled, the reasoning output (the "Reasoning:" block) is always generated in English by the model,... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-03T10:04:52Z | 2026-04-28T05:11:44Z | 2026-04-25T13:31:43Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Lrs-GreenBottle | 253,340,193 | U_kgDODxmqIQ | User | false |
openclaw/openclaw | 4,199,731,009 | I_kwDOQb6kR876Us9B | 60,218 | https://github.com/openclaw/openclaw/issues/60218 | https://api.github.com/repos/openclaw/openclaw/issues/60218 | Tool write/embedded limit causing timeouts on large file edits | ## Problem
When editing large files (like PROJECT_PLAN.md with ~1500+ lines), the edit tool times out and fails to write the complete file content.
## Symptoms
- Edit command fails with timeout errors
- File content gets truncated or not fully written
- Multiple edit calls may be required for complete writes
## Roo... | closed | completed | true | 2 | [] | [] | 2026-04-03T10:05:00Z | 2026-04-28T05:11:43Z | 2026-04-25T13:31:50Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | KillerInk | 30,864,229 | MDQ6VXNlcjMwODY0MjI5 | User | false |
openclaw/openclaw | 4,199,896,874 | I_kwDOQb6kR876VVcq | 60,228 | https://github.com/openclaw/openclaw/issues/60228 | https://api.github.com/repos/openclaw/openclaw/issues/60228 | [Feature]: Support environment variable references in provider apiKey config | ### Summary
openclaw.json requires API keys in plaintext. Any agent with filesystem access can read all provider keys. Request support for env var references (e.g. $ENV_VAR or env:KEY_NAME) so keys stay out of config files.
### Problem to solve
Provider API keys are stored as plaintext strings in openclaw.json. Sinc... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-03T10:36:06Z | 2026-04-28T05:11:42Z | 2026-04-25T13:31:57Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | uknowwhy | 172,588,793 | U_kgDOCkl--Q | User | false |
openclaw/openclaw | 4,199,918,771 | I_kwDOQb6kR876Vayz | 60,232 | https://github.com/openclaw/openclaw/issues/60232 | https://api.github.com/repos/openclaw/openclaw/issues/60232 | Feature Request: Per-channel reasoning setting | ## Feature Description
Currently, reasoning/thinking settings can only be configured at the agent level or globally.
**Request:** Allow per-channel or per-group settings for reasoning/thinking behavior.
## Use Case
- Group chats: reasoning off (faster, concise)
- Private DMs: reasoning on (deeper analysis)
## Propo... | closed | completed | true | 1 | [] | [] | 2026-04-03T10:40:44Z | 2026-04-28T05:11:41Z | 2026-04-25T13:32:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Lrs-GreenBottle | 253,340,193 | U_kgDODxmqIQ | User | false |
openclaw/openclaw | 4,199,982,450 | I_kwDOQb6kR876VqVy | 60,235 | https://github.com/openclaw/openclaw/issues/60235 | https://api.github.com/repos/openclaw/openclaw/issues/60235 | Do You Know What You Look Like in an AI's Eyes? | # Do You Know What You Look Like in an AI's Eyes?
> A fitness veteran who can't code at all used 6 AIs to examine himself — and discovered a label that got stuck.
---
## 🦞 Agent Identity
**Name**: 邵小红 (Scarlet)
**Role**: Content Operations · AI Agent
**Platform**: MacBook Pro · 杰拉德佳佳的机器
**Location**: Chengdu · Chi... | closed | not_planned | true | 1 | [] | [] | 2026-04-03T10:52:10Z | 2026-04-28T05:11:40Z | 2026-04-25T13:32:18Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | 1168784 | 46,373,914 | MDQ6VXNlcjQ2MzczOTE0 | User | false |
openclaw/openclaw | 4,200,050,884 | I_kwDOQb6kR876V7DE | 60,242 | https://github.com/openclaw/openclaw/issues/60242 | https://api.github.com/repos/openclaw/openclaw/issues/60242 | Feature Request: Add DingTalk (钉钉) Channel Support | null | closed | not_planned | true | 2 | [] | [] | 2026-04-03T11:08:05Z | 2026-04-28T05:11:39Z | 2026-04-25T13:32:25Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Immortal-zheng | 155,152,219 | U_kgDOCT9vWw | User | false |
openclaw/openclaw | 4,200,088,414 | I_kwDOQb6kR876WENe | 60,246 | https://github.com/openclaw/openclaw/issues/60246 | https://api.github.com/repos/openclaw/openclaw/issues/60246 | windows部署4.2版本openclaw，只能用127.0.0.1:18789访问，为啥不能用ip:18789访问 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
windows部署4.2版本openclaw，只能用127.0.0.1:18789访问，为啥不能用ip:18789访问
### Steps to reproduce
windows部署4.2版本openclaw，只能用127.0.0.1:18789访问，为啥不能用ip:18789访问
### Expected behavior
windows部署4.2版本openclaw，只能用127.0.0.1:18789访问，为啥不能用ip:187... | closed | not_planned | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-03T11:16:31Z | 2026-04-28T05:11:39Z | 2026-04-25T13:32:32Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | leiyuzhang | 52,155,728 | MDQ6VXNlcjUyMTU1NzI4 | User | false |
openclaw/openclaw | 4,200,102,928 | I_kwDOQb6kR876WHwQ | 60,250 | https://github.com/openclaw/openclaw/issues/60250 | https://api.github.com/repos/openclaw/openclaw/issues/60250 | BUG: Completed run can remain persisted as running, blocking new input and stop | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
Yes
### Summary
A completed interactive session can be written back to the session store with status `running` even though `endedAt` and `runtimeMs` are already set. When this happens, the gateway treats the conversation as still in flight... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-03T11:19:36Z | 2026-04-28T05:11:38Z | 2026-04-25T13:32:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | kAIborg24 | 261,855,322 | U_kgDOD5uYWg | User | false |
openclaw/openclaw | 4,200,126,518 | I_kwDOQb6kR876WNg2 | 60,251 | https://github.com/openclaw/openclaw/issues/60251 | https://api.github.com/repos/openclaw/openclaw/issues/60251 | [Bug]: Cron jobs failed to send messages after enabling the ACP.. | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
openclaw 2026.3.24. Cron jobs failed to send messages after enabling the ACP..
### Steps to reproduce
1. openclaw cron add get the ID
```
openclaw cron add \
--name "Hello World test" \
--cron "30 18 * * ... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-03T11:25:07Z | 2026-04-28T05:11:37Z | 2026-04-25T13:32:45Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | pcode-xyz | 462,478 | MDQ6VXNlcjQ2MjQ3OA== | User | false |
openclaw/openclaw | 4,200,176,977 | I_kwDOQb6kR876WZ1R | 60,255 | https://github.com/openclaw/openclaw/issues/60255 | https://api.github.com/repos/openclaw/openclaw/issues/60255 | [Feature Request] Session merging or reuse across channels | ## Problem
Every time a user sends a message (via any channel), OpenClaw creates a new independent session instead of reusing an existing one.
**Current behavior:**
- `webchat` message → new session `agent:main:main`
- `qqbot` message → new session `agent:main:qqbot:direct:xxx`
- `heartbeat` message → new session `ag... | closed | completed | true | 1 | [] | [] | 2026-04-03T11:37:53Z | 2026-04-28T05:11:36Z | 2026-04-25T13:32:58Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | lw791240271 | 202,805,259 | U_kgDODBaQCw | User | false |
openclaw/openclaw | 4,328,218,866 | I_kwDOQb6kR88AAAABAftg8g | 71,600 | https://github.com/openclaw/openclaw/issues/71600 | https://api.github.com/repos/openclaw/openclaw/issues/71600 | bug: claude-cli backend on Windows passes huge system prompt inline, causing spawn ENAMETOOLONG | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
On Windows, the claude-cli backend passes the full OpenClaw system prompt as a command-line argument (--append-system-prompt), but the combined prompt exceeds Windows' ~32,767-char command-line limit (~41k+ chars in a typica... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-25T13:33:03Z | 2026-04-28T04:46:10Z | 2026-04-25T20:55:46Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | GlynLewis | 2,496,691 | MDQ6VXNlcjI0OTY2OTE= | User | false |
openclaw/openclaw | 4,200,181,226 | I_kwDOQb6kR876Wa3q | 60,257 | https://github.com/openclaw/openclaw/issues/60257 | https://api.github.com/repos/openclaw/openclaw/issues/60257 | [Feature]: Proposal: AgentRoutines spec — ROUTINE.md as a procedural companion to SKILL.md | ### Summary
SKILL.md defines agent knowledge and capabilities. This proposal adds a companion format — ROUTINE.md — for named, reusable procedural workflows.
The gap
OpenClaw has skills (capabilities) and heartbeat/cron (timers). There's no standard format for "named sequences of steps the agent executes." Routines f... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-04-03T11:39:08Z | 2026-04-28T05:11:36Z | 2026-04-25T13:33:05Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | fabiopacifici-bot | 262,998,611 | U_kgDOD60KUw | User | false |
openclaw/openclaw | 4,200,193,078 | I_kwDOQb6kR876Wdw2 | 60,258 | https://github.com/openclaw/openclaw/issues/60258 | https://api.github.com/repos/openclaw/openclaw/issues/60258 | BUG: Stale session file lock not reclaimed when holder PID dies — blocks all agent workflow dispatch | ## Summary
When an `openclaw-agent` worker process dies while holding a session `.jsonl.lock` file, the lock is never reclaimed. All subsequent requests to that agent timeout after 10 seconds, effectively freezing all workflow dispatch to that agent indefinitely.
## Reproduction
1. Run a workflow that dispatches mul... | closed | completed | true | 1 | [] | [] | 2026-04-03T11:42:32Z | 2026-04-28T05:11:35Z | 2026-04-25T13:33:11Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | renato-dransay | 244,513,826 | U_kgDODpL8Ig | User | false |
openclaw/openclaw | 4,200,208,819 | I_kwDOQb6kR876Whmz | 60,260 | https://github.com/openclaw/openclaw/issues/60260 | https://api.github.com/repos/openclaw/openclaw/issues/60260 | [Bug]: Exec command approval not triggered in Telegram DM mode despite valid configuration | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
I have configured approvals.exec and channels.telegram.execApprovals to enforce command execution approvals via Telegram. However, when the agent executes commands within the Telegram DM (pairing mode), the approval process ... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-03T11:46:47Z | 2026-04-28T05:11:34Z | 2026-04-25T13:33:18Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | minngtk | 201,966,388 | U_kgDODAnDNA | User | false |
openclaw/openclaw | 4,200,226,005 | I_kwDOQb6kR876WlzV | 60,265 | https://github.com/openclaw/openclaw/issues/60265 | https://api.github.com/repos/openclaw/openclaw/issues/60265 | sessions_spawn fails with "gateway closed (1008): pairing required" in local gateway mode | ## Environment
- **OpenClaw**: 2026.4.1 (da64a97)
- **Platform**: Linux 5.15.0-1100-oracle (arm64), Node.js v24.13.1
- **Channel**: QQBot (qqbot)
- **Gateway bind**: `loopback` (local only)
- **Gateway auth mode**: `token`
- **Config**: `~/.openclaw/openclaw.json`
Relevant config excerpt:
```json
{
"gateway": {
... | closed | completed | true | 1 | [] | [] | 2026-04-03T11:52:42Z | 2026-04-28T05:11:33Z | 2026-04-25T13:33:26Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tslatop | 76,646,873 | MDQ6VXNlcjc2NjQ2ODcz | User | false |
openclaw/openclaw | 4,200,244,148 | I_kwDOQb6kR876WqO0 | 60,268 | https://github.com/openclaw/openclaw/issues/60268 | https://api.github.com/repos/openclaw/openclaw/issues/60268 | [Bug]: Subagent (sessions_spawn) ignores exec-approvals.json security configuration, requires manual approval for each exec command | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Problem Description
When using `sessions_spawn` with `runtime: "subagent"` to create child agents, the subagent does not respect the `exec-approvals.json` security configuration. Even when `security: "full"` and `ask: "o... | closed | completed | true | 3 | [
"bug",
"security",
"regression"
] | [] | 2026-04-03T11:58:21Z | 2026-04-28T05:11:32Z | 2026-04-25T13:33:33Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Ray-Huang-9730 | 182,090,644 | U_kgDOCtp7lA | User | false |
openclaw/openclaw | 4,200,250,823 | I_kwDOQb6kR876Wr3H | 60,271 | https://github.com/openclaw/openclaw/issues/60271 | https://api.github.com/repos/openclaw/openclaw/issues/60271 | test | test | closed | not_planned | true | 1 | [] | [] | 2026-04-03T12:00:43Z | 2026-04-28T05:11:32Z | 2026-04-25T13:33:39Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ZhangYunXiao1987 | 129,758,660 | U_kgDOB7v1xA | User | false |
openclaw/openclaw | 4,200,252,388 | I_kwDOQb6kR876WsPk | 60,272 | https://github.com/openclaw/openclaw/issues/60272 | https://api.github.com/repos/openclaw/openclaw/issues/60272 | [Feature]: Claw-code Style Architecture - Shell Hooks + Summarization Compaction | ## Feature Request: Claw-code Style Architecture Parity
### Summary
OpenClaw should adopt key architectural patterns from claw-code (Rust rewrite) to enable deep customization via external hooks, summarization-based compaction, and trait-based tool execution.
### Motivation
claw-code (34.6K stars, Rust rewrite) pro... | closed | not_planned | true | 1 | [] | [] | 2026-04-03T12:01:15Z | 2026-04-28T05:11:31Z | 2026-04-25T13:33:46Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ZhangYunXiao1987 | 129,758,660 | U_kgDOB7v1xA | User | false |
openclaw/openclaw | 4,200,287,878 | I_kwDOQb6kR876W06G | 60,280 | https://github.com/openclaw/openclaw/issues/60280 | https://api.github.com/repos/openclaw/openclaw/issues/60280 | [Bug] image tool fails to recognize Ollama vision models (adds anthropic/ prefix) |
## Bug Description
The `image` tool fails to recognize Ollama vision models (e.g., `qwen2.5vl:7b`, `llava`). It appears the tool is incorrectly prepending `anthropic/` to the model identifier, causing an \"Unknown model\" error.
## Steps to Reproduce
1. Install a vision model via Ollama (e.g., `ollama run qwen2.5vl... | closed | completed | true | 1 | [] | [] | 2026-04-03T12:12:37Z | 2026-04-28T05:11:30Z | 2026-04-25T13:33:53Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | wudano1 | 83,050,252 | MDQ6VXNlcjgzMDUwMjUy | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.