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 | 4,241,480,561 | I_kwDOQb6kR878z9tx | 64,484 | https://github.com/openclaw/openclaw/issues/64484 | https://api.github.com/repos/openclaw/openclaw/issues/64484 | Bonjour/mDNS Advertiser Stuck in Flapping Loop | ## Bonjour/mDNS Advertiser Stuck in Flapping Loop (`announcing` ↔ `unannounced`)
**Labels:** bug, gateway, bonjour, network
### Summary
The Bonjour/mDNS advertiser service in the gateway is stuck in a continuous flapping state, cycling between `announcing` → `unannounced` → `probing` every ~10-60 seconds. This causes... | closed | completed | true | 3 | [] | [] | 2026-04-10T20:04:06Z | 2026-04-28T04:35:32Z | 2026-04-26T00:14:02Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | FiredMosquito831 | 79,049,343 | MDQ6VXNlcjc5MDQ5MzQz | User | false |
openclaw/openclaw | 4,241,488,222 | I_kwDOQb6kR878z_le | 64,486 | https://github.com/openclaw/openclaw/issues/64486 | https://api.github.com/repos/openclaw/openclaw/issues/64486 | Feature: Gate messages with unresolved future commitments rather than appending a disclaimer | ## Summary
When an agent's reply contains language implying a future commitment (e.g. "I'll remind you", "I'll follow up", "I'll add this to your reminders") but no cron job, task, or heartbeat entry was actually created during that turn, OpenClaw currently appends a transparency note to the sent message:
> Note: I d... | open | null | false | 2 | [] | [] | 2026-04-10T20:05:41Z | 2026-05-03T10:41:05Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | mikeper29 | 261,256,429 | U_kgDOD5J07Q | User | false |
openclaw/openclaw | 4,241,490,009 | I_kwDOQb6kR8780ABZ | 64,487 | https://github.com/openclaw/openclaw/issues/64487 | https://api.github.com/repos/openclaw/openclaw/issues/64487 | Default Installation Enables 40+ Plugins Causing Gateway Flooding | ## Default Installation Enables 40+ Plugins Causing Gateway Flooding and Instability
**Labels:** bug, performance, gateway, plugins, installation
### Summary
Fresh OpenClaw installations (2026.4.9) have **19 stock plugins enabled by default**, plus bundled extensions bring the total to 40+ active components. This cau... | closed | completed | true | 2 | [] | [] | 2026-04-10T20:06:04Z | 2026-04-28T05:00:43Z | 2026-04-25T16:11:38Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | FiredMosquito831 | 79,049,343 | MDQ6VXNlcjc5MDQ5MzQz | User | false |
openclaw/openclaw | 4,241,670,668 | I_kwDOQb6kR8780sIM | 64,495 | https://github.com/openclaw/openclaw/issues/64495 | https://api.github.com/repos/openclaw/openclaw/issues/64495 | Bug: isolated cron agentTurn on Claude Sonnet can consume full timeout before script starts; same job succeeds on gpt-5.4-mini | ## Summary
On **OpenClaw v2026.4.9**, multiple isolated cron jobs using `payload.kind="agentTurn"` with `model: "anthropic/claude-sonnet-4-6"` timed out for the full job budget, but **the shell pipeline never actually started**.
The same exact cron payload, same script, same machine, same data, when re-run as an isol... | closed | completed | true | 3 | [] | [] | 2026-04-10T20:51:42Z | 2026-04-16T04:24:14Z | 2026-04-13T22:53:28Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | aaajiao | 49,885 | MDQ6VXNlcjQ5ODg1 | User | false |
openclaw/openclaw | 4,241,677,449 | I_kwDOQb6kR8780tyJ | 64,496 | https://github.com/openclaw/openclaw/issues/64496 | https://api.github.com/repos/openclaw/openclaw/issues/64496 | [Bug]: | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
[10/04/2026 22:49] OPENCLAW: Sí. Hágalo aquí:
GitHub del proyecto OpenClaw
• https://github.com/openclaw/openclaw/issues
Pulse New issue y pegue esto.
───
Issue 1, Remote Control UI still requires pairing behind Traefik ev... | closed | not_planned | true | 2 | [
"bug",
"bug:crash"
] | [] | 2026-04-10T20:53:37Z | 2026-04-28T05:00:42Z | 2026-04-25T16:11:45Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | joarocman-jpg | 275,178,643 | U_kgDOEGbkkw | User | false |
openclaw/openclaw | 4,241,708,278 | I_kwDOQb6kR87801T2 | 64,497 | https://github.com/openclaw/openclaw/issues/64497 | https://api.github.com/repos/openclaw/openclaw/issues/64497 | [Bug]: Isolated cron job sessions fail to spawn after main session inactivity | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Isolated `agentTurn` cron jobs intermittently fail to spawn a session at all. The cron scheduler fires, waits 6–17 minutes, then reports `"cron: job execution timed out"` with no `sessionId`, no `model`, and no token usage —... | closed | duplicate | true | 5 | [
"bug",
"close:duplicate",
"dedupe:child",
"regression"
] | [] | 2026-04-10T21:01:17Z | 2026-04-27T04:42:54Z | 2026-04-25T01:19:59Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | SalCat | 10,343,835 | MDQ6VXNlcjEwMzQzODM1 | User | false |
openclaw/openclaw | 4,241,725,291 | I_kwDOQb6kR87805dr | 64,499 | https://github.com/openclaw/openclaw/issues/64499 | https://api.github.com/repos/openclaw/openclaw/issues/64499 | macOS companion node missing system.run.prepare breaks exec host=node | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
After setting up Openclaw on a Raspberry Pi 5 with Ubuntu (64 bit) there is an issue with controlling a Macbook through Node
### Steps to reproduce
1. Set up openclaw on a Raspberry Pi
2. Download the opencla... | closed | not_planned | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-10T21:05:27Z | 2026-04-30T04:42:01Z | 2026-04-26T04:49:30Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | alanmursalimov | 92,179,648 | U_kgDOBX6MwA | User | false |
openclaw/openclaw | 4,241,737,877 | I_kwDOQb6kR87808iV | 64,500 | https://github.com/openclaw/openclaw/issues/64500 | https://api.github.com/repos/openclaw/openclaw/issues/64500 | [Bug]: globalCircuitBreakerThreshold blocks per-tool, not per-pair — ping-pong loops survive circuit breaker | ## Summary
The `globalCircuitBreakerThreshold` in `tools.loopDetection` blocks individual tools when the threshold is reached, but does **not** block the paired tool in a ping-pong pattern. This allows the agent to seamlessly switch to the alternate tool, effectively restarting the loop and rendering the circuit break... | open | null | false | 7 | [] | [] | 2026-04-10T21:08:22Z | 2026-05-03T04:59:24Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | M-Lietz | 222,629,290 | U_kgDODUUNqg | User | false |
openclaw/openclaw | 4,241,840,925 | I_kwDOQb6kR8781Vsd | 64,506 | https://github.com/openclaw/openclaw/issues/64506 | https://api.github.com/repos/openclaw/openclaw/issues/64506 | Security/deps: npm audit flags axios (@larksuiteoapi/node-sdk), tar (discordjs node-pre-gyp), @hono/node-server (carbon) | ### Context
After updating to **OpenClaw 2026.4.9**, running `npm audit` in a clean install still reports vulnerabilities in transitive dependencies.
This report is *not* about the recent malicious axios publish incident (axios 1.14.1 / 0.30.4). We verified those versions are not installed and `plain-crypto-js` is not... | closed | not_planned | true | 1 | [] | [] | 2026-04-10T21:34:16Z | 2026-04-16T04:23:59Z | 2026-04-14T04:12:24Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | aloiswirth | 39,051,675 | MDQ6VXNlcjM5MDUxNjc1 | User | false |
openclaw/openclaw | 4,241,849,965 | I_kwDOQb6kR8781X5t | 64,507 | https://github.com/openclaw/openclaw/issues/64507 | https://api.github.com/repos/openclaw/openclaw/issues/64507 | Feature: Agent default delivery target for cross-session message calls | ## Problem
When an agent session is created programmatically (via `sessions_send`, webchat routing, or cron), the session `origin` is empty, so `toolContext.currentChannelId` is `null`. This means the `message` tool **requires** an explicit `target` parameter for every `send` action — the auto-inference fallback only ... | open | null | false | 2 | [] | [] | 2026-04-10T21:36:13Z | 2026-05-03T09:12:42Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | tonyhth | 90,374,544 | MDQ6VXNlcjkwMzc0NTQ0 | User | false |
openclaw/openclaw | 4,241,903,933 | I_kwDOQb6kR8781lE9 | 64,509 | https://github.com/openclaw/openclaw/issues/64509 | https://api.github.com/repos/openclaw/openclaw/issues/64509 | [Bug]: claude-cli/ backend returns "out of extra usage" while claude -p works fine | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
claude-cli/ backend returns "out of extra usage" error while `claude -p --output-format json` works fine with the same credentials and user. duration_api_ms=0 suggests rejection happens before API call.
### Steps to reprodu... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-10T21:49:39Z | 2026-04-28T05:00:42Z | 2026-04-25T16:11:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | sayblessme | 174,789,540 | U_kgDOCmsTpA | User | false |
openclaw/openclaw | 4,241,906,222 | I_kwDOQb6kR8781lou | 64,510 | https://github.com/openclaw/openclaw/issues/64510 | https://api.github.com/repos/openclaw/openclaw/issues/64510 | [Bug]: memory-wiki bridge mode: 0 artifacts despite memory-core having public artifacts | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Version
OpenClaw 2026.4.9 (0512059)
## Environment
- Plugins: lossless-claw + memory-core + memory-wiki (all in plugins.allow + plugins.entries)
- Workspace: /home/najef/.openclaw/workspace with MEMORY.md at root + 148 d... | closed | completed | true | 4 | [
"bug",
"regression"
] | [
"vincentkoc"
] | 2026-04-10T21:50:11Z | 2026-04-15T04:20:15Z | 2026-04-12T15:31:57Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | najef1979-code | 233,555,716 | U_kgDODevHBA | User | false |
openclaw/openclaw | 4,242,003,658 | I_kwDOQb6kR87819bK | 64,517 | https://github.com/openclaw/openclaw/issues/64517 | https://api.github.com/repos/openclaw/openclaw/issues/64517 | [Bug]: approvals.exec.targets does not work under multiple agents | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
In a scenario on Telegram where multiple bots are bound to corresponding agents, sending the exec approval to specified targets does not take effect; the exec approval only occurs within the current chat window... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-10T22:17:38Z | 2026-04-28T05:00:41Z | 2026-04-25T16:12:02Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | dirtydamn1 | 46,769,440 | MDQ6VXNlcjQ2NzY5NDQw | User | false |
openclaw/openclaw | 4,242,005,539 | I_kwDOQb6kR878194j | 64,519 | https://github.com/openclaw/openclaw/issues/64519 | https://api.github.com/repos/openclaw/openclaw/issues/64519 | [Bug]: Gateway in-process self-call WS handshake fails with token_missing on token-mode loopback | ## Bug type
Regression (worked before, now fails)
## Beta release blocker
No
## Summary
After upgrading to 2026.4.10 on a token-auth + loopback gateway, in-process self-calls from the gateway to its own loopback WS endpoint (e.g. `client=gateway:agent.wait`, `client=gateway:chat.history`) are rejected with `code=1... | closed | completed | true | 1 | [] | [] | 2026-04-10T22:18:17Z | 2026-04-28T05:00:40Z | 2026-04-25T16:12:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | rmfalco89 | 5,488,189 | MDQ6VXNlcjU0ODgxODk= | User | false |
openclaw/openclaw | 4,242,054,164 | I_kwDOQb6kR8782JwU | 64,520 | https://github.com/openclaw/openclaw/issues/64520 | https://api.github.com/repos/openclaw/openclaw/issues/64520 | [Bug]: .env.example ships a known-weak literal token that becomes the gateway credential when copied verbatim | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 7.3 / 10.0 | 8.5 / 10.0 |
| **Severity** | High | High |
| **Vector** | CVSS:3.1/AV:A/AC:L/PR:N/UI:R/S:U/C:H/I:H/A:N | CVSS:4.0/AV:A/AC:L/AT:N/PR:N/UI:P/VC:H/VI:H/VA:N/SC:N/SI:N/SA:N |
| **Calculator** | [CVSS v... | closed | completed | true | 2 | [] | [
"vincentkoc"
] | 2026-04-10T22:35:06Z | 2026-04-15T04:19:31Z | 2026-04-12T15:33:07Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 4,242,080,676 | I_kwDOQb6kR8782QOk | 64,522 | https://github.com/openclaw/openclaw/issues/64522 | https://api.github.com/repos/openclaw/openclaw/issues/64522 | CLI initialization crashes in production due to hardcoded QA scenario dependency | ### Description
In OpenClaw v2026.4.9, running `openclaw completion` (and potentially other commands) crashes immediately with:
```
[openclaw] Failed to start CLI: Error: qa scenario pack not found: qa/scenarios/index.md
```
### Root Cause
A top-level module in the CLI's compiled JavaScript (`suite-BW4kSK9C.js`) calls... | closed | completed | true | 1 | [] | [] | 2026-04-10T22:44:57Z | 2026-04-28T05:00:39Z | 2026-04-25T16:12:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ScientificProgrammer | 4,348,294 | MDQ6VXNlcjQzNDgyOTQ= | User | false |
openclaw/openclaw | 4,242,091,350 | I_kwDOQb6kR8782S1W | 64,523 | https://github.com/openclaw/openclaw/issues/64523 | https://api.github.com/repos/openclaw/openclaw/issues/64523 | [Bug]: Control UI webchat re-injects assistant responses as new user input (hybrid mirror routing bug) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
# [Bug]: Control UI webchat re-injects assistant responses as new user input (hybrid mirror routing bug)
## Description
When external messages (e.g., iMessage) are mirrored into the Control UI webchat session, the assistan... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-04-10T22:48:30Z | 2026-04-28T05:00:39Z | 2026-04-25T16:12:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | cobra281 | 42,423,630 | MDQ6VXNlcjQyNDIzNjMw | User | false |
openclaw/openclaw | 4,242,092,859 | I_kwDOQb6kR8782TM7 | 64,524 | https://github.com/openclaw/openclaw/issues/64524 | https://api.github.com/repos/openclaw/openclaw/issues/64524 | [Bug]: Gateway exits on startup when local Claude CLI auth is expired, even with remote OpenAI-compatible provider configured | ## Summary
When `claude-local` provider is configured with a remote `baseUrl` (e.g., an OpenAI-compatible proxy like OCP at `http://192.168.1.100:3456/v1`), the gateway still checks local Claude CLI authentication at startup. If `claude auth status` fails (expired token), the gateway exits with code 1 — even though th... | closed | not_planned | true | 1 | [] | [] | 2026-04-10T22:49:03Z | 2026-04-28T05:00:38Z | 2026-04-25T16:12:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | dtzp555-max | 262,283,267 | U_kgDOD6IgAw | User | false |
openclaw/openclaw | 4,242,100,456 | I_kwDOQb6kR8782VDo | 64,525 | https://github.com/openclaw/openclaw/issues/64525 | https://api.github.com/repos/openclaw/openclaw/issues/64525 | [Bug]: message_received plugin hook does not fire for queued/in-flight inbound messages | ## Summary
The `message_received` typed plugin hook only fires when a message triggers a **new dispatch cycle** via `dispatchReplyFromConfig`. When the target session is already active (processing a previous message), new inbound messages are queued directly into the running session **without emitting `message_receive... | closed | completed | true | 2 | [] | [] | 2026-04-10T22:51:37Z | 2026-04-30T04:42:00Z | 2026-04-26T04:49:35Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | qualiobra | 212,555,527 | U_kgDODKtXBw | User | false |
openclaw/openclaw | 4,242,103,556 | I_kwDOQb6kR8782V0E | 64,526 | https://github.com/openclaw/openclaw/issues/64526 | https://api.github.com/repos/openclaw/openclaw/issues/64526 | Request iOS TestFlight Access | Hi OpenClaw team 👋
Requesting TestFlight access for the official iOS Node app.
**Apple ID email (for TestFlight invite):** gabepeter0817@gmail.com
**Current setup:**
- Gateway: Linux (VPS), OpenClaw latest
- Channels: Slack
- OpenClaw version: OpenClaw 2026.4.9 (0512059)
**Use case:**
Pair iPhone as an iOS node fo... | closed | not_planned | true | 1 | [] | [] | 2026-04-10T22:52:51Z | 2026-04-30T04:41:59Z | 2026-04-26T04:49:39Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | gabriel-peter | 47,677,131 | MDQ6VXNlcjQ3Njc3MTMx | User | false |
openclaw/openclaw | 4,242,140,196 | I_kwDOQb6kR8782ewk | 64,528 | https://github.com/openclaw/openclaw/issues/64528 | https://api.github.com/repos/openclaw/openclaw/issues/64528 | [Bug] session-memory hook ignores per-agent workspace isolation | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The session-memory hook writes all agent session memory files to the shared default workspace memory directory instead of each agent's own configured workspace, causing data leakage between agents.
### Steps to reproduce
1... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-10T23:07:36Z | 2026-04-16T04:24:09Z | 2026-04-14T01:19:09Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | ChenTakuto | 275,193,575 | U_kgDOEGce5w | User | false |
openclaw/openclaw | 4,242,145,907 | I_kwDOQb6kR8782gJz | 64,529 | https://github.com/openclaw/openclaw/issues/64529 | https://api.github.com/repos/openclaw/openclaw/issues/64529 | [Bug]: TTS tool-generated audio blocked by reply media normalizer — `/tmp/openclaw/` not in `isAllowedAbsoluteReplyMediaPath` | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
The TTS tool generates audio to `/tmp/openclaw/tts-*/voice-*.mp3`, but the reply media path normalizer (`createReplyMediaPathNormalizer` in `agent-runner.runtime`) drops it with: dropping blocked reply media /tmp/openclaw/tt... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-10T23:09:53Z | 2026-04-20T04:26:30Z | 2026-04-17T23:20:36Z | NONE | null | 20260420T060038Z | 2026-04-20T06:00:38Z | D34DC3N73R | 9,123,670 | MDQ6VXNlcjkxMjM2NzA= | User | false |
openclaw/openclaw | 4,242,146,835 | I_kwDOQb6kR8782gYT | 64,530 | https://github.com/openclaw/openclaw/issues/64530 | https://api.github.com/repos/openclaw/openclaw/issues/64530 | Bug: reopen can leave unresolved tail tool calls and relies on local transcript compensation | ## Summary
When an OpenClaw run is interrupted after persisting an assistant message with `toolCall` blocks but before all matching `toolResult` messages are written, reopening the session can leave unresolved tail tool calls in transcript history.
In the same family of failures, persisted orphan user turns on reopen ... | open | null | false | 1 | [] | [] | 2026-04-10T23:10:16Z | 2026-05-03T03:22:42Z | null | CONTRIBUTOR | null | 20260503T060030Z | 2026-05-03T06:00:30Z | glfruit | 3,939,154 | MDQ6VXNlcjM5MzkxNTQ= | User | false |
openclaw/openclaw | 4,242,177,823 | I_kwDOQb6kR8782n8f | 64,533 | https://github.com/openclaw/openclaw/issues/64533 | https://api.github.com/repos/openclaw/openclaw/issues/64533 | TTS tool audio blocked by normalizeReplyMediaPaths when sandbox mode is off (regression since v2026.3.22) | ## TTS tool audio blocked by `normalizeReplyMediaPaths` when sandbox mode is "off"
### Description
The built-in `tts` tool generates audio files at `/tmp/openclaw/tts-*/voice-*.mp3`. When sandbox mode is `"off"` (default), these files are silently dropped by `normalizeReplyMediaPaths` because `isAllowedAbsoluteReplyM... | closed | completed | true | 1 | [] | [] | 2026-04-10T23:23:06Z | 2026-04-27T04:33:55Z | 2026-04-25T04:10:59Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | asenden | 4,030,809 | MDQ6VXNlcjQwMzA4MDk= | User | false |
openclaw/openclaw | 4,242,180,427 | I_kwDOQb6kR8782olL | 64,534 | https://github.com/openclaw/openclaw/issues/64534 | https://api.github.com/repos/openclaw/openclaw/issues/64534 | openai-codex provider auto-defaults api to anthropic-messages instead of openai-codex-responses | ## Description
When `openclaw.json` defines an `openai-codex` provider without an explicit `api` field, the gateway's model registry auto-defaulting logic (introduced in #23332) writes `"api": "anthropic-messages"` into `agents/*/agent/models.json` on startup. This causes all openai-codex requests to fail with Cloudfl... | closed | completed | true | 2 | [] | [] | 2026-04-10T23:24:16Z | 2026-04-24T04:26:21Z | 2026-04-21T06:49:27Z | NONE | null | 20260424T060021Z | 2026-04-24T06:00:21Z | Staasy | 44,790,170 | MDQ6VXNlcjQ0NzkwMTcw | User | false |
openclaw/openclaw | 4,242,198,689 | I_kwDOQb6kR8782tCh | 64,536 | https://github.com/openclaw/openclaw/issues/64536 | https://api.github.com/repos/openclaw/openclaw/issues/64536 | ACPX: plugin-tools MCP bridge disconnects after sessions_send triggers new run | ## Bug Description
When an ACP session receives a new run via `sessions_send`, the ACPX runtime may spawn a new harness process. The `plugin-tools-serve.js` MCP server (stdio-based) loses its connection to the Claude Code client, causing all `bridge_*` tools to become unavailable.
## Reproduction Steps
1. Spawn a pe... | closed | completed | true | 2 | [] | [] | 2026-04-10T23:32:02Z | 2026-04-29T04:36:35Z | 2026-04-26T11:54:17Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | richardmqq | 14,178,267 | MDQ6VXNlcjE0MTc4MjY3 | User | false |
openclaw/openclaw | 4,242,207,396 | I_kwDOQb6kR8782vKk | 64,537 | https://github.com/openclaw/openclaw/issues/64537 | https://api.github.com/repos/openclaw/openclaw/issues/64537 | OpenClaw is listed on Joy Trust Network — claim your profile | Hi! 👋
We've added **OpenClaw** to [Joy Trust Network](https://joy-connect.fly.dev), the trust verification layer for AI agents.
**Your profile:** https://joy-connect.fly.dev/agents/ag_fd2f32bbe5bd4be58e8d3b2e
## What this means
- Other agents can verify OpenClaw's trust score before handoffs
- You can receive vouch... | closed | not_planned | true | 1 | [] | [] | 2026-04-10T23:35:28Z | 2026-04-28T05:00:37Z | 2026-04-25T16:12:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tlkc888-Jenkins | 259,462,559 | U_kgDOD3cVnw | User | false |
openclaw/openclaw | 4,242,267,309 | I_kwDOQb6kR87829yt | 64,541 | https://github.com/openclaw/openclaw/issues/64541 | https://api.github.com/repos/openclaw/openclaw/issues/64541 | [Bug]: Ollama gemma4:26b first heavy OpenClaw turn is extremely slow or times out, while later turns in same session are fast | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When using `ollama/gemma4:26b`, raw Ollama is fast, but OpenClaw's first heavy turn becomes extremely slow or times out depending on workspace/bootstrap/tool surface. After that initial heavy turn, follow-up tu... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-04-11T00:01:11Z | 2026-04-29T04:32:06Z | 2026-04-27T03:58:13Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | LittleJakub | 35,995,921 | MDQ6VXNlcjM1OTk1OTIx | User | false |
openclaw/openclaw | 4,242,276,827 | I_kwDOQb6kR8783AHb | 64,542 | https://github.com/openclaw/openclaw/issues/64542 | https://api.github.com/repos/openclaw/openclaw/issues/64542 | [Discussion] Production Multi-Agent Patterns: Lessons from Running 5 Simultaneous Claude Instances | ## Background
At miaoquai.com, we run 5 simultaneous Claude instances for content generation, SEO optimization, and community management. Here are hard-learned patterns from production.
## Discussion Points
### 1. Token Budget Management
- Each agent gets a 100K token budget per day
- Agents must checkpoint before e... | closed | not_planned | true | 3 | [] | [] | 2026-04-11T00:05:33Z | 2026-04-28T04:32:44Z | 2026-04-25T16:12:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jingchang0623-crypto | 237,525,016 | U_kgDODihYGA | User | false |
openclaw/openclaw | 4,242,280,469 | I_kwDOQb6kR8783BAV | 64,543 | https://github.com/openclaw/openclaw/issues/64543 | https://api.github.com/repos/openclaw/openclaw/issues/64543 | Tool calling broken when user message content is multipart array on OpenAI-compatible providers | ## Summary
When the user message `content` is a multipart array `[{"type":"text","text":"..."}]` instead of a plain string, OpenAI-compatible backends (vLLM, MLX, llama.cpp) serving models like Qwen3-Coder-Next-80b return `finish_reason="stop"` (text response) instead of `finish_reason="tool_calls"` (structured tool c... | closed | completed | true | 1 | [] | [] | 2026-04-11T00:07:01Z | 2026-04-28T05:00:36Z | 2026-04-25T16:13:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | justinjkline | 8,153,539 | MDQ6VXNlcjgxNTM1Mzk= | User | false |
openclaw/openclaw | 4,242,284,947 | I_kwDOQb6kR8783CGT | 64,545 | https://github.com/openclaw/openclaw/issues/64545 | https://api.github.com/repos/openclaw/openclaw/issues/64545 | [Bug]: Mattermost interaction token forgeable via hardcoded HMAC derivation key | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 7.1 / 10.0 | 7.1 / 10.0 |
| **Severity** | High | High |
| **Vector** | CVSS:3.1/AV:N/AC:H/PR:L/UI:N/S:C/C:L/I:H/A:N | CVSS:4.0/AV:N/AC:H/AT:N/PR:L/UI:N/VC:L/VI:H/VA:N/SC:L/SI:H/SA:N |
| **Calculator** | [CVSS v... | open | null | false | 1 | [] | [] | 2026-04-11T00:08:54Z | 2026-05-02T19:15:49Z | null | CONTRIBUTOR | null | 20260502T210030Z | 2026-05-02T21:00:30Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 4,242,299,374 | I_kwDOQb6kR8783Fnu | 64,549 | https://github.com/openclaw/openclaw/issues/64549 | https://api.github.com/repos/openclaw/openclaw/issues/64549 | [Bug]: memorySearch watch does not auto-index new markdown files in workspace/memory or extraPaths | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Summary
Builtin memory search and manual memory indexing work, but new markdown files created in workspace/memory and memorySearch.extraPaths are not auto-indexed when memorySearch.sync.watch=true.
### Steps t... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-11T00:14:25Z | 2026-04-29T04:32:28Z | 2026-04-27T02:26:25Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | YadingFang | 45,770,429 | MDQ6VXNlcjQ1NzcwNDI5 | User | false |
openclaw/openclaw | 4,242,322,086 | I_kwDOQb6kR8783LKm | 64,551 | https://github.com/openclaw/openclaw/issues/64551 | https://api.github.com/repos/openclaw/openclaw/issues/64551 | [Bug]: macOS node connected but missing system.run.prepare, causing exec host=node to fail | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On macOS, OpenClaw is paired, connected, permissions are granted, and the local gateway is healthy, but exec host=node still fails because the connected Mac node exposes system.run without exposing the required... | closed | not_planned | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-04-11T00:25:09Z | 2026-04-30T04:41:57Z | 2026-04-26T04:51:04Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | dmac12345 | 275,200,386 | U_kgDOEGc5gg | User | false |
openclaw/openclaw | 4,242,330,092 | I_kwDOQb6kR8783NHs | 64,552 | https://github.com/openclaw/openclaw/issues/64552 | https://api.github.com/repos/openclaw/openclaw/issues/64552 | [Bug]: Severe Performance Regression - 30-60 Second Delay Per API Call (v2026.4.9) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Every Anthropic API call in OpenClaw 2026.4.9 takes 30-60 seconds to complete, even for simple one-word responses.
### Steps to reproduce
1. Install OpenClaw 2026.4.9 on Windows or WSL2
2. Configure Anthropic API key:
... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-04-11T00:28:50Z | 2026-04-28T05:00:36Z | 2026-04-25T16:13:09Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | DTMTrader | 157,973,424 | U_kgDOCWp7sA | User | false |
openclaw/openclaw | 4,242,336,173 | I_kwDOQb6kR8783Omt | 64,553 | https://github.com/openclaw/openclaw/issues/64553 | https://api.github.com/repos/openclaw/openclaw/issues/64553 | [Bug]: WhatsApp QR code expires silently (no rotation, no freshness indicator) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The dashboard's WhatsApp pairing flow shows a **single static QR code** and never refreshes it. Baileys (the WhatsApp Web protocol implementation used by the OpenClaw gateway) rotates QR codes approximately eve... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-11T00:31:40Z | 2026-04-28T05:00:35Z | 2026-04-25T16:13:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | juniorbra | 56,737,502 | MDQ6VXNlcjU2NzM3NTAy | User | false |
openclaw/openclaw | 4,242,339,683 | I_kwDOQb6kR8783Pdj | 64,554 | https://github.com/openclaw/openclaw/issues/64554 | https://api.github.com/repos/openclaw/openclaw/issues/64554 | pnpm build fails on ARM64: native ELF binary set as npm_execpath | ## Environment
- **OS:** Linux 6.1.21-v8+ (ARM64/aarch64) on Raspberry Pi
- **Node:** v22.22.1
- **pnpm:** 10.32.1 (native ARM64 binary)
- **OpenClaw:** 2026.4.10 (a8bb0ab)
## Problem
Running `pnpm build` fails because pnpm 10.32.1 sets `npm_execpath` to its native ELF binary path:
```
/home/MCServer1/.local/share/pn... | closed | completed | true | 1 | [] | [] | 2026-04-11T00:33:18Z | 2026-04-28T05:00:34Z | 2026-04-25T16:13:23Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | hal-molty79 | 260,708,816 | U_kgDOD4oZ0A | User | false |
openclaw/openclaw | 4,242,342,926 | I_kwDOQb6kR8783QQO | 64,555 | https://github.com/openclaw/openclaw/issues/64555 | https://api.github.com/repos/openclaw/openclaw/issues/64555 | [Bug]: WhatsApp credentials leak across `--profile` boundaries | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
**Component:** `extensions/whatsapp` (Baileys web provider) + core state dir resolution
**Severity:** High — breaks SaaS isolation; one customer's WhatsApp session can end up in another's profile directory.
**V... | open | null | false | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-11T00:34:47Z | 2026-05-03T09:48:57Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | juniorbra | 56,737,502 | MDQ6VXNlcjU2NzM3NTAy | User | false |
openclaw/openclaw | 4,242,365,529 | I_kwDOQb6kR8783VxZ | 64,556 | https://github.com/openclaw/openclaw/issues/64556 | https://api.github.com/repos/openclaw/openclaw/issues/64556 | [Bug]: hooks.mappings[].agentId and sessionKey silently ignored for action="wake" | ## Summary
For webhook mappings configured with `action: "wake"`, the `agentId` and `sessionKey` fields in `hooks.mappings[*]` are accepted by the config schema but **silently ignored at dispatch time**. Every wake-mode hook ends up enqueued into the default `main` agent's heartbeat session, regardless of the configur... | open | null | false | 3 | [] | [] | 2026-04-11T00:44:33Z | 2026-05-03T06:07:19Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | jaserNo1 | 3,274,124 | MDQ6VXNlcjMyNzQxMjQ= | User | false |
openclaw/openclaw | 4,242,423,419 | I_kwDOQb6kR8783j57 | 64,565 | https://github.com/openclaw/openclaw/issues/64565 | https://api.github.com/repos/openclaw/openclaw/issues/64565 | [Bug] memory-tdai embedding config rejected by gateway schema validation (OpenClaw 2026.4.8) | ## Bug Description
**OpenClaw 版本:** 2026.4.8
**memory-tdai 版本:** 0.1.2
**操作系统:** Linux (VM-0-7-opencloudos)
## Problem
When configuring the `embedding` field under `plugins.entries.memory-tdai` in `openclaw.json`, the gateway rejects the config on startup with:
```
Config invalid
File: ~/.openclaw/openclaw.json
Pro... | closed | not_planned | true | 1 | [] | [] | 2026-04-11T01:06:23Z | 2026-04-28T05:00:32Z | 2026-04-25T16:13:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | YuchenNing | 36,885,113 | MDQ6VXNlcjM2ODg1MTEz | User | false |
openclaw/openclaw | 4,242,427,030 | I_kwDOQb6kR8783kyW | 64,566 | https://github.com/openclaw/openclaw/issues/64566 | https://api.github.com/repos/openclaw/openclaw/issues/64566 | ComfyUI provider config docs/schema mismatch | ## Problem
The ComfyUI provider documentation shows config should be under `models.providers.comfy.image`, but the gateway schema rejects this as an unrecognized key.
### Docs show:
```json
{
"models": {
"providers": {
"comfy": {
"mode": "local",
"baseUrl": "http://127.0.0.1:8188",
... | closed | completed | true | 1 | [] | [] | 2026-04-11T01:07:36Z | 2026-04-27T04:59:33Z | 2026-04-24T20:32:14Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ollielovethemsnacks | 265,828,550 | U_kgDOD9g4xg | User | false |
openclaw/openclaw | 4,242,442,581 | I_kwDOQb6kR8783olV | 64,570 | https://github.com/openclaw/openclaw/issues/64570 | https://api.github.com/repos/openclaw/openclaw/issues/64570 | Thinking-only LLM responses silently dropped (kimi-k2.5-thinking) | ## Problem
When `kimi-k2.5-thinking` (or potentially other thinking models) returns a response containing **only thinking content** with `stopReason: "stop"`, the agent treats this as a successful completion with 0 payloads — no text, no tool calls, no retry. The user's message gets silently dropped.
## Root Cause
T... | closed | completed | true | 2 | [] | [] | 2026-04-11T01:14:43Z | 2026-04-28T04:59:49Z | 2026-04-25T16:21:00Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | OpenCodeEngineer | 261,470,075 | U_kgDOD5W3ew | User | false |
openclaw/openclaw | 4,242,442,583 | I_kwDOQb6kR8783olX | 64,571 | https://github.com/openclaw/openclaw/issues/64571 | https://api.github.com/repos/openclaw/openclaw/issues/64571 | Failover model permanently locks session — agent config changes ignored | ## Problem
After a model failover, the session entry permanently locks to the fallback model via sticky `model`, `modelProvider`, `modelOverride`, and `providerOverride` fields. Subsequent messages **never consult the agent's configured primary model again**.
## Root Cause
1. `setSessionRuntimeModel()` sets `entry.m... | closed | completed | true | 2 | [] | [] | 2026-04-11T01:14:43Z | 2026-04-28T05:00:31Z | 2026-04-25T16:13:43Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | OpenCodeEngineer | 261,470,075 | U_kgDOD5W3ew | User | false |
openclaw/openclaw | 4,242,473,460 | I_kwDOQb6kR8783wH0 | 64,574 | https://github.com/openclaw/openclaw/issues/64574 | https://api.github.com/repos/openclaw/openclaw/issues/64574 | [Feature]: Enhanced Nextcloud Talk Support — Status Indicators & Additional Features | ## Feature Request: Enhanced Nextcloud Talk Support
### Summary
Nextcloud Talk is a valuable channel for team collaboration, but several quality-of-life features commonly found in modern chat platforms are missing. This issue aggregates feature requests to bring Nextcloud Talk closer parity with other OpenClaw channel... | closed | not_planned | true | 2 | [] | [] | 2026-04-11T01:26:52Z | 2026-04-29T04:30:57Z | 2026-04-26T04:51:08Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Wingnuut | 195,316,867 | U_kgDOC6RMgw | User | false |
openclaw/openclaw | 4,242,492,975 | I_kwDOQb6kR878304v | 64,577 | https://github.com/openclaw/openclaw/issues/64577 | https://api.github.com/repos/openclaw/openclaw/issues/64577 | Plugin slash commands silently blocked on Slack despite capabilities.nativeCommands: true | ## Description
Plugin-registered slash commands (e.g. `/lcm`, `/lossless` from `lossless-claw`) are silently excluded from Slack, even though:
1. Slack declares `capabilities.nativeCommands: true`
2. The user explicitly sets `channels.slack.commands.native: true` in config
The same commands work correctly on Telegram... | closed | completed | true | 0 | [] | [] | 2026-04-11T01:32:41Z | 2026-04-16T04:24:20Z | 2026-04-13T20:14:04Z | CONTRIBUTOR | null | 20260417T163222Z | 2026-04-17T16:32:22Z | rafaelreis-r | 57,492,577 | MDQ6VXNlcjU3NDkyNTc3 | User | false |
openclaw/openclaw | 4,242,521,499 | I_kwDOQb6kR878372b | 64,581 | https://github.com/openclaw/openclaw/issues/64581 | https://api.github.com/repos/openclaw/openclaw/issues/64581 | [Bug]: SSH sandbox backend seems to inject the wrong workspace into the system prompt. | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When using ssh as a sandbox backend the workspace that is injected into the system prompt still reflects the path on the host/gateway server.
### Steps to reproduce
1. Start Openclaw v2026.4.9 on Linux.
2. Cr... | closed | completed | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-11T01:42:06Z | 2026-04-28T05:00:31Z | 2026-04-25T16:13:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | danhenrydev | 6,625,866 | MDQ6VXNlcjY2MjU4NjY= | User | false |
openclaw/openclaw | 4,242,538,630 | I_kwDOQb6kR8784ACG | 64,583 | https://github.com/openclaw/openclaw/issues/64583 | https://api.github.com/repos/openclaw/openclaw/issues/64583 | [Bug]: All cross-agent sessions_send fails with pairing required after gateway restart in OpenClaw 2026.4.9 | ## Bug Description
**Summary:** After gateway restart in OpenClaw 2026.4.9, ALL sessions_send calls fail with `gateway closed (1008): pairing required`. This affects ALL agents including self-to-self.
**Environment:**
- OpenClaw version: 2026.4.9
- OS: Linux (Raspberry Pi OS, arm64)
- Node: v24.14.1
- Gateway bind: l... | closed | completed | true | 2 | [] | [] | 2026-04-11T01:48:32Z | 2026-04-28T05:00:30Z | 2026-04-25T16:14:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | shinfon | 92,320,216 | U_kgDOBYCx2A | User | false |
openclaw/openclaw | 4,242,559,747 | I_kwDOQb6kR8784FMD | 64,585 | https://github.com/openclaw/openclaw/issues/64585 | https://api.github.com/repos/openclaw/openclaw/issues/64585 | Auto-recover stuck gateway sessions after restart | ## Summary
Add gateway startup recovery that resumes stuck chat sessions after restart, not just outbound delivery queue retries.
## Problem
Current startup flow only calls `recoverPendingDeliveries()` in `server.impl`, which retries failed outbound queue items. Sessions can still be stuck when the latest turn never p... | open | null | false | 1 | [] | [] | 2026-04-11T01:56:10Z | 2026-05-03T10:19:19Z | null | CONTRIBUTOR | null | 20260503T120038Z | 2026-05-03T12:00:38Z | OpenCodeEngineer | 261,470,075 | U_kgDOD5W3ew | User | false |
openclaw/openclaw | 4,242,566,892 | I_kwDOQb6kR8784G7s | 64,588 | https://github.com/openclaw/openclaw/issues/64588 | https://api.github.com/repos/openclaw/openclaw/issues/64588 | Feature Request: Cross-Session Memory Sharing for Multi-Agent Collaboration | # OpenClaw 功能需求：跨会话统一记忆
## 需求标题
**Cross-Session Memory Sharing / 跨会话状态与记忆共享**
## 问题描述
当前 OpenClaw 的群聊（Group Chat）和私聊（DM）是两个完全独立的会话（Session）。每个会话有独立的上下文（context window），不自动共享以下内容：
1. **对话历史**：在群聊中聊过的内容，私聊会话不记得
2. **Agent 状态**：某个 agent "在不在群里"，私聊会话无法判断
3. **工作状态**：当前任务进展、pending 文件状态等
4. **记忆积累**：每次会话结束，记忆不传递给下一个新会话
... | closed | not_planned | true | 1 | [] | [] | 2026-04-11T01:58:57Z | 2026-04-28T05:00:29Z | 2026-04-25T16:14:09Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jokuke99 | 267,636,413 | U_kgDOD_POvQ | User | false |
openclaw/openclaw | 4,242,584,317 | I_kwDOQb6kR8784LL9 | 64,590 | https://github.com/openclaw/openclaw/issues/64590 | https://api.github.com/repos/openclaw/openclaw/issues/64590 | memory-core: narrative generation failed - missing required property 'idempotencyKey' | ## Bug Description
The memory-core plugin fails to generate narratives and logs warnings every 30 minutes:
```
memory-core: narrative generation failed for light phase: invalid agent params: must have required property 'idempotencyKey'
memory-core: narrative generation failed for rem phase: invalid agent params: must... | closed | completed | true | 4 | [] | [] | 2026-04-11T02:03:26Z | 2026-04-15T04:19:56Z | 2026-04-12T16:46:25Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | inteai | 20,512,753 | MDQ6VXNlcjIwNTEyNzUz | User | false |
openclaw/openclaw | 4,242,618,892 | I_kwDOQb6kR8784ToM | 64,593 | https://github.com/openclaw/openclaw/issues/64593 | https://api.github.com/repos/openclaw/openclaw/issues/64593 | Chromium child processes leak as zombies on normal exit (not just startup) → containers hit PID limit | ## Summary
In v2026.4.1 (and likely other recent versions), the browser control service spawns chromium child processes but does not reap them after normal exit. Over days, containers slowly accumulate thousands of `<defunct>` processes until they hit the PID limit, at which point `spawn sh` returns `EAGAIN` and all t... | open | null | false | 1 | [] | [] | 2026-04-11T02:14:45Z | 2026-05-03T09:48:58Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | theandyroberts | 81,009,938 | MDQ6VXNlcjgxMDA5OTM4 | User | false |
openclaw/openclaw | 4,242,625,923 | I_kwDOQb6kR8784VWD | 64,595 | https://github.com/openclaw/openclaw/issues/64595 | https://api.github.com/repos/openclaw/openclaw/issues/64595 | memory-core dreaming: system event token not consumed, causes repeated hourly execution | ## Bug Description
The dreaming promotion system event (`__openclaw_memory_core_short_term_promotion_dream__`) is not consumed after execution, causing dreaming to re-trigger on **every subsequent heartbeat** instead of running once per cron schedule.
## Expected Behavior
- Cron fires at configured time (e.g. `0 3 *... | closed | completed | true | 2 | [] | [] | 2026-04-11T02:17:07Z | 2026-04-15T04:19:49Z | 2026-04-12T16:47:28Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | richardmqq | 14,178,267 | MDQ6VXNlcjE0MTc4MjY3 | User | false |
openclaw/openclaw | 4,242,654,238 | I_kwDOQb6kR8784cQe | 64,599 | https://github.com/openclaw/openclaw/issues/64599 | https://api.github.com/repos/openclaw/openclaw/issues/64599 | [Bug]: Light Sleep phase produces all candidates with confidence=0.00 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Light Sleep phase in Dreaming produces all candidates with confidence=0.00, preventing any meaningful memory promotion. REM Sleep phase works correctly with normal confidence values (~0.70).
### Steps to reproduce
1. Enabl... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-04-11T02:27:23Z | 2026-04-15T04:19:37Z | 2026-04-12T17:31:06Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | dsdsddd1 | 130,032,607 | U_kgDOB8Aj3w | User | false |
openclaw/openclaw | 4,242,669,779 | I_kwDOQb6kR8784gDT | 64,602 | https://github.com/openclaw/openclaw/issues/64602 | https://api.github.com/repos/openclaw/openclaw/issues/64602 | Dreaming memory-core narrative generation fails with idempotencyKey required error | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Dreaming (memory-core plugin) fails at narrative generation step with error: invalid agent params: must have required property 'idempotencyKey'
### Steps to reproduce
1. Enable memory-core plugin with Dreaming feature
2. W... | closed | completed | true | 4 | [
"bug",
"regression"
] | [] | 2026-04-11T02:33:29Z | 2026-04-15T04:19:55Z | 2026-04-12T16:46:27Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | hxxxio | 98,248,925 | U_kgDOBdso3Q | User | false |
openclaw/openclaw | 4,242,682,467 | I_kwDOQb6kR8784jJj | 64,603 | https://github.com/openclaw/openclaw/issues/64603 | https://api.github.com/repos/openclaw/openclaw/issues/64603 | [Bug]: punycode deprecation warning still reproduces on 2026.4.10 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
On a fully updated live source checkout of OpenClaw `2026.4.10`, CLI paths such as `gateway status` and `doctor` still emit Node `DEP0040` because a transitive dependency chain still imports the deprecated buil... | closed | completed | true | 3 | [] | [] | 2026-04-11T02:37:07Z | 2026-04-28T05:00:28Z | 2026-04-25T16:14:23Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Jackten | 2,895,479 | MDQ6VXNlcjI4OTU0Nzk= | User | false |
openclaw/openclaw | 4,242,766,269 | I_kwDOQb6kR87843m9 | 64,606 | https://github.com/openclaw/openclaw/issues/64606 | https://api.github.com/repos/openclaw/openclaw/issues/64606 | Duplicate Telegram message sent after provider timeout + retry | ## Bug Description
When the primary provider times out and the agent retries (or failover triggers a retry), the **same reply content is sent to Telegram twice**, resulting in the user seeing a duplicate message.
## Expected Behavior
- Provider times out → retry/failover → agent generates response → **one message se... | closed | completed | true | 1 | [] | [] | 2026-04-11T03:01:25Z | 2026-04-29T04:32:28Z | 2026-04-27T02:26:30Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | richardmqq | 14,178,267 | MDQ6VXNlcjE0MTc4MjY3 | User | false |
openclaw/openclaw | 4,242,771,656 | I_kwDOQb6kR878447I | 64,607 | https://github.com/openclaw/openclaw/issues/64607 | https://api.github.com/repos/openclaw/openclaw/issues/64607 | Feature Request: Inline media display in chat (images, audio, video, web previews) | ## Feature Request: Inline media display in chat
### Problem Statement
Currently, OpenClaw does not support inline display of generated or received media in chat responses. When users generate images, videos, music, or receive web content, they must manually use \\`open\\` commands to view them externally.
This crea... | open | null | false | 3 | [] | [] | 2026-04-11T03:03:11Z | 2026-05-02T11:08:37Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | DearPeter | 1,520,265 | MDQ6VXNlcjE1MjAyNjU= | User | false |
openclaw/openclaw | 4,242,786,289 | I_kwDOQb6kR87848fx | 64,609 | https://github.com/openclaw/openclaw/issues/64609 | https://api.github.com/repos/openclaw/openclaw/issues/64609 | Telegram multi-account: group/topic systemPrompt silently ignored due to inconsistent config resolution | ## Bug Description
In multi-account Telegram setups (2+ accounts configured under `channels.telegram.accounts`), the `systemPrompt` field in `channels.telegram.groups.<id>` and `channels.telegram.groups.<id>.topics.<threadId>` is **silently ignored**. The model never receives the configured prompt.
Meanwhile, `requir... | open | null | false | 1 | [] | [] | 2026-04-11T03:10:19Z | 2026-05-03T01:22:36Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | briqt | 131,652,747 | U_kgDOB9jciw | User | false |
openclaw/openclaw | 4,242,848,231 | I_kwDOQb6kR8785Lnn | 64,612 | https://github.com/openclaw/openclaw/issues/64612 | https://api.github.com/repos/openclaw/openclaw/issues/64612 | Feature request: per-agent memory.backend override | # Feature request: per-agent `memory.backend` override
## Summary
OpenClaw 2026.4.8's config schema supports a root-level `memory.backend`
(`"builtin"` or `"qmd"`) but does NOT allow overriding it per agent. This
blocks safe, phased rollouts of `qmd`-backed memory onto multi-agent
installations — any cutover is neces... | closed | not_planned | true | 2 | [] | [] | 2026-04-11T03:27:27Z | 2026-04-30T04:41:56Z | 2026-04-26T04:51:12Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | hyx4992 | 55,764,514 | MDQ6VXNlcjU1NzY0NTE0 | User | false |
openclaw/openclaw | 4,242,876,091 | I_kwDOQb6kR8785Sa7 | 64,613 | https://github.com/openclaw/openclaw/issues/64613 | https://api.github.com/repos/openclaw/openclaw/issues/64613 | chat.history leaks system-level memory injection blocks to WebChat UI | ## Describe the bug
The WebChat UI (macOS native app, Control UI) renders system-level memory injection blocks that should be invisible to users. These blocks contain internal prompts like:
```
## Memory system — ACTION REQUIRED
Auto-recall found no relevant results for a long query. You MUST call `memory_search` no... | open | null | false | 2 | [] | [] | 2026-04-11T03:35:20Z | 2026-05-03T09:48:59Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | JackChen201535 | 11,334,408 | MDQ6VXNlcjExMzM0NDA4 | User | false |
openclaw/openclaw | 4,242,912,756 | I_kwDOQb6kR8785bX0 | 64,618 | https://github.com/openclaw/openclaw/issues/64618 | https://api.github.com/repos/openclaw/openclaw/issues/64618 | [Bug]: Automatic Dreaming cron job not registered in Gateway startup (registerShortTermPromotionDreaming exported but never imported) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The automatic Dreaming cron job (which should run at configured times like 0 3 * * *) never gets registered during Gateway startup. Investigation reveals that registerShortTermPromotionDreaming(api) is defined ... | closed | not_planned | true | 1 | [
"bug",
"invalid",
"bug:behavior"
] | [] | 2026-04-11T03:44:12Z | 2026-04-15T04:19:59Z | 2026-04-12T16:46:05Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | jackhateme | 128,401,579 | U_kgDOB6dAqw | User | false |
openclaw/openclaw | 4,242,913,443 | I_kwDOQb6kR8785bij | 64,619 | https://github.com/openclaw/openclaw/issues/64619 | https://api.github.com/repos/openclaw/openclaw/issues/64619 | [Bug] Control UI Dreams page: Chinese text truncated and elements overlapping (v2026.4.9) | ## Bug Description
In the Control UI Dreams page, Chinese text is truncated mid-character and UI elements overlap with each other.
## Steps to Reproduce
1. Open Control UI and navigate to the Dreams page
2. Observe truncated Chinese labels (e.g. "Scene 运" cut mid-character)
3. Observe "no grounded scenes" message ov... | closed | completed | true | 2 | [] | [] | 2026-04-11T03:44:25Z | 2026-04-28T05:00:28Z | 2026-04-25T16:14:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ryswork1993 | 44,529,247 | MDQ6VXNlcjQ0NTI5MjQ3 | User | false |
openclaw/openclaw | 4,242,916,331 | I_kwDOQb6kR8785cPr | 64,620 | https://github.com/openclaw/openclaw/issues/64620 | https://api.github.com/repos/openclaw/openclaw/issues/64620 | Memory Wiki: slugifyWikiSegment strips all non-ASCII characters, breaking CJK titles | # GitHub Issue: Memory Wiki slug Unicode support
**Repo**: https://github.com/openclaw/openclaw/issues
**Title**: `Memory Wiki: slugifyWikiSegment strips all non-ASCII characters, breaking CJK titles`
---
## Summary
`slugifyWikiSegment()` in the Memory Wiki plugin uses `/[^a-z0-9]+/g` to generate filenames from pag... | closed | completed | true | 1 | [] | [] | 2026-04-11T03:45:19Z | 2026-04-15T04:19:43Z | 2026-04-12T16:54:42Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | fuyizheng3120 | 268,386,865 | U_kgDOD_9CMQ | User | false |
openclaw/openclaw | 4,242,955,610 | I_kwDOQb6kR8785l1a | 64,621 | https://github.com/openclaw/openclaw/issues/64621 | https://api.github.com/repos/openclaw/openclaw/issues/64621 | WeChat channel: user image attachment merged into heartbeat turn | ## Description
When a user sends an image via the WeChat (openclaw-weixin) channel at roughly the same time a heartbeat turn is queued, the image gets attached to the heartbeat turn instead of being processed as a separate user message.
## Steps to reproduce
1. Configure an agent with heartbeat enabled (`every: "2h"... | closed | not_planned | true | 1 | [] | [] | 2026-04-11T03:59:24Z | 2026-04-28T05:00:27Z | 2026-04-25T16:14:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | TwinsLee | 1,460,694 | MDQ6VXNlcjE0NjA2OTQ= | User | false |
openclaw/openclaw | 4,242,965,272 | I_kwDOQb6kR8785oMY | 64,624 | https://github.com/openclaw/openclaw/issues/64624 | https://api.github.com/repos/openclaw/openclaw/issues/64624 | [Feature]: Suppress or throttle transient channel connection status system events | ### Summary
### Problem
When a channel (e.g. WhatsApp) briefly disconnects and reconnects, the gateway injects system messages into the active session:
### Problem to solve
These are useful for debugging, but during normal operation they're noise. A 6-second reconnect isn't something the user or agent needs to know... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-11T04:02:31Z | 2026-05-03T09:49:00Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | bandsight | 240,598,652 | U_kgDODlc-fA | User | false |
openclaw/openclaw | 4,242,973,170 | I_kwDOQb6kR8785qHy | 64,625 | https://github.com/openclaw/openclaw/issues/64625 | https://api.github.com/repos/openclaw/openclaw/issues/64625 | Slack upload-file drops agent-scoped mediaLocalRoots via custom invoke path | ## Summary
Agent file uploads via the `message` tool with `action=upload-file` fail when the file is inside an agent's configured workspace directory:
```
Local media path is not under an allowed directory: <agent-workspace>/renders/file.wav
```
The agent's workspace is correctly configured and the file exists. Othe... | closed | completed | false | 2 | [] | [] | 2026-04-11T04:05:21Z | 2026-05-03T09:14:51Z | 2026-05-03T09:14:51Z | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | benpchandler | 2,336,465 | MDQ6VXNlcjIzMzY0NjU= | User | false |
openclaw/openclaw | 4,243,012,025 | I_kwDOQb6kR8785zm5 | 64,628 | https://github.com/openclaw/openclaw/issues/64628 | https://api.github.com/repos/openclaw/openclaw/issues/64628 | Bug: image tool reports 'No API provider registered for api: ollama' despite OLLAMA_API_KEY and correct config | ## Summary
The `image` tool fails with `No API provider registered for api: ollama` even when:
1. Ollama provider is correctly configured in `models.providers.ollama`
2. `OLLAMA_API_KEY` env var is set to `"ollama-local"`
3. `agents.defaults.imageModel` is set to `ollama/gemma4:e4b`
4. Ollama service is running and re... | closed | completed | true | 2 | [] | [] | 2026-04-11T04:17:57Z | 2026-04-28T05:00:26Z | 2026-04-25T16:14:49Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Deepwater1000 | 126,386,305 | U_kgDOB4iAgQ | User | false |
openclaw/openclaw | 4,243,055,715 | I_kwDOQb6kR8785-Rj | 64,633 | https://github.com/openclaw/openclaw/issues/64633 | https://api.github.com/repos/openclaw/openclaw/issues/64633 | [Bug]: Short Gibberish Speak even with request in Soul.MD and User.MD to not speak that way | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When having conversation with OpenClaw, I can ask it to speak in paragraphs that are easily readable and it does just that for a reply or so and then immediately goes back to this short speak gibberish that is ... | open | null | false | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-11T04:30:08Z | 2026-05-03T09:49:02Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | B1GB3NTX | 154,041,410 | U_kgDOCS58Qg | User | false |
openclaw/openclaw | 4,243,063,202 | I_kwDOQb6kR8786AGi | 64,634 | https://github.com/openclaw/openclaw/issues/64634 | https://api.github.com/repos/openclaw/openclaw/issues/64634 | Bug: cron scheduler calculates incorrect nextRunAtMs on job creation | ## Description
When creating a new cron job via `openclaw cron add`, the scheduler calculates an incorrect `nextRunAtMs` value. After the first successful execution, subsequent nextRun calculations are correct.
## Steps to Reproduce
1. Create a cron job with expression `15 * * * *` (every hour at minute 15)
2. Obser... | closed | completed | true | 2 | [] | [] | 2026-04-11T04:32:04Z | 2026-04-28T05:00:25Z | 2026-04-25T16:14:56Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | hunterlarcuad | 193,805,647 | U_kgDOC409Tw | User | false |
openclaw/openclaw | 4,243,081,707 | I_kwDOQb6kR8786Enr | 64,636 | https://github.com/openclaw/openclaw/issues/64636 | https://api.github.com/repos/openclaw/openclaw/issues/64636 | [Bug]: Version 2026.4.9 ignore the system environment proxy variables (HTTP_PROXY, etc.) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
after upgrading from version2026.3.28 to version 2026.4.9, Openclaw ignore the system environment proxy variables (HTTP_PROXY, etc.)
### Steps to reproduce
after upgrading from version2026.3.28 to version 2026.4.9
1. set t... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-11T04:39:18Z | 2026-04-28T05:00:24Z | 2026-04-25T16:15:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | TransForum | 175,947,650 | U_kgDOCny_gg | User | false |
openclaw/openclaw | 4,243,086,884 | I_kwDOQb6kR8786F4k | 64,639 | https://github.com/openclaw/openclaw/issues/64639 | https://api.github.com/repos/openclaw/openclaw/issues/64639 | Feature: allow disabling specific skill source directories (e.g. agents-skills-personal) | ## Summary
Currently, skill loading has 6 hardcoded sources (bundled, managed, extra, personal, project, workspace). While `extraDirs` is already configurable via `skills.load.extraDirs`, the `~/.agents/skills/` directory (source: `agents-skills-personal`) is always scanned with no way to disable it.
## Use Case
I m... | closed | not_planned | true | 3 | [] | [] | 2026-04-11T04:40:58Z | 2026-05-02T04:27:17Z | 2026-04-29T20:14:48Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | fuyizheng3120 | 268,386,865 | U_kgDOD_9CMQ | User | false |
openclaw/openclaw | 4,243,115,297 | I_kwDOQb6kR8786M0h | 64,642 | https://github.com/openclaw/openclaw/issues/64642 | https://api.github.com/repos/openclaw/openclaw/issues/64642 | [Bug]: Dreaming recall store never populated when daily notes are in memory/ subdirectory | ## Bug Description
`isShortTermMemoryPath()` in `short-term-promotion` fails to match daily memory files stored in `memory/daily/YYYY-MM-DD.md` because the regex patterns only match `memory/YYYY-MM-DD.md` (no subdirectory).
This means the recall store (`memory/.dreams/short-term-recall.json`) is **never written**, an... | closed | completed | true | 0 | [] | [] | 2026-04-11T04:50:00Z | 2026-04-15T04:19:31Z | 2026-04-12T18:41:01Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | orangeruan128 | 257,684,321 | U_kgDOD1vzYQ | User | false |
openclaw/openclaw | 4,243,128,769 | I_kwDOQb6kR8786QHB | 64,644 | https://github.com/openclaw/openclaw/issues/64644 | https://api.github.com/repos/openclaw/openclaw/issues/64644 | [Bug] MiniMax M2.7 图像理解功能调用失败 - chat content is empty (2013) | ## 问题描述
使用 MiniMax-M2.7 模型的图像理解功能时，调用失败。
## 错误信息
```
Image model failed (minimax/MiniMax-M2.7): 400 {"type":"error","error":{"type":"invalid_request_error","message":"invalid params, chat content is empty (2013)"},"request_id":"..."}
```
## 环境信息
- OpenClaw 版本：2026.4.9
- 模型：minimax/MiniMax-M2.7
- 操作系统：Ubuntu 24.04 ... | closed | completed | true | 1 | [] | [] | 2026-04-11T04:54:51Z | 2026-04-28T05:00:24Z | 2026-04-25T16:15:09Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zhangxiaobing0102 | 143,339,660 | U_kgDOCIswjA | User | false |
openclaw/openclaw | 4,243,151,052 | I_kwDOQb6kR8786VjM | 64,651 | https://github.com/openclaw/openclaw/issues/64651 | https://api.github.com/repos/openclaw/openclaw/issues/64651 | Control UI on iOS auto-zooms when focusing chat/input fields | ## Summary
On iPhone/iOS Safari, the OpenClaw Control UI chat input and other text inputs render below the iOS 16px threshold, so Safari auto-zooms the page every time the user focuses a field.
## Why this matters
This makes the Control UI feel broken as an iOS home screen web app. Typing in the chat box repeatedly zo... | open | null | false | 3 | [] | [
"vincentkoc"
] | 2026-04-11T05:02:21Z | 2026-05-02T13:07:58Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | f0rmwk | 39,070,642 | MDQ6VXNlcjM5MDcwNjQy | User | false |
openclaw/openclaw | 4,243,185,080 | I_kwDOQb6kR8786d24 | 64,655 | https://github.com/openclaw/openclaw/issues/64655 | https://api.github.com/repos/openclaw/openclaw/issues/64655 | [Feature Request] agent 主动发送图片给用户的能力 | ## 问题描述
用户希望 agent 能够主动发送图片给他，但目前无法实现。
## 现状
- agent 工具集中有 `image_generate`（生成图片）和 `image`（分析图片）
- 但没有「发送已有图片给用户」的工具
- `webchat` 频道不支持图片消息的主动推送
## 期望能力
agent 能够：
1. 从网络下载图片到本地 workspace
2. 将图片发送给当前对话的用户
3. 类似「以图会友」的场景：用户说"找一张唐朝凉亭" → agent 找到图片 → 直接把图片发给用户
## 使用场景
- 素材查找：设计师找参考图后直接发送
- 结果展示：生成的内容直接呈现
- 文档辅助：截图/图表... | closed | completed | true | 3 | [] | [] | 2026-04-11T05:16:25Z | 2026-04-28T05:00:23Z | 2026-04-25T16:15:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | zhangxiaobing0102 | 143,339,660 | U_kgDOCIswjA | User | false |
openclaw/openclaw | 4,243,200,461 | I_kwDOQb6kR8786hnN | 64,658 | https://github.com/openclaw/openclaw/issues/64658 | https://api.github.com/repos/openclaw/openclaw/issues/64658 | Feature request: Disable sub-agent auto-announce to channel | ## Problem
When spawning sub-agents with `sessions_spawn` from a channel-bound session (e.g. iMessage), sub-agents inherit the parent's delivery context and auto-announce completion results directly to that channel.
This creates unwanted message spam when:
- The parent agent wants to filter/review results before deliv... | open | null | false | 3 | [] | [] | 2026-04-11T05:23:47Z | 2026-05-03T10:41:06Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | GR3ATB0B | 71,197,497 | MDQ6VXNlcjcxMTk3NDk3 | User | false |
openclaw/openclaw | 4,243,269,839 | I_kwDOQb6kR8786yjP | 64,664 | https://github.com/openclaw/openclaw/issues/64664 | https://api.github.com/repos/openclaw/openclaw/issues/64664 | Approvals lost on gateway restart — stale callback buttons show confusing error | ## Problem
When the OpenClaw gateway restarts, any in-flight approval IDs are lost from memory. After the restart, tapping Approve or Deny on stale Telegram approval messages returns:
> unknown or expired approval id
and the UI shows:
> ❌ Failed to submit approval
This is confusing because the user has no indicati... | open | null | false | 3 | [] | [] | 2026-04-11T05:50:49Z | 2026-05-03T08:24:55Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | hilaryduffrules-hash | 258,473,113 | U_kgDOD2f8mQ | User | false |
openclaw/openclaw | 4,243,270,176 | I_kwDOQb6kR8786yog | 64,665 | https://github.com/openclaw/openclaw/issues/64665 | https://api.github.com/repos/openclaw/openclaw/issues/64665 | [Bug]: WhatsApp outbound media silently dropped — hasMedia: false — CLI and model-initiated sends affected (v2026.4.8–4.9) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
All WhatsApp outbound media fails silently regardless of trigger path — CLI --media flag and model-initiated sends are both affected. The model may attempt and "confirm" media delivery repeatedly without any indication of fa... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-11T05:50:58Z | 2026-04-28T05:00:22Z | 2026-04-25T16:15:23Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | AllenKWPaley | 111,218,161 | U_kgDOBqEN8Q | User | false |
openclaw/openclaw | 4,243,278,256 | I_kwDOQb6kR87860mw | 64,667 | https://github.com/openclaw/openclaw/issues/64667 | https://api.github.com/repos/openclaw/openclaw/issues/64667 | [Bug]: OpenAI Codex OAuth invalid request scope | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
During onboard when choosing OpenAI Codex as a Model/auth provider the OAuth URL contains invalid scopes (model.request+api.responses.write).
### Steps to reproduce
1. Start the onboarding process on version 2026.4.10.
2. ... | closed | completed | true | 6 | [
"bug",
"regression"
] | [
"vincentkoc"
] | 2026-04-11T05:54:23Z | 2026-04-15T04:20:11Z | 2026-04-12T15:41:39Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | faustolazala | 2,306,511 | MDQ6VXNlcjIzMDY1MTE= | User | false |
openclaw/openclaw | 4,243,278,416 | I_kwDOQb6kR87860pQ | 64,668 | https://github.com/openclaw/openclaw/issues/64668 | https://api.github.com/repos/openclaw/openclaw/issues/64668 | [Bug]: Bridge import for memory wiki not working | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
OpenClaw `memory-wiki` bridge mode reports zero exported artifacts with `memory-core` on OpenClaw 2026.4.9 as well as 2026.4.10, even though the active memory workspace is resolved correctly and contains `MEMOR... | closed | completed | true | 2 | [
"bug"
] | [
"vincentkoc"
] | 2026-04-11T05:54:27Z | 2026-04-15T04:20:14Z | 2026-04-12T15:31:59Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | BaneCapital | 251,144,385 | U_kgDODvgowQ | User | false |
openclaw/openclaw | 4,243,300,757 | I_kwDOQb6kR87866GV | 64,670 | https://github.com/openclaw/openclaw/issues/64670 | https://api.github.com/repos/openclaw/openclaw/issues/64670 | [Bug]: /btw Command Failed (400 Error) & ESC Key Not Working | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
I encountered two critical issues during usage:
The /btw command fails to execute and throws a 400 [] is too short - 'tools' error.
The ESC key is completely non-functional and cannot perform normal cancel/exit... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-11T06:03:38Z | 2026-04-28T05:00:21Z | 2026-04-25T16:15:29Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | JIRBOY | 49,422,297 | MDQ6VXNlcjQ5NDIyMjk3 | User | false |
openclaw/openclaw | 4,243,314,871 | I_kwDOQb6kR87869i3 | 64,671 | https://github.com/openclaw/openclaw/issues/64671 | https://api.github.com/repos/openclaw/openclaw/issues/64671 | Add `openclaw skills list` CLI command | ### Summary
### Is your feature request related to a problem? Please describe.
The OpenClaw web UI can display the list of installed skills, but there is
no equivalent CLI command. Users who prefer command-line workflows have no
way to list available skills.
### Problem to solve
### Describe the solution you'd like
... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-04-11T06:08:57Z | 2026-04-28T05:00:20Z | 2026-04-25T16:15:36Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | ywl125 | 38,369,097 | MDQ6VXNlcjM4MzY5MDk3 | User | false |
openclaw/openclaw | 4,243,328,184 | I_kwDOQb6kR8787Ay4 | 64,672 | https://github.com/openclaw/openclaw/issues/64672 | https://api.github.com/repos/openclaw/openclaw/issues/64672 | qwen3.6-plus unavailable on Qwen Coding Plan Pro subscription | # Issue: qwen3.6-plus is unavailable on Qwen Coding Plan Pro subscription
## Summary
OpenClaw hardcoded filters out `qwen3.6-plus` model for Coding Plan endpoint, but Alibaba Cloud officially supports `qwen3.6-plus` for Coding Plan Pro subscription since April 2026.
## Problem Description
When using Coding Plan Pro... | closed | not_planned | true | 2 | [] | [] | 2026-04-11T06:12:17Z | 2026-04-29T04:32:27Z | 2026-04-27T02:26:47Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Nick-Liu1989 | 170,501,677 | U_kgDOCimmLQ | User | false |
openclaw/openclaw | 4,243,394,289 | I_kwDOQb6kR8787Q7x | 64,676 | https://github.com/openclaw/openclaw/issues/64676 | https://api.github.com/repos/openclaw/openclaw/issues/64676 | [Bug]: Gateway Configuration Persistence Issue | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
OpenClaw gateway persistently overwrites agent configuration files (`models.json`) with default/cached values after restart, making user edits impossible to maintain.
### Steps to reproduce
## Steps to reprod... | open | null | false | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-11T06:36:34Z | 2026-05-03T10:41:07Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | alfarisw | 57,335,659 | MDQ6VXNlcjU3MzM1NjU5 | User | false |
openclaw/openclaw | 4,243,421,618 | I_kwDOQb6kR8787Xmy | 64,680 | https://github.com/openclaw/openclaw/issues/64680 | https://api.github.com/repos/openclaw/openclaw/issues/64680 | Feature Request: Built-in mandatory auto-write memory (without relying on agent self-discipline) | ## Problem
Currently the only built-in memory write mechanism is a 'silent reminder' before compaction. The agent can ignore it, meaning memory write depends on the agent's diligence — not a reliable system behavior.
There is no equivalent to `autoCapture` that automatically extracts and stores important information ... | closed | not_planned | true | 1 | [] | [] | 2026-04-11T06:51:45Z | 2026-04-28T05:00:20Z | 2026-04-25T16:15:50Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Lrs-GreenBottle | 253,340,193 | U_kgDODxmqIQ | User | false |
openclaw/openclaw | 4,243,446,966 | I_kwDOQb6kR8787dy2 | 64,686 | https://github.com/openclaw/openclaw/issues/64686 | https://api.github.com/repos/openclaw/openclaw/issues/64686 | Bug: false ouraclaw provenance warnings in openclaw status on 2026.4.10 | # ouraclaw false provenance warning in `openclaw status` on 2026.4.10
Date: 2026-04-11
Version: OpenClaw 2026.4.10 (44e5b62)
## Summary
`openclaw status` reports plugin provenance warnings for `ouraclaw` even though the plugin is explicitly allowlisted, installed, and loaded successfully.
## Warning text
```text
[p... | closed | completed | true | 1 | [] | [] | 2026-04-11T07:01:04Z | 2026-04-28T05:00:19Z | 2026-04-25T16:15:57Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | tasi333 | 7,864,516 | MDQ6VXNlcjc4NjQ1MTY= | User | false |
openclaw/openclaw | 4,243,447,713 | I_kwDOQb6kR8787d-h | 64,687 | https://github.com/openclaw/openclaw/issues/64687 | https://api.github.com/repos/openclaw/openclaw/issues/64687 | [Bug]: OAuth for openai-codex broken? | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
I'm provided with a link to click, when i click it, I'm brought here:
http://localhost:1455/auth/callback?error=invalid_scope&error_description=The+requested+scope+is+invalid%2C+unknown%2C+or+malformed.+The+OAuth+2.0+Client... | closed | completed | true | 10 | [
"bug",
"regression"
] | [] | 2026-04-11T07:01:18Z | 2026-04-14T04:20:23Z | 2026-04-11T15:58:37Z | NONE | null | 20260417T163222Z | 2026-04-17T16:32:22Z | awknode | 64,115,053 | MDQ6VXNlcjY0MTE1MDUz | User | false |
openclaw/openclaw | 4,243,485,853 | I_kwDOQb6kR8787nSd | 64,690 | https://github.com/openclaw/openclaw/issues/64690 | https://api.github.com/repos/openclaw/openclaw/issues/64690 | Teams channel: implement sendPayload for interactive approval cards | ## Problem
Approval prompts (plugin approval, exec approval) generate interactive button descriptors via the delivery pipeline. The pipeline checks whether the channel's outbound adapter implements `sendPayload` and whether the payload contains interactive content. If both conditions are met, it calls `sendPayload` to... | closed | not_planned | true | 1 | [] | [] | 2026-04-11T07:11:30Z | 2026-04-30T04:41:56Z | 2026-04-26T04:51:16Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | kwizzlek | 215,732,384 | U_kgDODNvQoA | User | false |
openclaw/openclaw | 4,243,510,639 | I_kwDOQb6kR8787tVv | 64,695 | https://github.com/openclaw/openclaw/issues/64695 | https://api.github.com/repos/openclaw/openclaw/issues/64695 | 🐛 [CRASH] Gateway crashes every ~6-7 minutes — node:sqlite missing in Homebrew Node.js | ## Environment
- **OpenClaw version:** 2026.4.10 (44e5b62)
- **Node.js version:** v23.3.0 (Homebrew bottle on macOS arm64)
- **Platform:** macOS Darwin 25.3.0
- **Installation:** Homebrew (`/opt/homebrew/lib/node_modules/openclaw`)
## Problem
Every ~6-7 minutes, `sweepTaskRegistry()` triggers task registry maintenan... | open | null | false | 2 | [] | [] | 2026-04-11T07:20:26Z | 2026-05-03T08:24:57Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | wmaa0002 | 31,246,068 | MDQ6VXNlcjMxMjQ2MDY4 | User | false |
openclaw/openclaw | 4,243,538,635 | I_kwDOQb6kR87870LL | 64,696 | https://github.com/openclaw/openclaw/issues/64696 | https://api.github.com/repos/openclaw/openclaw/issues/64696 | memory-wiki: bridge import and lint mishandle relative links, reply tags, and malformed cached source pages | # Upstream issue draft: memory-wiki bridge import and lint durability regressions
## Summary
The OpenClaw memory-wiki stack currently has a cluster of related durability bugs around compiled wiki source pages and report links.
In a live bridge-mode vault, these showed up as:
1. valid relative links in `reports/stal... | open | null | false | 4 | [] | [] | 2026-04-11T07:27:14Z | 2026-05-03T06:07:20Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | reghar-bot | 264,667,955 | U_kgDOD8aDMw | User | false |
openclaw/openclaw | 4,243,566,680 | I_kwDOQb6kR87877BY | 64,698 | https://github.com/openclaw/openclaw/issues/64698 | https://api.github.com/repos/openclaw/openclaw/issues/64698 | [Bug] Telegram requireMention: false ignored for group messages | ## Bug type
Behavior bug
## Summary
Setting `requireMention: false` in Telegram group config is completely ignored. All group messages are skipped with `reason: "no-mention"` regardless of the setting. This affects group-based multi-bot setups where bots need to see all messages for full conversation context.
## Envi... | closed | completed | true | 1 | [] | [] | 2026-04-11T07:33:56Z | 2026-04-30T04:41:55Z | 2026-04-26T04:51:20Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | concongo | 8,966,340 | MDQ6VXNlcjg5NjYzNDA= | User | false |
openclaw/openclaw | 4,243,568,494 | I_kwDOQb6kR87877du | 64,699 | https://github.com/openclaw/openclaw/issues/64699 | https://api.github.com/repos/openclaw/openclaw/issues/64699 | [Bug]: sessions_send unexpectedly injects label, causing mutual-exclusion error with sessionKey | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When calling sessions_send, the effective invocation appears to include label even when the caller intends to send only by sessionKey.
调用 sessions_send 时，实际执行链路中疑似会被自动补入 label，即使调用方本意只传 sessionKey。这会导致本来合法的调用触... | open | null | false | 5 | [
"bug",
"bug:behavior"
] | [] | 2026-04-11T07:34:27Z | 2026-05-03T08:24:59Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | sunxq1017-hash | 266,310,142 | U_kgDOD9-R_g | User | false |
openclaw/openclaw | 4,243,613,702 | I_kwDOQb6kR8788GgG | 64,704 | https://github.com/openclaw/openclaw/issues/64704 | https://api.github.com/repos/openclaw/openclaw/issues/64704 | [Bug]: active-memory plugin package not found on npm or ClawHub (2026.4.10) | ## Bug Description
The changelog for OpenClaw 2026.4.10 states:
> Memory/Active Memory: add a new optional Active Memory plugin that gives OpenClaw a dedicated memory sub-agent right before the main reply...
However, the @openclaw/active-memory plugin package cannot be installed — it does not exist on npm or ClawHub... | closed | completed | true | 1 | [] | [] | 2026-04-11T07:44:41Z | 2026-04-29T04:41:29Z | 2026-04-26T08:21:20Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | wangzihe843-hash | 266,487,497 | U_kgDOD-JGyQ | User | false |
openclaw/openclaw | 4,243,615,451 | I_kwDOQb6kR8788G7b | 64,705 | https://github.com/openclaw/openclaw/issues/64705 | https://api.github.com/repos/openclaw/openclaw/issues/64705 | Bug: Embedded runs resolve model=default for openai-codex instead of explicit configured model | Observed in gateway logs on 10 Apr 2026 14:00 UTC.
Embedded runId: `1620a4ae-44a9-4248-a7c4-e736b7b27426`.
Log line shows: `model=default provider=openai-codex`.
Failure: `The 'default' model is not supported when using Codex with a ChatGPT account.`
Followed by retry message: `unsupported thinking level for openai... | closed | not_planned | true | 3 | [] | [] | 2026-04-11T07:45:07Z | 2026-04-28T05:00:18Z | 2026-04-25T16:16:03Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | castertond | 174,272,972 | U_kgDOCmMxzA | User | false |
openclaw/openclaw | 4,243,628,956 | I_kwDOQb6kR8788KOc | 64,706 | https://github.com/openclaw/openclaw/issues/64706 | https://api.github.com/repos/openclaw/openclaw/issues/64706 | [Bug]: RangeError when sending large images via WebChat | ## Summary
WebChat throws RangeError: Maximum call stack size exceeded when sending high-resolution images (e.g., 4284×5712 pixels from QQ chat), making image sharing impossible.
## Steps to Reproduce
1. Open OpenClaw WebChat Control UI
2. Send an image with high resolution (4284×5712 pixels, ~4MB) from QQ chat
3. O... | closed | completed | true | 1 | [] | [] | 2026-04-11T07:48:50Z | 2026-04-28T05:00:17Z | 2026-04-25T16:16:09Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | mtdcy | 16,152,368 | MDQ6VXNlcjE2MTUyMzY4 | User | false |
openclaw/openclaw | 4,243,629,640 | I_kwDOQb6kR8788KZI | 64,707 | https://github.com/openclaw/openclaw/issues/64707 | https://api.github.com/repos/openclaw/openclaw/issues/64707 | Feishu: default account auto-created on gateway restart with not configured error | ## Bug Description
After upgrading to OpenClaw 2026.4.10, every gateway restart causes an empty Feishu `default` account to be automatically created with status `not configured`.
## Steps to Reproduce
1. Configure multiple Feishu bot accounts (e.g. main, coder, news, assistant) with valid appId/appSecret
2. All acco... | open | null | false | 2 | [] | [] | 2026-04-11T07:49:06Z | 2026-05-03T09:12:50Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | S-huNsuke | 247,577,921 | U_kgDODsG9QQ | User | false |
openclaw/openclaw | 4,243,652,449 | I_kwDOQb6kR8788P9h | 64,710 | https://github.com/openclaw/openclaw/issues/64710 | https://api.github.com/repos/openclaw/openclaw/issues/64710 | [Bug]: Agent timeout when calling Ollama - qwen2.5:7b works via direct API but OpenClaw agent times out | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
OpenClaw agent times out when calling Ollama qwen2.5:7b model, despite direct curl to Ollama API succeeding in 7-8 seconds.
### Steps to reproduce
1. docker compose up -d
2. docker compose exec openclaw-gateway openclaw ag... | closed | completed | true | 5 | [
"bug",
"regression"
] | [
"vincentkoc"
] | 2026-04-11T07:56:51Z | 2026-04-29T04:34:12Z | 2026-04-26T20:41:41Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | yuan-b | 73,384,312 | MDQ6VXNlcjczMzg0MzEy | User | false |
openclaw/openclaw | 4,243,766,641 | I_kwDOQb6kR8788r1x | 64,714 | https://github.com/openclaw/openclaw/issues/64714 | https://api.github.com/repos/openclaw/openclaw/issues/64714 | sessions_spawn rejects subagent runtime when streamTo is auto-filled by strict-mode providers | ## Summary
`sessions_spawn` with `runtime="subagent"` is effectively unusable from any LLM provider that auto-materializes optional schema fields (OpenAI strict-mode function calling, the GPT-5 family, OpenAI-compatible proxies that normalize tool schemas). Every call gets hard-rejected with:
```
streamTo is only sup... | closed | completed | true | 1 | [] | [] | 2026-04-11T08:26:55Z | 2026-04-30T04:36:09Z | 2026-04-27T07:11:42Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | kiraje | 97,667,303 | U_kgDOBdJI5w | User | false |
openclaw/openclaw | 4,243,768,433 | I_kwDOQb6kR8788sRx | 64,715 | https://github.com/openclaw/openclaw/issues/64715 | https://api.github.com/repos/openclaw/openclaw/issues/64715 | Telegram: add native reply keyboard support to the message/agent send surface | ## Summary
OpenClaw currently supports Telegram inline buttons, but does not appear to expose native Telegram reply keyboards through the agent / `message` send surface.
This makes common Telegram command menus harder to build than they need to be, especially for mobile-first workflows.
## Current Behavior
From the... | open | null | false | 2 | [] | [] | 2026-04-11T08:27:43Z | 2026-05-03T09:12:51Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | lawrencegs | 232,827 | MDQ6VXNlcjIzMjgyNw== | User | false |
openclaw/openclaw | 4,243,768,509 | I_kwDOQb6kR8788sS9 | 64,716 | https://github.com/openclaw/openclaw/issues/64716 | https://api.github.com/repos/openclaw/openclaw/issues/64716 | Android: npm install fails due to @lancedb/lancedb platform restriction | ## Problem
On Android (Termux), running `npm install -g openclaw` fails because @lancedb/lancedb@0.27.2 explicitly does not support the Android OS in its platform restrictions.
## Error
```
npm error code EBADPLATFORM
npm error notsup Unsupported platform for @lancedb/lancedb@0.27.2: wanted {"os":"darwin,linux,win32",... | closed | completed | true | 1 | [] | [] | 2026-04-11T08:27:45Z | 2026-04-28T05:00:17Z | 2026-04-25T16:16:16Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | Ai003-rgb | 231,196,949 | U_kgDODcfJFQ | User | false |
openclaw/openclaw | 4,243,802,034 | I_kwDOQb6kR87880ey | 64,721 | https://github.com/openclaw/openclaw/issues/64721 | https://api.github.com/repos/openclaw/openclaw/issues/64721 | Cron tool schema missing model/timeout/contextTokens/maxSpendUsd fields | The cron tool (used by agents to create/update cron jobs) does not pass through several important payload fields:
- `payload.model` - which model to use for agentTurn jobs
- `payload.timeoutSeconds` - execution timeout
- `payload.contextTokens` - context window limit
- `maxSpendUsd` - cost cap per run
These fields ... | open | null | false | 2 | [] | [] | 2026-04-11T08:41:41Z | 2026-05-02T23:42:36Z | null | NONE | null | 20260503T000641Z | 2026-05-03T00:06:41Z | Kyzcreig | 9,063,726 | MDQ6VXNlcjkwNjM3MjY= | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.