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 | 4,344,570,308 | I_kwDOQb6kR88AAAABAvThxA | 73,660 | https://github.com/openclaw/openclaw/issues/73660 | https://api.github.com/repos/openclaw/openclaw/issues/73660 | toAnthropicModelRef silently rewrites claude-cli/ model prefix to anthropic/, ignoring explicit provider intent | **Version:** 2026.4.25 (6dd1488db3f4)
## Description
`cli-migration.js:toAnthropicModelRef()` unconditionally rewrites any Claude model reference to `anthropic/` prefix, even when the user explicitly sets `claude-cli/` in their config:
```js
// dist/extensions/anthropic/cli-migration.js lines 7-14
function toAnthrop... | closed | completed | true | 2 | [] | [] | 2026-04-28T16:05:35Z | 2026-05-01T04:40:25Z | 2026-04-29T04:08:14Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | spartman | 399,099 | MDQ6VXNlcjM5OTA5OQ== | User | false |
openclaw/openclaw | 4,344,616,447 | I_kwDOQb6kR88AAAABAvWV_w | 73,661 | https://github.com/openclaw/openclaw/issues/73661 | https://api.github.com/repos/openclaw/openclaw/issues/73661 | [Bug]:TUI cannot connect to gateway after version mismatch (device identity required) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
TUI fails to connect to gateway after upgrading gateway version, returning "device identity required".
### Steps to reproduce
1. Install OpenClaw 2026.3.15:
npm install -g openclaw@2026.3.15
2. Upgrade to 2026.4.24:
... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-28T16:12:30Z | 2026-05-01T04:42:56Z | 2026-04-28T16:17:27Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | oldsix-cell | 189,520,719 | U_kgDOC0vbTw | User | false |
openclaw/openclaw | 4,344,662,202 | I_kwDOQb6kR88AAAABAvZIug | 73,666 | https://github.com/openclaw/openclaw/issues/73666 | https://api.github.com/repos/openclaw/openclaw/issues/73666 | [Bug]：TUI enters infinite reconnect loop on WebSocket close (code 1006) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
TUI repeatedly reconnects on WebSocket close (code 1006) without stopping or showing the underlying error.
### Steps to reproduce
1. Start gateway:
openclaw gateway
2. Use a client that cannot authenticat... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-04-28T16:19:37Z | 2026-05-01T04:42:55Z | 2026-04-28T16:24:06Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | oldsix-cell | 189,520,719 | U_kgDOC0vbTw | User | false |
openclaw/openclaw | 4,344,656,948 | I_kwDOQb6kR88AAAABAvY0NA | 73,665 | https://github.com/openclaw/openclaw/issues/73665 | https://api.github.com/repos/openclaw/openclaw/issues/73665 | Image optimization regression after 2026.4.26: MiniMax-M2.7 image understanding fails with 'Failed to optimize image' | ## Problem
Image understanding stopped working after updating to 2026.4.26. Previously (2026-04-09) images were successfully resized and sent to the model. Now all image inputs fail at the optimization step.
## Environment
- OpenClaw: 2026.4.26
- macOS: 26.2 (arm64)
- Node: 22.22.2
- Channel: WeChat (openclaw-weixin... | closed | completed | true | 1 | [] | [] | 2026-04-28T16:18:48Z | 2026-05-01T04:42:55Z | 2026-04-28T16:24:18Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | renxiangtw | 129,622,322 | U_kgDOB7nhMg | User | false |
openclaw/openclaw | 4,171,410,516 | I_kwDOQb6kR874oqxU | 57,756 | https://github.com/openclaw/openclaw/issues/57756 | https://api.github.com/repos/openclaw/openclaw/issues/57756 | [Bug]: session-key-based session access is not scoped to the calling operator client/device | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Session-key-based gateway APIs and event subscriptions appear to be scoped only by method-level operator permissions, not by the calling client/device’s relationship to the target session.
In the current imple... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-03-30T14:49:47Z | 2026-05-01T04:42:54Z | 2026-04-28T16:26:53Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | xhl2513 | 270,969,872 | U_kgDOECasEA | User | false |
openclaw/openclaw | 4,174,923,671 | I_kwDOQb6kR8742EeX | 58,021 | https://github.com/openclaw/openclaw/issues/58021 | https://api.github.com/repos/openclaw/openclaw/issues/58021 | Feature: `transform_tool_result` plugin hook for modifying tool results before model context | # Feature Proposal: `transform_tool_result` plugin hook
## Summary
Add a new plugin hook that fires after a tool executes but before the result enters the model's context, allowing plugins to modify what the model sees.
## Current state
The plugin hook lifecycle around tool calls:
```
before_tool_call → tool exe... | closed | not_planned | true | 3 | [] | [] | 2026-03-31T00:49:31Z | 2026-05-01T04:42:53Z | 2026-04-28T16:26:59Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Hasnonname | 27,864,846 | MDQ6VXNlcjI3ODY0ODQ2 | User | false |
openclaw/openclaw | 4,157,089,427 | I_kwDOQb6kR873yCaT | 56,132 | https://github.com/openclaw/openclaw/issues/56132 | https://api.github.com/repos/openclaw/openclaw/issues/56132 | [Feature]: | ### Summary
[github-issue.md](https://github.com/user-attachments/files/26318831/github-issue.md)
### Problem to solve
!!!!
### Proposed solution
[github-issue.md](https://github.com/user-attachments/files/26318834/github-issue.md)
### Alternatives considered
[github-issue.md](https://github.com/user-attachments... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-03-28T02:56:31Z | 2026-05-03T08:24:29Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | odlxun | 264,320,025 | U_kgDOD8E0GQ | User | false |
openclaw/openclaw | 4,158,079,453 | I_kwDOQb6kR87310Hd | 56,228 | https://github.com/openclaw/openclaw/issues/56228 | https://api.github.com/repos/openclaw/openclaw/issues/56228 | Discord: Components v2 text not extracted from referenced_message in reply context | ## Problem
When a user replies to a message sent with Discord Components v2 (containers, text display blocks), the reply context body is empty. This is because `resolveDiscordMessageText()` in `route-resolution-*.js` only checks `content` and `embeds[0]` — both of which are empty by design for v2 messages.
## Root Cau... | closed | completed | true | 2 | [] | [] | 2026-03-28T06:29:33Z | 2026-05-04T04:34:36Z | 2026-05-02T01:56:46Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | HollandDrive | 172,116,498 | U_kgDOCkJKEg | User | false |
openclaw/openclaw | 4,161,489,463 | I_kwDOQb6kR874C0o3 | 56,594 | https://github.com/openclaw/openclaw/issues/56594 | https://api.github.com/repos/openclaw/openclaw/issues/56594 | [Bug 2026.3.24] All model calls return "404 status code (no body)" after doctor command | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After the update to 2026.3.24, all model calls return 404 status code (no body). Both OpenAI and Google APIs work correctly via direct curl. The issue is internal to OpenClaw. The doctor command was run today which triggered... | open | null | false | 1 | [
"bug",
"regression"
] | [] | 2026-03-28T19:58:44Z | 2026-05-03T08:24:31Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | alessandromittone | 271,833,988 | U_kgDOEDPbhA | User | false |
openclaw/openclaw | 4,161,687,152 | I_kwDOQb6kR874Dk5w | 56,621 | https://github.com/openclaw/openclaw/issues/56621 | https://api.github.com/repos/openclaw/openclaw/issues/56621 | Feature: pairing list-approved and pairing revoke commands | ## Problem
There is no CLI command to:
1. **List approved/paired senders** for a channel
2. **Revoke (de-pair)** a previously approved sender
## Current state
- `openclaw pairing list <channel>` — only shows **pending** requests, not approved ones
- `openclaw pairing approve <channel> <code>` — works fine
- `opencla... | open | null | false | 1 | [] | [] | 2026-03-28T21:17:54Z | 2026-05-03T08:38:54Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Pibot-Dev | 270,189,658 | U_kgDOEBrEWg | User | false |
openclaw/openclaw | 4,119,638,500 | I_kwDOQb6kR871jLHk | 52,767 | https://github.com/openclaw/openclaw/issues/52767 | https://api.github.com/repos/openclaw/openclaw/issues/52767 | boot-md hook re-runs BOOT.md on every gateway restart | ## Bug Description
boot-md is a gateway internal hook that runs BOOT.md every time the gateway starts. However, BOOT.md is designed for first-run initialization — it should only run once, not on every startup.
## Actual Behavior
Every gateway restart unconditionally executes BOOT.md, causing unnecessary re-execution... | open | null | false | 3 | [] | [] | 2026-03-23T09:54:35Z | 2026-05-04T21:00:43Z | null | NONE | null | 20260505T000034Z | 2026-05-05T00:00:34Z | skyline7hd | 103,162,920 | U_kgDOBiYkKA | User | false |
openclaw/openclaw | 4,162,621,752 | I_kwDOQb6kR874HJE4 | 56,813 | https://github.com/openclaw/openclaw/issues/56813 | https://api.github.com/repos/openclaw/openclaw/issues/56813 | Bundled openshell module crashes exec tool — requires TELEGRAM_BOT_TOKEN even when standalone CLI is used | **Version:** OpenClaw 2026.3.24 and 2026.3.28
**Environment:** Ubuntu 22.04, OpenShell 0.0.16 (standalone CLI at `~/.local/bin/openshell`), OpenClaw installed globally via `npm install -g openclaw`
## Description
The bundled `openshell` npm module at `/usr/lib/node_modules/openclaw/node_modules/openshell` requires `... | open | null | false | 2 | [] | [] | 2026-03-29T05:04:04Z | 2026-05-02T16:40:24Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | teosborne | 4,041,297 | MDQ6VXNlcjQwNDEyOTc= | User | false |
openclaw/openclaw | 4,163,218,946 | I_kwDOQb6kR874Ja4C | 56,862 | https://github.com/openclaw/openclaw/issues/56862 | https://api.github.com/repos/openclaw/openclaw/issues/56862 | builtin memory_search can return empty results until builtin index is forced to reindex | ## Summary
Built-in `memory_search` can return empty results even when the builtin sqlite memory index already exists and manual FTS queries match the expected content.
In our case, forcing a builtin full reindex immediately fixed the issue, without changing provider/model/endpoint.
This suggests the problem is like... | open | null | false | 2 | [] | [] | 2026-03-29T07:43:14Z | 2026-05-02T17:54:53Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | CadilarcZhang | 39,903,338 | MDQ6VXNlcjM5OTAzMzM4 | User | false |
openclaw/openclaw | 4,163,453,607 | I_kwDOQb6kR874KUKn | 56,897 | https://github.com/openclaw/openclaw/issues/56897 | https://api.github.com/repos/openclaw/openclaw/issues/56897 | xurl skill: add confirm-before-post enforcement and audit log instructions | ## Summary
The `xurl` skill (Twitter/X API CLI wrapper) currently relies on soft agent-level guardrails in its SKILL.md. Given that xurl can post public tweets, reply, DM, and perform other high-impact actions on behalf of users, the skill should enforce stronger safety patterns at the SKILL.md instruction level.
## ... | open | null | false | 1 | [] | [] | 2026-03-29T09:00:26Z | 2026-05-02T17:30:34Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | xinbenlv | 640,325 | MDQ6VXNlcjY0MDMyNQ== | User | false |
openclaw/openclaw | 4,163,992,584 | I_kwDOQb6kR874MXwI | 56,994 | https://github.com/openclaw/openclaw/issues/56994 | https://api.github.com/repos/openclaw/openclaw/issues/56994 | fix: use atomic writes for crash-safety on exec approvals, restart sentinel, shell profile, and saveJsonFile | ## Summary
Several critical state files are written with non-atomic `writeFileSync` or `fs.writeFile`, risking data loss on crash. The codebase already has `writeJsonAtomic` (temp + fsync + rename + dir-sync) in `src/infra/json-files.ts` and a sync `writeTextFileAtomic` in `src/secrets/shared.ts`, but these paths don'... | open | null | false | 2 | [] | [] | 2026-03-29T12:03:21Z | 2026-05-03T07:04:10Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | jamie8johnson | 85,816,907 | MDQ6VXNlcjg1ODE2OTA3 | User | false |
openclaw/openclaw | 4,163,997,861 | I_kwDOQb6kR874MZCl | 56,997 | https://github.com/openclaw/openclaw/issues/56997 | https://api.github.com/repos/openclaw/openclaw/issues/56997 | [Feature] Control UI: Support channel-based session view toggle | Currently, the Control UI only shows the main webchat session. Conversations from other channels (WeChat, Feishu, QQ) are in separate isolated sessions that cannot be viewed in the Control UI.
**Feature Request:**
Add a session view mode in the Control UI that allows users to:
1. **Separated view** — Show each chann... | open | null | false | 1 | [] | [] | 2026-03-29T12:05:22Z | 2026-05-03T07:44:34Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | lisuicong | 271,954,608 | U_kgDOEDWysA | User | false |
openclaw/openclaw | 4,164,091,295 | I_kwDOQb6kR874Mv2f | 57,019 | https://github.com/openclaw/openclaw/issues/57019 | https://api.github.com/repos/openclaw/openclaw/issues/57019 | fix: session write lock race — async release can delete newly-acquired lock | ## Session write lock race: release can destroy a freshly acquired lock
### Summary
In `src/agents/session-write-lock.ts`, `releaseHeldLock` deletes the session key from the in-memory `HELD_LOCKS` Map (line 149) **before** starting the async file cleanup (`handle.close()` + `fs.rm()`, lines 150-161). This opens a rac... | open | null | false | 2 | [] | [] | 2026-03-29T12:32:26Z | 2026-05-04T00:21:49Z | null | NONE | null | 20260504T030031Z | 2026-05-04T03:00:31Z | jamie8johnson | 85,816,907 | MDQ6VXNlcjg1ODE2OTA3 | User | false |
openclaw/openclaw | 4,164,125,134 | I_kwDOQb6kR874M4HO | 57,031 | https://github.com/openclaw/openclaw/issues/57031 | https://api.github.com/repos/openclaw/openclaw/issues/57031 | fix: QueuedFileWriter silently swallows all write errors | ## Summary
`QueuedFileWriter` in `queued-file-writer.ts` catches all write errors with `.catch(() => undefined)` and continues silently. When the disk fills up, all payload log and audit trail writes fail with no indication — the queue keeps processing but producing no output.
## Location
`src/agents/queued-file-wri... | open | null | false | 2 | [] | [] | 2026-03-29T12:42:24Z | 2026-05-03T06:49:20Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | jamie8johnson | 85,816,907 | MDQ6VXNlcjg1ODE2OTA3 | User | false |
openclaw/openclaw | 4,164,748,284 | I_kwDOQb6kR874PQP8 | 57,119 | https://github.com/openclaw/openclaw/issues/57119 | https://api.github.com/repos/openclaw/openclaw/issues/57119 | Feature Request: Enforce skill execution sandbox when enabled: false | ## Problem
Setting `skills.entries.<name>.enabled: false` (including the UI toggle) only removes the skill from auto-matching. It does **not** prevent the agent from:
1. Reading the skill's `SKILL.md` via the `read` tool
2. Executing the skill's scripts via the `exec` tool
This means the "disable" toggle is **securi... | open | null | false | 1 | [] | [] | 2026-03-29T15:23:37Z | 2026-05-03T08:38:55Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | LittleCadet | 40,167,192 | MDQ6VXNlcjQwMTY3MTky | User | false |
openclaw/openclaw | 4,164,762,256 | I_kwDOQb6kR874PTqQ | 57,123 | https://github.com/openclaw/openclaw/issues/57123 | https://api.github.com/repos/openclaw/openclaw/issues/57123 | Feature: Incremental/adaptive inbound message debounce | ## Problem
Currently `messages.inbound.debounceMs` is a fixed timer — every new message resets the same static delay. This is suboptimal for both scenarios:
- **Single message**: User waits the full debounce (e.g. 2s) unnecessarily before getting a response.
- **Rapid-fire messages**: The fixed window may be too shor... | open | null | false | 2 | [] | [] | 2026-03-29T15:28:19Z | 2026-05-03T10:18:47Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | K-NRS | 20,605,111 | MDQ6VXNlcjIwNjA1MTEx | User | false |
openclaw/openclaw | 4,164,913,850 | I_kwDOQb6kR874P4q6 | 57,143 | https://github.com/openclaw/openclaw/issues/57143 | https://api.github.com/repos/openclaw/openclaw/issues/57143 | [Bug]: gh-issues skill references hardcoded /data/.clawdbot/ path (doesn't exist on standard installs) | ## Description
The bundled `gh-issues` skill (`SKILL.md`) contains **9 hardcoded references** to `/data/.clawdbot/` which is a legacy path from the old ClawdBot branding. On standard OpenClaw installations, this path does not exist — the correct base path is `~/.openclaw/`.
## Affected paths in SKILL.md
| Hardcoded ... | open | null | false | 1 | [] | [] | 2026-03-29T16:10:10Z | 2026-05-02T18:31:04Z | null | NONE | null | 20260502T210030Z | 2026-05-02T21:00:30Z | mvalentsev | 2,305,718 | MDQ6VXNlcjIzMDU3MTg= | User | false |
openclaw/openclaw | 4,165,221,412 | I_kwDOQb6kR874RDwk | 57,202 | https://github.com/openclaw/openclaw/issues/57202 | https://api.github.com/repos/openclaw/openclaw/issues/57202 | [Feature]: 分布式部署的可能性？ | ### Summary
如果用两台OpenClaw网关节点连在一起，一台控制多agents ，一台保存记忆、处理记忆，这个方案可行吗？
### Problem to solve
解决单台云服务器性能不够，无法处理大量数据向量嵌入
### Proposed solution
如果能实现分布式节点，那就可以多台网关多agen共享同一套记忆库，理论上就没有性能瓶颈了。
### Alternatives considered
_No response_
### Impact
单台网关性能不足的情况
### Evidence/examples
_No response_
### Additional informati... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-03-29T17:30:56Z | 2026-05-03T08:38:56Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Gavincent1 | 34,804,116 | MDQ6VXNlcjM0ODA0MTE2 | User | false |
openclaw/openclaw | 4,165,239,911 | I_kwDOQb6kR874RIRn | 57,204 | https://github.com/openclaw/openclaw/issues/57204 | https://api.github.com/repos/openclaw/openclaw/issues/57204 | perf(dev): gateway:watch triggers full rebuild on every startup when working tree is dirty | ## Summary
When running `pnpm gateway:watch` with any uncommitted changes in `src/` or `extensions/`, the startup script unconditionally triggers a full TypeScript rebuild (~5 minutes) on every invocation — even when the dist output is already up to date.
## Root Cause
`scripts/run-node.mjs` `shouldBuild()` uses `ha... | open | null | false | 1 | [] | [] | 2026-03-29T17:37:05Z | 2026-05-03T10:40:50Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | oNo500 | 24,784,662 | MDQ6VXNlcjI0Nzg0NjYy | User | false |
openclaw/openclaw | 4,165,270,924 | I_kwDOQb6kR874RP2M | 57,210 | https://github.com/openclaw/openclaw/issues/57210 | https://api.github.com/repos/openclaw/openclaw/issues/57210 | [Feature Request] Multi-Session / New Chat Support | ## Feature Request: Multi-Session / New Chat Support
### Problem
Currently Control UI has only one global session. All conversations share the same context, causing:
1. Mixed contexts between different topics
2. No isolation between work and personal chats
3. Performance issues with long conversation history
4. Cannot... | closed | not_planned | true | 1 | [] | [] | 2026-03-29T17:44:47Z | 2026-05-01T04:42:22Z | 2026-04-28T18:07:01Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | kumikoda110 | 15,316,691 | MDQ6VXNlcjE1MzE2Njkx | User | false |
openclaw/openclaw | 4,165,515,287 | I_kwDOQb6kR874SLgX | 57,241 | https://github.com/openclaw/openclaw/issues/57241 | https://api.github.com/repos/openclaw/openclaw/issues/57241 | [Bug]: tailscale.mode "serve" overwrites Funnel to tailnet-only on every gateway restart; tailscale.mode "funnel" fails with token auth | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
When gateway.tailscale.mode is set to "serve", the gateway overwrites any existing Tailscale Funnel configuration on every restart, reverting public paths to tailnet-only. When set to "funnel" (the intended fix), the gateway... | open | null | false | 1 | [
"bug",
"regression"
] | [] | 2026-03-29T18:56:28Z | 2026-05-03T08:24:35Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Neo-Creator-1 | 261,085,700 | U_kgDOD4_aBA | User | false |
openclaw/openclaw | 4,165,623,729 | I_kwDOQb6kR874Sl-x | 57,259 | https://github.com/openclaw/openclaw/issues/57259 | https://api.github.com/repos/openclaw/openclaw/issues/57259 | Add github-copilot as media-understanding image provider | ## Feature Request
### Problem
GitHub Copilot models (e.g. \github-copilot/gpt-4o\, \github-copilot/claude-opus-4.6\) support \ ext+image\ input, but \github-copilot\ is not registered as a media-understanding provider for image capability.
Currently the provider support matrix for Image is:
- OpenAI, Anthropic, Goog... | open | null | false | 1 | [] | [] | 2026-03-29T19:40:13Z | 2026-05-03T09:12:28Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | vvshakirov | 49,247,118 | MDQ6VXNlcjQ5MjQ3MTE4 | User | false |
openclaw/openclaw | 4,165,647,661 | I_kwDOQb6kR874Sr0t | 57,270 | https://github.com/openclaw/openclaw/issues/57270 | https://api.github.com/repos/openclaw/openclaw/issues/57270 | tools.exec.approvalRunningNoticeMs is documented but rejected by the 2026.3.24 config schema | ## Summary
The OpenClaw docs describe `tools.exec.approvalRunningNoticeMs` as a supported config key for approval-gated exec follow-up notices, but the local `2026.3.24 (cff6dc9)` build rejects that field as an unrecognized key under `tools.exec`.
This looks like a docs/schema mismatch rather than an operator mistake... | open | null | false | 1 | [] | [] | 2026-03-29T19:51:39Z | 2026-05-03T11:49:05Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | glorifiedautocomplete | 271,297,221 | U_kgDOECuqxQ | User | false |
openclaw/openclaw | 4,165,656,352 | I_kwDOQb6kR874St8g | 57,271 | https://github.com/openclaw/openclaw/issues/57271 | https://api.github.com/repos/openclaw/openclaw/issues/57271 | docker.binds rejected for paths outside per-session sandbox dir | ## Bug
`sandbox.docker.binds` rejects bind mount sources outside the per-session sandbox directory and agent workspace, even though the [sandboxing docs](https://docs.openclaw.ai/gateway/sandboxing#custom-bind-mounts) show examples with arbitrary host paths (`/home/user/source`, `/var/data/myapp`, `/mnt/cache`).
## E... | open | null | false | 1 | [] | [] | 2026-03-29T19:55:44Z | 2026-05-02T18:31:06Z | null | NONE | null | 20260502T210030Z | 2026-05-02T21:00:30Z | Belthezar46 | 85,139,005 | MDQ6VXNlcjg1MTM5MDA1 | User | false |
openclaw/openclaw | 4,165,686,041 | I_kwDOQb6kR874S1MZ | 57,278 | https://github.com/openclaw/openclaw/issues/57278 | https://api.github.com/repos/openclaw/openclaw/issues/57278 | bug(telegram): reply to bot photo injects bot-sent media as inbound user attachment | When a user replies to a bot-sent photo, `resolveReplyMediaForMessage` downloads the original image and injects it as inbound user media. The agent incorrectly receives and reacts to images it previously generated.
## Root Cause
`isSelfAuthoredTelegramMessage` exists and is correctly used at the top-level inbound han... | open | null | false | 1 | [] | [] | 2026-03-29T20:09:30Z | 2026-05-03T08:24:39Z | null | CONTRIBUTOR | null | 20260503T090142Z | 2026-05-03T09:01:42Z | amittell | 1,388,680 | MDQ6VXNlcjEzODg2ODA= | User | false |
openclaw/openclaw | 4,165,691,102 | I_kwDOQb6kR874S2be | 57,281 | https://github.com/openclaw/openclaw/issues/57281 | https://api.github.com/repos/openclaw/openclaw/issues/57281 | Auth profile failover blocked by file lock contention on rate_limit errors | ## Bug
When an auth profile hits a 429 `rate_limit_error`, the gateway logs `decision=rotate_profile` but never executes the rotation to the next profile. The run terminates as failed. Timeout errors rotate correctly.
## Root Cause
In `src/agents/pi-embedded-runner/run.ts`, the `shouldRotate` block calls `await mayb... | open | null | false | 1 | [] | [] | 2026-03-29T20:11:53Z | 2026-05-03T08:38:58Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | jeremyknows | 237,305,675 | U_kgDODiT_Sw | User | false |
openclaw/openclaw | 4,165,700,760 | I_kwDOQb6kR874S4yY | 57,284 | https://github.com/openclaw/openclaw/issues/57284 | https://api.github.com/repos/openclaw/openclaw/issues/57284 | [Bug]: openclaw agent --deliver silently drops messages (exit 0, no delivery) | ### Bug Summary
`openclaw agent --deliver --channel signal --to "+1234567890" --message "notification text"` reports success (exit 0) but the message is never delivered to the target channel. The command spawns a full LLM agent turn that processes the notification text as a user message, and the agent either responds ... | open | null | false | 1 | [
"bug",
"cli",
"commands",
"agents"
] | [
"vincentkoc"
] | 2026-03-29T20:16:13Z | 2026-05-03T07:15:05Z | null | CONTRIBUTOR | null | 20260503T090142Z | 2026-05-03T09:01:42Z | joeykrug | 5,925,937 | MDQ6VXNlcjU5MjU5Mzc= | User | false |
openclaw/openclaw | 4,165,797,094 | I_kwDOQb6kR874TQTm | 57,296 | https://github.com/openclaw/openclaw/issues/57296 | https://api.github.com/repos/openclaw/openclaw/issues/57296 | WebChat TTS reports success but produces no audible playback or downloadable audio (Telegram works) | # WebChat TTS reports success but produces no audible playback or downloadable audio (Telegram works)
## Problem
In WebChat, TTS tool calls return success ("Generated audio reply"), but no audio is heard and no media/file is surfaced to the user.
## Environment
- OpenClaw `2026.3.28` (updated from `2026.3.13`)
- macO... | open | null | false | 2 | [] | [] | 2026-03-29T21:03:53Z | 2026-05-05T02:38:17Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | andhai | 20,992,727 | MDQ6VXNlcjIwOTkyNzI3 | User | false |
openclaw/openclaw | 4,165,807,060 | I_kwDOQb6kR874TSvU | 57,297 | https://github.com/openclaw/openclaw/issues/57297 | https://api.github.com/repos/openclaw/openclaw/issues/57297 | WebChat TTS: add explicit routing/playback logs and user-visible delivery status | # WebChat TTS: add explicit routing/playback logs and user-visible delivery status
## Problem
When TTS is triggered from WebChat, operators may see a tool-level success signal but still get no audible output and no downloadable media. Current logs do not clearly show where the WebChat TTS pipeline failed (generation v... | open | null | false | 1 | [] | [] | 2026-03-29T21:08:48Z | 2026-05-03T08:24:43Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | andhai | 20,992,727 | MDQ6VXNlcjIwOTkyNzI3 | User | false |
openclaw/openclaw | 4,165,876,824 | I_kwDOQb6kR874TjxY | 57,307 | https://github.com/openclaw/openclaw/issues/57307 | https://api.github.com/repos/openclaw/openclaw/issues/57307 | [Feature]: Memory importance scoring + time decay for long-term memory | ### Summary
Add importance scoring (1-10) at memory write time and Weibull time-decay so that old/trivial memories fade while important ones persist, improving retrieval quality for long-running assistants.
### Problem to solve
All memory entries have equal weight regardless of age or content significance. A note ab... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-03-29T21:45:22Z | 2026-05-03T06:49:23Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | kophysty | 34,319,062 | MDQ6VXNlcjM0MzE5MDYy | User | false |
openclaw/openclaw | 4,165,914,687 | I_kwDOQb6kR874TtA_ | 57,313 | https://github.com/openclaw/openclaw/issues/57313 | https://api.github.com/repos/openclaw/openclaw/issues/57313 | Plugin SDK: expose sessionModelOverride in PluginHookAgentContext (before_model_resolve) | ## Summary
When a user manually sets a model override via `/model`, the `before_model_resolve` hook has no way to detect this and correctly yields. The hook currently fires on every message and may overwrite the user-intended model with its own routing decision.
## Current state
- `PluginHookBeforeModelResolveEvent`... | open | null | false | 1 | [] | [] | 2026-03-29T22:06:06Z | 2026-05-02T17:30:36Z | null | NONE | null | 20260502T180029Z | 2026-05-02T18:00:29Z | SmithU64 | 117,943,562 | U_kgDOBwetCg | User | false |
openclaw/openclaw | 4,166,273,000 | I_kwDOQb6kR874VEfo | 57,369 | https://github.com/openclaw/openclaw/issues/57369 | https://api.github.com/repos/openclaw/openclaw/issues/57369 | session.reset: support mode: "never" for manual-only session reset behavior | ## Summary
Please add a proper `mode: "never"` option to session reset configuration so sessions can be configured for true manual-only reset behavior.
Right now the closest workaround is to use `mode: "idle"` with a very large `idleMinutes` value, but that is only an approximation and not a real "never reset automat... | open | null | false | 1 | [] | [] | 2026-03-30T00:51:41Z | 2026-05-03T10:18:49Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | aa7469 | 265,133,328 | U_kgDOD82dEA | User | false |
openclaw/openclaw | 4,166,323,687 | I_kwDOQb6kR874VQ3n | 57,376 | https://github.com/openclaw/openclaw/issues/57376 | https://api.github.com/repos/openclaw/openclaw/issues/57376 | sessions.list returns phantom store-key entry with null updatedAt and no sessionId | ## Description
`sessions.list` returns a phantom entry whose key matches the agent's sessions store path (e.g. `agent:sean:sessions`) with `updatedAt: null` and no `sessionId` or `origin`. This appears to be a store-level placeholder leaking into the session list response.
## Repro
Call `sessions.list` with `agentId... | open | null | false | 1 | [] | [] | 2026-03-30T01:10:38Z | 2026-05-03T09:12:31Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | smarchetti | 3,430,072 | MDQ6VXNlcjM0MzAwNzI= | User | false |
openclaw/openclaw | 4,166,368,721 | I_kwDOQb6kR874Vb3R | 57,387 | https://github.com/openclaw/openclaw/issues/57387 | https://api.github.com/repos/openclaw/openclaw/issues/57387 | Feature: HMAC/token-based mutual auth for inter-agent sessions_send | ## Problem
In multi-agent deployments, any agent can forge a `sessions_send` message claiming to originate from any other agent. There is no cryptographic verification of sender identity.
In a 19-agent swarm scenario, a compromised agent (e.g., a coding agent processing untrusted external content) could fabricate ins... | open | null | false | 2 | [] | [] | 2026-03-30T01:25:07Z | 2026-05-03T10:40:51Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | hubertmodularshift-cmyk | 261,700,715 | U_kgDOD5k8aw | User | false |
openclaw/openclaw | 4,166,405,857 | I_kwDOQb6kR874Vk7h | 57,389 | https://github.com/openclaw/openclaw/issues/57389 | https://api.github.com/repos/openclaw/openclaw/issues/57389 | Bug: auth-profiles.json field name `api_key` silently ignored — correct field is `key` | ## Summary
When configuring `auth-profiles.json` with `"api_key"` as the field name for API key credentials, the key resolution silently fails. The correct field name is `"key"`, but this is not documented and there is no validation error — the profile is simply skipped.
This causes downstream features (e.g., LCM plu... | open | null | false | 1 | [] | [] | 2026-03-30T01:36:11Z | 2026-05-03T06:07:10Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | zeeyuu222 | 267,768,244 | U_kgDOD_XRtA | User | false |
openclaw/openclaw | 4,166,507,563 | I_kwDOQb6kR874V9wr | 57,404 | https://github.com/openclaw/openclaw/issues/57404 | https://api.github.com/repos/openclaw/openclaw/issues/57404 | [Feature]: Expose per-run token usage on the WebSocket lifecycle event stream | ## Problem
The lifecycle `end` event contains only timing data (`startedAt`, `endedAt`, `aborted`, `stopReason`). Token and cost metrics collected in `agentMeta` during a run are persisted to the session store but never surfaced on the WebSocket event stream. This means external observers (dashboards, monitoring tools... | open | null | false | 2 | [] | [] | 2026-03-30T02:06:28Z | 2026-05-03T08:38:59Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Studio729 | 34,817,975 | MDQ6VXNlcjM0ODE3OTc1 | User | false |
openclaw/openclaw | 4,166,540,776 | I_kwDOQb6kR874WF3o | 57,408 | https://github.com/openclaw/openclaw/issues/57408 | https://api.github.com/repos/openclaw/openclaw/issues/57408 | [Bug]: project-local .env is ignored in 2026.3.28, OPENCLAW_CONFIG_PATH falls back to ~/.openclaw/openclaw.json | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
There appears to be a regression between 2026.3.24 and 2026.3.28.
In 2026.3.28, the project-local .env is not picked up during CLI bootstrap.
OPENCLAW_CONFIG_PATH set in the project .env is ignored, and OpenClaw falls back ... | open | null | false | 2 | [
"bug"
] | [] | 2026-03-30T02:15:09Z | 2026-05-03T08:24:45Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | chaoscaros | 31,836,404 | MDQ6VXNlcjMxODM2NDA0 | User | false |
openclaw/openclaw | 4,344,701,524 | I_kwDOQb6kR88AAAABAvbiVA | 73,670 | https://github.com/openclaw/openclaw/issues/73670 | https://api.github.com/repos/openclaw/openclaw/issues/73670 | [Bug]: Mac App overwrites user config fields on save, removing existing entries | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Saving configuration in the Mac App removes existing fields from the config file without warning.
### Steps to reproduce
1. Initialize config using the Mac App
2. Manually edit:
~/.openclaw/config.json
3... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-28T16:26:25Z | 2026-05-01T04:42:52Z | 2026-04-28T16:31:49Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | oldsix-cell | 189,520,719 | U_kgDOC0vbTw | User | false |
openclaw/openclaw | 4,344,719,418 | I_kwDOQb6kR88AAAABAvcoOg | 73,673 | https://github.com/openclaw/openclaw/issues/73673 | https://api.github.com/repos/openclaw/openclaw/issues/73673 | [Bug]: Gateway repeatedly restarts under launchd after crash, making it difficult to debug | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When running under launchd, gateway crashes cause immediate restarts, leading to continuous restart loops and making debugging difficult.
### Steps to reproduce
1. Configure OpenClaw to run via launchd (Launc... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-28T16:29:31Z | 2026-05-03T04:44:02Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | oldsix-cell | 189,520,719 | U_kgDOC0vbTw | User | false |
openclaw/openclaw | 4,344,463,460 | I_kwDOQb6kR88AAAABAvNAZA | 73,655 | https://github.com/openclaw/openclaw/issues/73655 | https://api.github.com/repos/openclaw/openclaw/issues/73655 | Gateway leak triad on plugin restart: Manifest EADDRINUSE retry loop, signal-handler accumulation, sync I/O on session JSONL → WS handshake starvation | ## Summary
After a normal post-update gateway restart on 2026.4.26 (homebrew install, gateway PID 39890), three independent leaks compounded over ~10 minutes and starved the WS upgrade handler. Result: every WebSocket open from a client (Paperclip's adapter, in our case) timed out at 26s with `closed before connect` /... | open | null | false | 15 | [] | [] | 2026-04-28T15:48:51Z | 2026-05-04T18:15:56Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | bryangauvin | 36,140,319 | MDQ6VXNlcjM2MTQwMzE5 | User | false |
openclaw/openclaw | 4,344,751,687 | I_kwDOQb6kR88AAAABAvemRw | 73,677 | https://github.com/openclaw/openclaw/issues/73677 | https://api.github.com/repos/openclaw/openclaw/issues/73677 | [Bug]: Feature Request: Convert Feishu custom stickers to Unicode emoji or text representation | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Problem
When using OpenClaw with Feishu (Lark), custom stickers (e.g., [OK], [完成], [调皮]) are not visible to the AI agent. The agent only receives standard Unicode emoji (👍 ✅ 🎉) but not Feishu's custom bracket-style stic... | open | null | false | 1 | [
"bug"
] | [] | 2026-04-28T16:35:18Z | 2026-05-03T07:45:28Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | xzzzewennn | 132,631,242 | U_kgDOB-fKyg | User | false |
openclaw/openclaw | 4,344,751,037 | I_kwDOQb6kR88AAAABAvejvQ | 73,676 | https://github.com/openclaw/openclaw/issues/73676 | https://api.github.com/repos/openclaw/openclaw/issues/73676 | [Bug]: CLI crashes when current working directory is deleted (uv_cwd error not handled) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
CLI exits with uv_cwd error when the current working directory is deleted, without handling or fallback.
### Steps to reproduce
1. Create and enter a directory:
mkdir test-dir
cd test-dir
2. Delete the... | open | null | false | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-28T16:35:10Z | 2026-05-03T09:49:28Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | oldsix-cell | 189,520,719 | U_kgDOC0vbTw | User | false |
openclaw/openclaw | 4,344,394,559 | I_kwDOQb6kR88AAAABAvIzPw | 73,647 | https://github.com/openclaw/openclaw/issues/73647 | https://api.github.com/repos/openclaw/openclaw/issues/73647 | [Bug]: 2026.4.26 Telegram-only gateway spins CPU in channel runtime / staged runtime deps | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Summary
OpenClaw 2026.4.26 still pegs CPU in a Telegram-only gateway after the 4.26 runtime-deps fix. The previous `openclaw-unknown-*` runtime-deps directories are gone, but the gateway now loops around channel runtime ... | closed | not_planned | true | 7 | [
"bug",
"regression"
] | [] | 2026-04-28T15:38:44Z | 2026-05-01T04:41:33Z | 2026-04-28T21:41:12Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | mattmcintyre | 3,009,067 | MDQ6VXNlcjMwMDkwNjc= | User | false |
openclaw/openclaw | 4,344,780,785 | I_kwDOQb6kR88AAAABAvgX8Q | 73,680 | https://github.com/openclaw/openclaw/issues/73680 | https://api.github.com/repos/openclaw/openclaw/issues/73680 | Per-agent `verboseDefault` / `elevatedDefault` rejected by config schema, despite resolver and SDK types supporting them | ## Summary
The runtime treats `verboseDefault` and `elevatedDefault` as valid per-agent overrides (resolver, plugin-SDK types, status text, command handlers), but the strict zod entry schema does not declare them, so `openclaw config set agents.list.<idx>.verboseDefault on` (or `elevatedDefault`) is rejected with `Unr... | open | null | false | 1 | [] | [] | 2026-04-28T16:40:33Z | 2026-05-03T07:45:29Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | kevinheinrichs | 113,271,657 | U_kgDOBsBjaQ | User | false |
openclaw/openclaw | 4,344,783,812 | I_kwDOQb6kR88AAAABAvgjxA | 73,681 | https://github.com/openclaw/openclaw/issues/73681 | https://api.github.com/repos/openclaw/openclaw/issues/73681 | googlechat: REACTION_ADDED / REACTION_REMOVED webhook events are silently dropped | ## Problem
The Google Chat webhook handler in `extensions/googlechat/src/monitor.ts:58-59` and `extensions/googlechat/src/monitor-webhook.ts:81-93` only proceeds when `eventType === "MESSAGE"`. Any `REACTION_ADDED` or `REACTION_REMOVED` events Google Chat could deliver (when the Chat app subscribes to those event type... | open | null | false | 1 | [] | [] | 2026-04-28T16:41:09Z | 2026-05-03T12:31:52Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | ScientificProgrammer | 4,348,294 | MDQ6VXNlcjQzNDgyOTQ= | User | false |
openclaw/openclaw | 4,344,795,511 | I_kwDOQb6kR88AAAABAvhRdw | 73,684 | https://github.com/openclaw/openclaw/issues/73684 | https://api.github.com/repos/openclaw/openclaw/issues/73684 | [Bug]: Session-file lock cascade — single stale .lock blocks every subsequent agent invocation in container | ## Summary
When OpenClaw's GRAEAE agent crashes or is killed mid-invocation, it leaves a stale `.lock` file under `~/.openclaw/agents/main/sessions/`. Every subsequent `openclaw agent` invocation in the same container then fails with the same lock-timeout error, falls back to the embedded model chain (which itself fai... | closed | not_planned | true | 1 | [] | [] | 2026-04-28T16:43:07Z | 2026-05-01T04:42:52Z | 2026-04-28T16:46:03Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | perlowja | 162,375,972 | U_kgDOCa2pJA | User | false |
openclaw/openclaw | 4,344,790,733 | I_kwDOQb6kR88AAAABAvg-zQ | 73,683 | https://github.com/openclaw/openclaw/issues/73683 | https://api.github.com/repos/openclaw/openclaw/issues/73683 | Session bootstrap race condition sends zero messages to zai/glm-5.1 causing 400 error | ## Bug Description
When a new webchat session starts, the first API request to the LLM provider is sent **before** the bootstrap user message is written to the session file. This results in `prompt.submitted` having `messages: []` (zero conversation messages, only a system prompt).
Models that tolerate a system-only ... | closed | completed | true | 2 | [] | [] | 2026-04-28T16:42:19Z | 2026-05-03T04:35:40Z | 2026-04-30T17:59:59Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | lhtpluto | 132,045,861 | U_kgDOB97cJQ | User | false |
openclaw/openclaw | 4,336,781,835 | I_kwDOQb6kR88AAAABAn4KCw | 72,865 | https://github.com/openclaw/openclaw/issues/72865 | https://api.github.com/repos/openclaw/openclaw/issues/72865 | [Bug] 4.25 regression: Bedrock Claude 4.6 "assistant message prefill" error on first message (Feishu channel) | ## Bug type
Regression (worked before, now fails)
## Summary
After upgrading from OpenClaw 2026.4.24 to 2026.4.25, all new sessions on the Feishu channel fail on the **first user message** with:
> Validation error: The model returned the following errors: This model does not support assistant message prefill. The con... | closed | completed | true | 2 | [] | [] | 2026-04-27T15:09:04Z | 2026-05-03T04:36:44Z | 2026-04-30T14:54:02Z | CONTRIBUTOR | null | 20260503T060030Z | 2026-05-03T06:00:30Z | wujiaming88 | 7,887,545 | MDQ6VXNlcjc4ODc1NDU= | User | false |
openclaw/openclaw | 4,344,820,181 | I_kwDOQb6kR88AAAABAvix1Q | 73,687 | https://github.com/openclaw/openclaw/issues/73687 | https://api.github.com/repos/openclaw/openclaw/issues/73687 | [Bug]: Default fallback chain includes models that fail the minimum context-window check, making them effectively unusable | ## Summary
The default `agents.defaults.model.fallbacks` chain ships with `google/gemma-4-31b-it` (`contextWindow: 8192` per `modelsConfig`), but the agent enforces a minimum of 16000 tokens before it will use a model. So when fallback fires (e.g. primary returns 429 RESOURCE_EXHAUSTED), the chain advances to `gemma-4... | closed | completed | true | 2 | [] | [] | 2026-04-28T16:46:53Z | 2026-05-01T04:42:19Z | 2026-04-28T16:51:45Z | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | perlowja | 162,375,972 | U_kgDOCa2pJA | User | false |
openclaw/openclaw | 4,344,804,450 | I_kwDOQb6kR88AAAABAvh0Yg | 73,685 | https://github.com/openclaw/openclaw/issues/73685 | https://api.github.com/repos/openclaw/openclaw/issues/73685 | [Bug]: @openclaw/discord@2026.3.13 fails on openclaw@2026.4.26: missing plugin-sdk/discord export | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading OpenClaw core from `2026.4.25` to `2026.4.26`, the external Discord plugin fails to load during gateway startup.
`@openclaw/discord@2026.3.13` imports `openclaw/plugin-sdk/discord`, but `openclaw@2026.4.26` ... | closed | completed | true | 4 | [
"bug"
] | [] | 2026-04-28T16:44:24Z | 2026-05-01T04:39:51Z | 2026-04-28T19:59:45Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | rderickson9 | 14,813,558 | MDQ6VXNlcjE0ODEzNTU4 | User | false |
openclaw/openclaw | 4,344,835,106 | I_kwDOQb6kR88AAAABAvjsIg | 73,688 | https://github.com/openclaw/openclaw/issues/73688 | https://api.github.com/repos/openclaw/openclaw/issues/73688 | [Bug] /new session creation fails with ZAI/GLM: provider rejects system-only messages (400) | ## Bug Description
**`/new` slash command and fresh session creation fails with ZAI (GLM) provider** when the default model is any GLM variant (`glm-5.1`, `glm-5-turbo`, etc.). The error surfaces as:
```
⚠️ Something went wrong while processing your request. Please try again, or use /new to start a fresh session.
```... | closed | completed | true | 2 | [] | [] | 2026-04-28T16:49:34Z | 2026-05-01T04:41:57Z | 2026-04-28T16:53:06Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Ahmedsalehsaleh | 126,225,437 | U_kgDOB4YMHQ | User | false |
openclaw/openclaw | 4,344,845,101 | I_kwDOQb6kR88AAAABAvkTLQ | 73,690 | https://github.com/openclaw/openclaw/issues/73690 | https://api.github.com/repos/openclaw/openclaw/issues/73690 | [Bug]: | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
On macOS native in OpenClaw 2026.4.26, openclaw agent calls can take 100s+ to return even when the embedded agent session itself completes in under 1 second.
### Steps to reproduce
1. Configure a local native OpenClaw install... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-28T16:51:27Z | 2026-05-01T04:42:49Z | 2026-04-28T16:56:50Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Noxiffow | 130,965,846 | U_kgDOB85hVg | User | false |
openclaw/openclaw | 4,344,788,745 | I_kwDOQb6kR88AAAABAvg3CQ | 73,682 | https://github.com/openclaw/openclaw/issues/73682 | https://api.github.com/repos/openclaw/openclaw/issues/73682 | [Bug]: "No credentials found for profile anthropic:claude-cli" fires in cron lane despite fresh keychain — #71026 fix not reaching runtime path | ## Summary
In \`2026.4.26\`, the cron lane fails with \`Error: No credentials found for profile "anthropic:claude-cli"\` when the persisted \`auth-profiles.json\` token has expired, even though the macOS Keychain (\`Claude Code-credentials\`) holds fresh OAuth tokens. This is despite \`EXTERNAL_CLI_SYNC_PROVIDERS\` in... | closed | completed | true | 4 | [] | [] | 2026-04-28T16:42:01Z | 2026-05-01T04:41:59Z | 2026-04-28T19:42:36Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | RyanSandoval | 7,387,444 | MDQ6VXNlcjczODc0NDQ= | User | false |
openclaw/openclaw | 4,195,723,199 | I_kwDOQb6kR876Fae_ | 59,797 | https://github.com/openclaw/openclaw/issues/59797 | https://api.github.com/repos/openclaw/openclaw/issues/59797 | [Bug]: 4.1 openclaw-weixin Failed to start login | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Windows下openclaw 2026.4.1 安装 openclaw-weixin 2.1.3，一键安装失败，手动openclaw channels login --channel openclaw-weixin，正在启动微信扫码登录...
报错：Failed to start login: AbortError: This operation was aborted
Channel login failed: Error: Failed... | closed | not_planned | true | 3 | [
"bug",
"regression"
] | [] | 2026-04-02T16:36:22Z | 2026-05-01T04:42:48Z | 2026-04-28T16:58:55Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | kingos8888 | 182,330,138 | U_kgDOCt4jGg | User | false |
openclaw/openclaw | 4,196,425,227 | I_kwDOQb6kR876IF4L | 59,868 | https://github.com/openclaw/openclaw/issues/59868 | https://api.github.com/repos/openclaw/openclaw/issues/59868 | edit tool shows false positive error when oldText match fails after file was already fixed |
## Problem
When using the edit tool, if the oldText parameter does not match (e.g., due to minor whitespace differences or encoding), the tool returns an error like:
```
Error: Tool did not complete. Edit: in ~/.openclaw/openclaw.json (398 chars) failed
```
However, if the user then manually corrects the file (e.g.... | closed | not_planned | true | 1 | [] | [] | 2026-04-02T19:01:10Z | 2026-05-01T04:42:48Z | 2026-04-28T16:59:02Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | zhuyankarl | 7,082,758 | MDQ6VXNlcjcwODI3NTg= | User | false |
openclaw/openclaw | 4,196,709,377 | I_kwDOQb6kR876JLQB | 59,888 | https://github.com/openclaw/openclaw/issues/59888 | https://api.github.com/repos/openclaw/openclaw/issues/59888 | [Bug]: WhatsApp web-auto-reply dispatches stale/cached reply to subsequent inbound message (5ms response, no LLM call) | ## Summary
When two WhatsApp messages arrive within ~90 seconds to the same session, the second message receives an **identical copy** of the first message's reply. The second reply completes in **5ms** (vs 597ms for the first), proving no LLM call was made.
## Steps to reproduce
1. Send a WhatsApp message to the bo... | closed | completed | true | 2 | [] | [] | 2026-04-02T20:03:48Z | 2026-05-01T04:42:47Z | 2026-04-28T16:59:09Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Freddy-Cach | 215,536,142 | U_kgDODNjSDg | User | false |
openclaw/openclaw | 4,197,504,859 | I_kwDOQb6kR876MNdb | 59,957 | https://github.com/openclaw/openclaw/issues/59957 | https://api.github.com/repos/openclaw/openclaw/issues/59957 | [Bug]: Tool output unknown error after forced update to v2026.4.2 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Tool output unknown error after forced update to v2026.4.2
Why is OpenClaw updating itself when not requested! I installed V2026.3.28 since it actually works and then when I ran it, it ran that dam package manager and update... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-04-02T23:53:26Z | 2026-05-01T04:42:46Z | 2026-04-28T16:59:15Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | jdc4429 | 19,594,564 | MDQ6VXNlcjE5NTk0NTY0 | User | false |
openclaw/openclaw | 4,197,657,785 | I_kwDOQb6kR876Myy5 | 59,970 | https://github.com/openclaw/openclaw/issues/59970 | https://api.github.com/repos/openclaw/openclaw/issues/59970 | Proposal: reply-first chat behavior, retrieval guardrails, and result-oriented sub-agent reporting | ## Summary
I’d like to propose a small bundle of behavior/routing improvements that significantly improves chat quality for direct-message and reply-based workflows, especially on Telegram-like surfaces.
These changes are not about adding new model capabilities. They are mainly about better prioritization, cleaner re... | closed | not_planned | true | 2 | [] | [] | 2026-04-03T00:53:38Z | 2026-05-01T04:42:45Z | 2026-04-28T16:59:22Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Litton45 | 87,689,927 | MDQ6VXNlcjg3Njg5OTI3 | User | false |
openclaw/openclaw | 4,204,590,119 | I_kwDOQb6kR876nPQn | 60,819 | https://github.com/openclaw/openclaw/issues/60819 | https://api.github.com/repos/openclaw/openclaw/issues/60819 | doctor/status memory false positive after mem0 runtime recovery | # Bug: `openclaw doctor` / `openclaw status` report memory unavailable even when mem0 runtime is healthy
## Summary
After fixing a real `openclaw-mem0` runtime regression locally, mem0 runtime recovered and is actively working again, but OpenClaw still reports memory as unavailable/unregistered in diagnostics.
This a... | closed | not_planned | true | 1 | [] | [] | 2026-04-04T10:54:46Z | 2026-05-01T04:42:45Z | 2026-04-28T17:00:10Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | duclamvan | 5,202,937 | MDQ6VXNlcjUyMDI5Mzc= | User | false |
openclaw/openclaw | 4,344,870,407 | I_kwDOQb6kR88AAAABAvl2Bw | 73,691 | https://github.com/openclaw/openclaw/issues/73691 | https://api.github.com/repos/openclaw/openclaw/issues/73691 | [Bug]: MEMORY.md grows unbounded → bootstrap overflow → Gateway freeze | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
MEMORY.md grows unbounded with no size guard. After extended use, exceeds bootstrap hard limit (default 12KB per file / 60KB total), causing session write-lock timeout and Gateway freeze.
### Steps to reproduce
1. Enable d... | open | null | false | 1 | [
"bug"
] | [] | 2026-04-28T16:56:06Z | 2026-05-03T12:31:53Z | null | NONE | null | 20260503T150031Z | 2026-05-03T15:00:31Z | yangdoom007 | 267,955,226 | U_kgDOD_isGg | User | false |
openclaw/openclaw | 4,344,342,361 | I_kwDOQb6kR88AAAABAvFnWQ | 73,638 | https://github.com/openclaw/openclaw/issues/73638 | https://api.github.com/repos/openclaw/openclaw/issues/73638 | [Feature]: Non-interactive onboarding support for trusted-proxy gateway auth | ## Summary
The \`onboard\` command cannot configure \`trusted-proxy\` gateway auth non-interactively, blocking automated and scripted deployments behind a reverse proxy.
## Problem to solve
The \`onboard\` command supports non-interactive setup for \`token\` and \`password\` gateway auth via flags like \`--gateway-a... | open | null | false | 1 | [] | [] | 2026-04-28T15:30:55Z | 2026-05-03T03:37:45Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | mattf | 112,653 | MDQ6VXNlcjExMjY1Mw== | User | false |
openclaw/openclaw | 4,344,344,940 | I_kwDOQb6kR88AAAABAvFxbA | 73,639 | https://github.com/openclaw/openclaw/issues/73639 | https://api.github.com/repos/openclaw/openclaw/issues/73639 | [Feature]: Non-interactive onboarding support for trusted-proxy gateway auth | ## Summary
The \`onboard\` command cannot configure \`trusted-proxy\` gateway auth non-interactively, blocking automated and scripted deployments behind a reverse proxy.
## Problem to solve
The \`onboard\` command supports non-interactive setup for \`token\` and \`password\` gateway auth via flags like \`--gateway-a... | open | null | false | 2 | [] | [] | 2026-04-28T15:31:20Z | 2026-05-03T06:50:37Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | mattf | 112,653 | MDQ6VXNlcjExMjY1Mw== | User | false |
openclaw/openclaw | 4,344,345,617 | I_kwDOQb6kR88AAAABAvF0EQ | 73,640 | https://github.com/openclaw/openclaw/issues/73640 | https://api.github.com/repos/openclaw/openclaw/issues/73640 | Anthropic 400: empty text ContentBlock not filtered on user/toolResult/assistant-error/images-only/preserveSignatures paths | ## Summary
Anthropic API rejects requests with `Validation error: The text field in the ContentBlock object at messages.N.content.0 is blank. Add text to the text field, and try again.` when the session transcript contains an empty-text content block that was not caught by sanitization.
## Observed
```
Validation er... | closed | completed | true | 2 | [] | [] | 2026-04-28T15:31:26Z | 2026-05-01T04:42:07Z | 2026-04-28T19:42:38Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | jowhee327 | 67,151,378 | MDQ6VXNlcjY3MTUxMzc4 | User | false |
openclaw/openclaw | 4,344,364,855 | I_kwDOQb6kR88AAAABAvG_Nw | 73,643 | https://github.com/openclaw/openclaw/issues/73643 | https://api.github.com/repos/openclaw/openclaw/issues/73643 | Feature Request: Add option to disable missed jobs catch-up on gateway restart |
## Problem
When the OpenClaw gateway restarts (e.g. after a crash or manual restart), it automatically runs all missed cron jobs that were scheduled during the downtime. This happens regardless of whether those jobs are still relevant — for example, A股 trading monitoring crons that were supposed to run at 11:00am fir... | closed | not_planned | true | 1 | [] | [] | 2026-04-28T15:34:23Z | 2026-05-03T04:34:08Z | 2026-05-01T03:21:04Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Barrett-Ethan | 276,641,461 | U_kgDOEH02tQ | User | false |
openclaw/openclaw | 4,344,381,793 | I_kwDOQb6kR88AAAABAvIBYQ | 73,646 | https://github.com/openclaw/openclaw/issues/73646 | https://api.github.com/repos/openclaw/openclaw/issues/73646 | [Bug]: `openclaw crestodian` exits 0 with "needs interactive TTY" error in non-TTY contexts; sibling subcommands correctly exit non-zero on the same input | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
`pnpm openclaw crestodian` (and equivalently `node ./dist/index.js crestodian`) detects when stdin is not a TTY, prints `Crestodian needs an interactive TTY. Use --message for one command.` to stdout, then exit... | closed | completed | true | 3 | [
"bug",
"bug:behavior"
] | [] | 2026-04-28T15:36:54Z | 2026-05-05T04:28:15Z | 2026-05-02T07:41:51Z | CONTRIBUTOR | null | 20260505T060034Z | 2026-05-05T06:00:34Z | bittoby | 218,712,309 | U_kgDODQlI9Q | User | false |
openclaw/openclaw | 4,344,421,059 | I_kwDOQb6kR88AAAABAvKaww | 73,652 | https://github.com/openclaw/openclaw/issues/73652 | https://api.github.com/repos/openclaw/openclaw/issues/73652 | Gateway accepts connections before ready signal, causing handshake timeouts on startup | ## Problem
Gateway listens on TCP port and accepts WebSocket connections before it is internally ready, causing all clients that connect during the ~15-22s startup window to experience handshake timeouts.
## Steps to Reproduce
1. Start OpenClaw Gateway (or trigger a config reload / restart)
2. Immediately open contr... | closed | completed | true | 2 | [] | [] | 2026-04-28T15:42:34Z | 2026-05-02T04:28:39Z | 2026-04-29T09:32:48Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | spenceryang1996-dot | 251,943,929 | U_kgDODwRb-Q | User | false |
openclaw/openclaw | 4,344,966,874 | I_kwDOQb6kR88AAAABAvru2g | 73,699 | https://github.com/openclaw/openclaw/issues/73699 | https://api.github.com/repos/openclaw/openclaw/issues/73699 | Bridge Discord voice channel I/O to text-channel agent session (voice-as-IO) | # Feature Request: Bridge Discord voice channel I/O to text-channel agent session
## Summary
Today, when Discord voice is enabled (`channels.discord.voice.enabled = true`) and the bot joins a voice channel, OpenClaw spins up a **separate isolated session** for that voice channel. Voice transcripts go to that session,... | open | null | false | 3 | [] | [] | 2026-04-28T17:13:00Z | 2026-05-03T08:25:56Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | mdpoirier-abbey | 280,093,496 | U_kgDOELHjOA | User | false |
openclaw/openclaw | 4,344,366,165 | I_kwDOQb6kR88AAAABAvHEVQ | 73,644 | https://github.com/openclaw/openclaw/issues/73644 | https://api.github.com/repos/openclaw/openclaw/issues/73644 | Feature Request: Add option to disable missed jobs catch-up on gateway restart |
## Problem
When the OpenClaw gateway restarts (e.g. after a crash or manual restart), it automatically runs all missed cron jobs that were scheduled during the downtime. This happens regardless of whether those jobs are still relevant — for example, A股 trading monitoring crons that were supposed to run at 11:00am fir... | closed | not_planned | true | 1 | [] | [] | 2026-04-28T15:34:35Z | 2026-05-01T04:42:41Z | 2026-04-28T17:23:00Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Barrett-Ethan | 276,641,461 | U_kgDOEH02tQ | User | false |
openclaw/openclaw | 4,345,001,703 | I_kwDOQb6kR88AAAABAvt25w | 73,701 | https://github.com/openclaw/openclaw/issues/73701 | https://api.github.com/repos/openclaw/openclaw/issues/73701 | [Bug]: `openclaw models --agent <id> auth login` writes OAuth tokens to the default agent's profile, ignoring `--agent` | ## Summary
`openclaw models --agent <id> auth login --provider <p>` silently writes the resulting OAuth credentials to the **default agent's** `auth-profiles.json` instead of the agent specified by `--agent`. The flag is honoured by other paths (e.g. `openclaw agent --agent <id> --message ...` routes correctly), but t... | closed | completed | true | 1 | [] | [] | 2026-04-28T17:19:51Z | 2026-05-01T04:42:40Z | 2026-04-28T17:24:56Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | MattConstantine | 85,278,990 | MDQ6VXNlcjg1Mjc4OTkw | User | false |
openclaw/openclaw | 4,098,480,802 | I_kwDOQb6kR870Sdqi | 50,096 | https://github.com/openclaw/openclaw/issues/50096 | https://api.github.com/repos/openclaw/openclaw/issues/50096 | Long-Term Memory & Knowledge Management | An agent's value is directly proportional to what it can remember. What good is a partner who starts every conversation from zero? The promise of OpenClaw is continuity — an agent that learns, recalls, and evolves with you.
The plumbing is fighting that promise.
Driftnet has been listening to the memory thread. The c... | open | null | false | 12 | [] | [] | 2026-03-19T00:32:01Z | 2026-05-04T20:22:46Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | ocdlmv1 | 266,897,137 | U_kgDOD-iG8Q | User | false |
openclaw/openclaw | 4,111,563,759 | I_kwDOQb6kR871EXvv | 51,451 | https://github.com/openclaw/openclaw/issues/51451 | https://api.github.com/repos/openclaw/openclaw/issues/51451 | feat: Per-request workspace override via HTTP header (x-openclaw-workspace) | ## 🚀 Feature Request
### Summary
Add support for a per-request workspace override via an HTTP header (`x-openclaw-workspace`) on the `/v1/chat/completions` endpoint.
---
### Problem / Motivation
When deploying OpenClaw as a **multi-tenant SaaS backend** (e.g., serving thousands of merchants/users from a shared cl... | open | null | false | 4 | [] | [] | 2026-03-21T06:05:20Z | 2026-05-04T18:47:24Z | null | NONE | null | 20260504T210031Z | 2026-05-04T21:00:31Z | david1228 | 2,298,080 | MDQ6VXNlcjIyOTgwODA= | User | false |
openclaw/openclaw | 4,184,779,696 | I_kwDOQb6kR875bquw | 58,822 | https://github.com/openclaw/openclaw/issues/58822 | https://api.github.com/repos/openclaw/openclaw/issues/58822 | fix(agents): subagent model precedence — global default shadowed by parent agent's own model | ## Bug
PR #58003 (commit e394262bd8, merged 2026-03-31) reordered the fallback chain in `resolveSubagentConfiguredModelSelection` to prefer `agentConfig.model` (the agent's **own** model) over `agents.defaults.subagents.model` (the **global** subagent default).
This is correct for named agents with an explicit `subag... | open | null | false | 2 | [] | [] | 2026-04-01T07:25:04Z | 2026-05-01T00:44:49Z | null | CONTRIBUTOR | null | 20260501T113057Z | 2026-05-01T11:30:57Z | joeykrug | 5,925,937 | MDQ6VXNlcjU5MjU5Mzc= | User | false |
openclaw/openclaw | 4,345,004,992 | I_kwDOQb6kR88AAAABAvuDwA | 73,702 | https://github.com/openclaw/openclaw/issues/73702 | https://api.github.com/repos/openclaw/openclaw/issues/73702 | sessions_send reply forwarding creates cross-session prompt-injection vector | # Issue: `sessions_send` reply forwarding creates cross-session prompt-injection vector
**Target:** openclaw/openclaw
**Severity:** medium (security / agent integrity)
**Reporter:** alvelda
---
## Title
```
sessions_send reply text injects into caller as bare role=user, bypassing sender envelope checks
```
## Body... | closed | completed | true | 2 | [] | [] | 2026-04-28T17:20:32Z | 2026-05-01T04:41:32Z | 2026-04-28T21:41:20Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | alvelda | 1,775,986 | MDQ6VXNlcjE3NzU5ODY= | User | false |
openclaw/openclaw | 4,192,201,009 | I_kwDOQb6kR8753-kx | 59,509 | https://github.com/openclaw/openclaw/issues/59509 | https://api.github.com/repos/openclaw/openclaw/issues/59509 | feat: support per-model location override for google-vertex provider | ## Problem
When using Google Vertex AI, different models require different regional endpoints:
- **Preview models** (e.g. `gemini-3.1-pro-preview`, `gemini-3-flash-preview`) are only available on the `global` endpoint (`global-aiplatform.googleapis.com`)
- **GA models** (e.g. `gemini-2.5-pro`, `gemini-2.5-flash`) requ... | open | null | false | 1 | [] | [] | 2026-04-02T06:54:37Z | 2026-05-01T00:45:00Z | null | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | pure-maple | 54,388,301 | MDQ6VXNlcjU0Mzg4MzAx | User | false |
openclaw/openclaw | 4,192,519,865 | I_kwDOQb6kR8755Ma5 | 59,528 | https://github.com/openclaw/openclaw/issues/59528 | https://api.github.com/repos/openclaw/openclaw/issues/59528 | [Bug]: The bundled plugins openshell does not work correctly since version 2026.03.13 | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
After I enable the bundled plugin openshell by the [offical docs](https://docs.openclaw.ai/gateway/openshell), the openshell does not create sandbox correctly . The status of the openshell sandbox is always sto... | open | null | false | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T07:45:06Z | 2026-05-02T07:16:08Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | HiddenWorld-lk | 58,210,041 | MDQ6VXNlcjU4MjEwMDQx | User | false |
openclaw/openclaw | 4,193,630,720 | I_kwDOQb6kR8759boA | 59,613 | https://github.com/openclaw/openclaw/issues/59613 | https://api.github.com/repos/openclaw/openclaw/issues/59613 | [Bug]: sandbox container lack of permission to read write files | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
deploy openclaw in docker and with sandbox enabled ( https://docs.openclaw.ai/install/docker#agent-sandbox ).
the created sandbox container does not have permission to read write `/workspace` files
### Steps ... | closed | not_planned | true | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-02T10:35:27Z | 2026-05-03T04:34:25Z | 2026-05-01T01:27:04Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | chanjarster | 612,413 | MDQ6VXNlcjYxMjQxMw== | User | false |
openclaw/openclaw | 4,194,920,708 | I_kwDOQb6kR876CWkE | 59,710 | https://github.com/openclaw/openclaw/issues/59710 | https://api.github.com/repos/openclaw/openclaw/issues/59710 | Heartbeat silently stops after ~20h despite #52270 fix in 2026.3.28 | ### Bug Description
Heartbeat interval timer appears to stop being re-armed after extended runtime, even after the supposed fix in #52270 (2026.3.28).
### Environment
- **OpenClaw version:** 2026.3.28 (f9b1079)
- **OS:** Linux 6.17.0-19-generic (x64)
- **Node:** v24.14.0
- **Heartbeat config:** 30m interval (default)
... | open | null | false | 2 | [] | [] | 2026-04-02T14:17:48Z | 2026-05-02T07:49:33Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | dxyang42 | 270,384,353 | U_kgDOEB284Q | User | false |
openclaw/openclaw | 4,194,962,119 | I_kwDOQb6kR876CgrH | 59,718 | https://github.com/openclaw/openclaw/issues/59718 | https://api.github.com/repos/openclaw/openclaw/issues/59718 | [Feature]: Allow writable custom binds with workspaceAccess: "none" for least-privilege sandbox access | ### Summary
Allow sandboxed agents to edit explicitly writable custom bind mounts while workspaceAccess remains "none".
### Problem to solve
I want to keep the main agent workspace isolated so the sandbox cannot write broadly to the real workspace, but I still need the agent to modify a small, explicit subset of rea... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-02T14:23:55Z | 2026-05-02T09:29:37Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | bartlomiejwolk | 325,348 | MDQ6VXNlcjMyNTM0OA== | User | false |
openclaw/openclaw | 4,195,096,745 | I_kwDOQb6kR876DBip | 59,740 | https://github.com/openclaw/openclaw/issues/59740 | https://api.github.com/repos/openclaw/openclaw/issues/59740 | Architecture: Consolidate ~15 custom boundary check scripts into unified tooling | ## Problem
OpenClaw maintains approximately 15 custom lint/boundary check scripts for enforcing architectural rules:
```
scripts/
├── check-architecture-smells.mjs
├── check-channel-agnostic-boundaries.mjs
├── check-cli-startup-memory.mjs
├── check-extension-plugin-sdk-boundary.mjs
├── check-no-conflict-markers.mjs
├... | open | null | false | 1 | [] | [] | 2026-04-02T14:45:18Z | 2026-05-02T11:08:12Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | dboone323 | 65,318,437 | MDQ6VXNlcjY1MzE4NDM3 | User | false |
openclaw/openclaw | 4,195,594,077 | I_kwDOQb6kR876E69d | 59,782 | https://github.com/openclaw/openclaw/issues/59782 | https://api.github.com/repos/openclaw/openclaw/issues/59782 | feat: Add silentErrorReplies / suppressAgentErrors support for WhatsApp channel | ## Feature Request
### Problem
When the primary model (e.g., Anthropic Claude) is rate-limited or overloaded and the fallback model also fails, OpenClaw sends a generic error message to the user:
> ⚠️ Something went wrong while processing your request. Please try again, or use /new to start a fresh session.
This is ... | open | null | false | 1 | [] | [] | 2026-04-02T16:11:02Z | 2026-05-02T07:49:37Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | smccaffrey | 11,462,398 | MDQ6VXNlcjExNDYyMzk4 | User | false |
openclaw/openclaw | 4,195,800,323 | I_kwDOQb6kR876FtUD | 59,808 | https://github.com/openclaw/openclaw/issues/59808 | https://api.github.com/repos/openclaw/openclaw/issues/59808 | fix: qmd vector search silently falls back to FTS on first query after gateway restart | ## Problem
After every gateway restart, the **first** `memory_search` call silently degrades to FTS-only mode instead of using the qmd vector search path. Subsequent calls work correctly.
**Symptom**:
```json
{
"provider": "none",
"fallback": {
"from": "qmd",
"reason": "Unexpected token 'm', \"[mcporter] ... | open | null | false | 1 | [] | [] | 2026-04-02T16:53:51Z | 2026-05-02T08:59:57Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | Silver-Aurora | 179,392,834 | U_kgDOCrFRQg | User | false |
openclaw/openclaw | 4,195,831,671 | I_kwDOQb6kR876F093 | 59,817 | https://github.com/openclaw/openclaw/issues/59817 | https://api.github.com/repos/openclaw/openclaw/issues/59817 | Discord DMs can fork into duplicate channel sessions during channel-info/network failures | ## Summary
Discord direct messages can be recorded under multiple session keys for the same conversation when DM channel classification fails during inbound processing.
Expected: a DM with one user should consistently use a single direct session key.
Observed: the same DM ended up split across these three keys:
- `ag... | closed | completed | true | 3 | [] | [] | 2026-04-02T17:00:34Z | 2026-05-04T04:34:45Z | 2026-05-02T01:30:17Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | DooPeePey | 266,929,590 | U_kgDOD-kFtg | User | false |
openclaw/openclaw | 4,196,584,916 | I_kwDOQb6kR876Is3U | 59,881 | https://github.com/openclaw/openclaw/issues/59881 | https://api.github.com/repos/openclaw/openclaw/issues/59881 | Plugin hook ctx.channelId returns provider name instead of actual channel ID | ## Bug: Plugin hook context `ctx.channelId` returns provider name instead of actual channel ID
### Environment
- OpenClaw version: 2026.4.1
- Channel: Discord
- Plugin: `@vectorize-io/hindsight-openclaw@0.4.19`
### Description
In plugin hooks (`before_prompt_build`, `agent_end`), `ctx.channelId` resolves to the provi... | closed | completed | true | 3 | [] | [] | 2026-04-02T19:36:50Z | 2026-05-04T04:34:20Z | 2026-05-02T03:06:10Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | bradfreels | 179,389,909 | U_kgDOCrFF1Q | User | false |
openclaw/openclaw | 4,197,006,227 | I_kwDOQb6kR876KTuT | 59,913 | https://github.com/openclaw/openclaw/issues/59913 | https://api.github.com/repos/openclaw/openclaw/issues/59913 | OpenAI compat gateway images bypass sanitization/resize, hit Anthropic 5MB limit | ## Summary
Images sent via the OpenAI-compatible `/v1/chat/completions` endpoint (as `image_url` content parts) bypass the image sanitization/resize pipeline. Large images (e.g., Mac Retina screenshots) are forwarded to Anthropic at full resolution and size, hitting Anthropic's 5MB per-image limit and returning `Could... | open | null | false | 1 | [] | [] | 2026-04-02T21:10:09Z | 2026-05-02T10:36:11Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | htplbc | 5,052,967 | MDQ6VXNlcjUwNTI5Njc= | User | false |
openclaw/openclaw | 4,197,389,900 | I_kwDOQb6kR876LxZM | 59,945 | https://github.com/openclaw/openclaw/issues/59945 | https://api.github.com/repos/openclaw/openclaw/issues/59945 | Slack socket mode fails with "reserved redaction sentinel" error despite valid tokens — infinite restart loop | ## Summary
Slack channel fails to connect in socket mode with valid tokens. Gateway enters infinite restart loop with `reserved redaction sentinel "__OPENCLAW_REDACTED__" is not valid config data` error.
## Environment
- OpenClaw version: 2026.4.2 (d74a122)
- OS: macOS 15.4 (arm64)
- Node: v25.8.2
- Slack plugin: @ope... | open | null | false | 2 | [] | [] | 2026-04-02T23:06:55Z | 2026-05-02T08:59:59Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | MarkkayTavion | 66,388,320 | MDQ6VXNlcjY2Mzg4MzIw | User | false |
openclaw/openclaw | 4,197,605,525 | I_kwDOQb6kR876MmCV | 59,966 | https://github.com/openclaw/openclaw/issues/59966 | https://api.github.com/repos/openclaw/openclaw/issues/59966 | Hardcoded 120s timeout in channel layer overrides configurable timeoutSeconds | ## Bug Description
The `agents.defaults.timeoutSeconds` setting in `openclaw.json` is overridden by a hardcoded 120-second timeout in the compiled channel code at `dist/channel-D5chgcOT.js`.
The relevant line:
```js
const timeoutPromise = new Promise((_, reject) => setTimeout(() => reject(new Error("Agent response ti... | open | null | false | 1 | [] | [] | 2026-04-03T00:32:40Z | 2026-05-02T09:22:01Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | unw1red | 45,374,291 | MDQ6VXNlcjQ1Mzc0Mjkx | User | false |
openclaw/openclaw | 4,198,063,041 | I_kwDOQb6kR876OVvB | 60,034 | https://github.com/openclaw/openclaw/issues/60034 | https://api.github.com/repos/openclaw/openclaw/issues/60034 | [Feature]: `session-memory` hook does not comply with AGENTS.md workspace directory convention | ### Summary
The built-in `session-memory` hook hardcodes output to `memory/<YYYY-MM-DD-slug>.md`, and does not comply with the directory structure defined in `AGENTS.md`. It is recommended to add `outputPath` + `filename` configuration options to the hook to make the output path fully customizable.
### Problem to sol... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-03T03:25:23Z | 2026-05-02T07:49:39Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | 9692922 | 237,369,701 | U_kgDODiX5ZQ | User | false |
openclaw/openclaw | 4,198,081,300 | I_kwDOQb6kR876OaMU | 60,041 | https://github.com/openclaw/openclaw/issues/60041 | https://api.github.com/repos/openclaw/openclaw/issues/60041 | [Bug]: Twitch channel enabled but never connects after update (schema/config mismatch, no clear validation path) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
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 glob... | open | null | false | 1 | [
"bug",
"regression"
] | [] | 2026-04-03T03:31:41Z | 2026-05-02T09:29:41Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | nhaener | 16,600,979 | MDQ6VXNlcjE2NjAwOTc5 | User | false |
openclaw/openclaw | 4,198,321,067 | I_kwDOQb6kR876PUur | 60,078 | https://github.com/openclaw/openclaw/issues/60078 | https://api.github.com/repos/openclaw/openclaw/issues/60078 | [Bug]: Announce delivery ignores modelByChannel, always uses agent default model | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Subagent announce delivery resolves the LLM model via `resolveDefaultModelForAgent()` which returns `agents.defaults.model.primary`, ignoring `channels.modelByChannel` overrides configured for the parent sessio... | open | null | false | 1 | [] | [] | 2026-04-03T04:47:56Z | 2026-05-02T07:16:12Z | null | NONE | null | 20260502T090030Z | 2026-05-02T09:00:30Z | davidbordenwi | 43,112,759 | MDQ6VXNlcjQzMTEyNzU5 | User | false |
openclaw/openclaw | 4,198,388,125 | I_kwDOQb6kR876PlGd | 60,084 | https://github.com/openclaw/openclaw/issues/60084 | https://api.github.com/repos/openclaw/openclaw/issues/60084 | node:sqlite fails on ARM64 (aarch64) in containers — SQLITE_BUSY on all file-backed databases | ## Summary
OpenClaw's memory search is completely broken on ARM64 (aarch64) when running in Kubernetes containers. The root cause is that `node:sqlite`'s `DatabaseSync` cannot acquire POSIX advisory file locks (`fcntl F_SETLK`) on ARM64 Linux in containerized environments. Every SQL operation on file-backed databases ... | open | null | false | 1 | [] | [] | 2026-04-03T05:10:09Z | 2026-05-02T09:05:44Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | spencer2211 | 28,957,500 | MDQ6VXNlcjI4OTU3NTAw | User | false |
openclaw/openclaw | 4,198,623,387 | I_kwDOQb6kR876Qeib | 60,103 | https://github.com/openclaw/openclaw/issues/60103 | https://api.github.com/repos/openclaw/openclaw/issues/60103 | [Feishu] Auto-normalize @mention syntax across message formats (card/text/post) | ## Problem
Feishu has 3 different message formats with incompatible @mention syntax:
| Format | Syntax | Used when |
|--------|--------|-----------|
| **text** | `<at user_id="ou_xxx">Name</at>` | Non-card text messages |
| **card** (interactive) | `<at id=ou_xxx></at>` | Streaming replies, card messages |
| **post** (... | open | null | false | 1 | [] | [] | 2026-04-03T06:11:35Z | 2026-05-02T09:29:45Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | guizhe614-lgtm | 229,044,981 | U_kgDODaby9Q | User | false |
openclaw/openclaw | 4,199,157,139 | I_kwDOQb6kR876Sg2T | 60,154 | https://github.com/openclaw/openclaw/issues/60154 | https://api.github.com/repos/openclaw/openclaw/issues/60154 | [Feature]: Per-account media inbound directory for multi-account Feishu | ### Summary
Per-account media inbound directory for multi-account Feishu
### Problem to solve
When multiple Feishu accounts are configured in OpenClaw (e.g. Gauss + Rita as separate enterprise apps), all downloaded media files go to the same global directory: ... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-03T08:15:26Z | 2026-05-02T10:36:13Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | yangericjie | 273,303,864 | U_kgDOEEpJOA | User | false |
openclaw/openclaw | 4,199,371,569 | I_kwDOQb6kR876TVMx | 60,190 | https://github.com/openclaw/openclaw/issues/60190 | https://api.github.com/repos/openclaw/openclaw/issues/60190 | [Feature]: support delayed aggregation for multi-part user inputs | ### Summary
When a user sends an image followed by text in a channel (e.g., QQ), the bot replies to each message immediately instead of waiting to understand the combined intent. A configurable aggregation delay (e.g., 3–5s) before processing would allow the bot to capture related messages and handle them as one conve... | open | null | false | 1 | [
"enhancement"
] | [] | 2026-04-03T09:00:44Z | 2026-05-02T10:36:14Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | Auroar | 55,086,404 | MDQ6VXNlcjU1MDg2NDA0 | User | false |
openclaw/openclaw | 4,199,374,649 | I_kwDOQb6kR876TV85 | 60,191 | https://github.com/openclaw/openclaw/issues/60191 | https://api.github.com/repos/openclaw/openclaw/issues/60191 | Fallback does not advance on OpenRouter 403 API key budget limit exceeded during /new | # Fallback does not advance on OpenRouter `403 API key budget limit exceeded` during `/new`
## Problem
When the primary model is `openrouter/xiaomi/mimo-v2-pro` and it returns:
`403 API key budget limit exceeded (monthly limit). Contact your org admin.`
OpenClaw does **not** advance to the next model in `agents.def... | open | null | false | 5 | [] | [] | 2026-04-03T09:01:15Z | 2026-05-02T09:22:04Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | omgitsgela | 25,420,944 | MDQ6VXNlcjI1NDIwOTQ0 | User | false |
openclaw/openclaw | 4,199,567,980 | I_kwDOQb6kR876UFJs | 60,204 | https://github.com/openclaw/openclaw/issues/60204 | https://api.github.com/repos/openclaw/openclaw/issues/60204 | Detect or retire legacy ensure-whatsapp cron checks that misreport gateway health | ## Summary
A legacy user crontab entry that runs `~/.openclaw/bin/ensure-whatsapp.sh` can produce misleading `~/.openclaw/logs/whatsapp-health.log` entries such as `WARN: Gateway inactive, starting via systemd` even while `openclaw-gateway.service` is healthy.
## What we observed
- The script is launched from user cro... | closed | completed | true | 3 | [] | [] | 2026-04-03T09:39:19Z | 2026-05-04T04:35:19Z | 2026-05-01T21:46:08Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | mySebbe | 14,988,278 | MDQ6VXNlcjE0OTg4Mjc4 | User | false |
openclaw/openclaw | 4,199,848,300 | I_kwDOQb6kR876VJls | 60,225 | https://github.com/openclaw/openclaw/issues/60225 | https://api.github.com/repos/openclaw/openclaw/issues/60225 | Regression: operator scopes cleared again for shared-auth API clients (v2026.4.1) | ## Context
Follow-up to #27494 (fixed in 9c14299, closed as completed). The fix regressed in v2026.4.1.
## What happened
The original fix (9c14299) correctly added `&& !sharedAuthOk` to `clearUnboundScopes()`. However, commit 8b88b927cb (`gateway: clear unbound scopes for trusted-proxy auth`, #57692) refactored the ... | open | null | false | 1 | [] | [] | 2026-04-03T10:24:49Z | 2026-05-02T09:22:06Z | null | CONTRIBUTOR | null | 20260502T120029Z | 2026-05-02T12:00:29Z | tobalsan | 2,358,453 | MDQ6VXNlcjIzNTg0NTM= | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.