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 | 4,362,218,697 | I_kwDOQb6kR88AAAABBAIsyQ | 75,328 | https://github.com/openclaw/openclaw/issues/75328 | https://api.github.com/repos/openclaw/openclaw/issues/75328 | Telegram channel crashes: Cannot find package 'json5' (ESM import failure in bundled deps) | ## Version
OpenClaw 2026.4.29 (a448042)
## Environment
- OS: CachyOS (Arch Linux), Kernel 6.19.12-1
- Node: v25.9.0
- Install: npm global (`npm install -g openclaw@2026.4.29`)
- Hardware: Dell Latitude 5300
## Symptoms
Telegram channel continuously crashes on startup. Auto-restart runs through all 10 attempts with ex... | closed | completed | true | 1 | [] | [] | 2026-05-01T01:03:56Z | 2026-05-03T04:34:32Z | 2026-05-01T01:08:35Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | erforschtbot-cmyk | 272,903,176 | U_kgDOEEQsCA | User | false |
openclaw/openclaw | 4,361,496,094 | I_kwDOQb6kR88AAAABA_cmHg | 75,277 | https://github.com/openclaw/openclaw/issues/75277 | https://api.github.com/repos/openclaw/openclaw/issues/75277 | QMD memory search may surface generated dreaming reports as ordinary memory results | ### Summary
I wanted to raise a question about the interaction between QMD-backed memory search and dreaming-generated report files.
This may be expected behavior, but I noticed that when QMD is used as the memory backend and default memory indexing includes `memory/**/*.md`, generated dreaming reports under paths su... | closed | not_planned | true | 1 | [] | [] | 2026-04-30T21:55:09Z | 2026-05-03T04:34:29Z | 2026-05-01T01:14:45Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | wit-oc | 258,550,756 | U_kgDOD2kr5A | User | false |
openclaw/openclaw | 4,362,157,912 | I_kwDOQb6kR88AAAABBAE_WA | 75,321 | https://github.com/openclaw/openclaw/issues/75321 | https://api.github.com/repos/openclaw/openclaw/issues/75321 | [Bug]: `write` tool has no append mode; append-only wrapper is scoped to a memory-flush subsystem and unavailable to general agent runs (cron skills, monitor-and-react workflows) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The standard `write` tool unconditionally calls `fs.writeFile`, overwriting any existing content. Append capability exists in OpenClaw via the `wrapToolMemoryFlushAppendOnlyWrite` wrapper but is reachable *only... | closed | not_planned | true | 1 | [
"bug"
] | [] | 2026-05-01T00:44:35Z | 2026-05-03T04:34:28Z | 2026-05-01T01:14:55Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | camerono | 1,108,491 | MDQ6VXNlcjExMDg0OTE= | User | false |
openclaw/openclaw | 4,362,265,507 | I_kwDOQb6kR88AAAABBALjow | 75,333 | https://github.com/openclaw/openclaw/issues/75333 | https://api.github.com/repos/openclaw/openclaw/issues/75333 | Apply ruff formatting to Python files in skills/ | The `skills/` directory contains Python scripts that need a clean `ruff format` pass to comply with the project's formatting rules. See task #18bcdd2d7657. | closed | completed | true | 2 | [] | [] | 2026-05-01T01:19:35Z | 2026-05-03T04:34:27Z | 2026-05-01T01:24:20Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | jotas6 | 80,299,581 | MDQ6VXNlcjgwMjk5NTgx | User | false |
openclaw/openclaw | 4,362,263,088 | I_kwDOQb6kR88AAAABBALaMA | 75,332 | https://github.com/openclaw/openclaw/issues/75332 | https://api.github.com/repos/openclaw/openclaw/issues/75332 | [Bug]: HTTP 429 from GitHub Copilot is still misclassified as idle timeout on 2026.4.27 | ### Bug type
Regression (or incomplete fix after #71120)
### Beta release blocker
No
### Summary
GitHub Copilot `/responses` returns HTTP 429 for an exhausted weekly quota in under 1 second, but OpenClaw `2026.4.27` still does not surface the provider HTTP error promptly.
Instead, the model stream produces no ass... | closed | completed | true | 2 | [] | [] | 2026-05-01T01:18:45Z | 2026-05-03T04:34:27Z | 2026-05-01T01:23:37Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | zzl360 | 100,331,525 | U_kgDOBfrwBQ | User | false |
openclaw/openclaw | 4,191,067,005 | I_kwDOQb6kR875zpt9 | 59,361 | https://github.com/openclaw/openclaw/issues/59361 | https://api.github.com/repos/openclaw/openclaw/issues/59361 | [Bug]: openclaw agent exits 0 with no output on v2026.3.31; lobster CLI unregistered as command | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
OpenClaw version: 2026.3.31 (213a704)
OS: Ubuntu on AWS Lightsail (headless, no desktop)
Previous working version: 2026.3.2
Bug 1: openclaw agent produces no output
openclaw agent --message "hello" --json exits 0 but outpu... | open | null | false | 2 | [
"bug",
"regression"
] | [] | 2026-04-02T02:46:30Z | 2026-05-01T01:26:07Z | null | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | eddie-schick | 221,229,210 | U_kgDODS-wmg | User | false |
openclaw/openclaw | 4,202,031,228 | I_kwDOQb6kR876deh8 | 60,509 | https://github.com/openclaw/openclaw/issues/60509 | https://api.github.com/repos/openclaw/openclaw/issues/60509 | Feature Request: Expose user message in agent:bootstrap hook context for query-conditioned context compression | ## Summary
The `agent:bootstrap` hook event currently provides `workspaceDir`, `cfg`, `sessionKey`, and `bootstrapFiles` in its context, but **does not include the incoming user message**. Adding the user message (or at minimum a query/topic string) would enable powerful context compression techniques that can reduce ... | open | null | false | 1 | [] | [] | 2026-04-03T20:08:48Z | 2026-05-02T11:08:18Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | MoritzGan | 44,712,740 | MDQ6VXNlcjQ0NzEyNzQw | User | false |
openclaw/openclaw | 4,203,171,087 | I_kwDOQb6kR876h00P | 60,652 | https://github.com/openclaw/openclaw/issues/60652 | https://api.github.com/repos/openclaw/openclaw/issues/60652 | [Bug]: Discord message tool never advertises upload-file/sendAttachment — agents cannot discover file-send capability | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Discord's describeDiscordMessageTool never adds upload-file or sendAttachment to the action set it returns, so the agent's message tool action enum never includes a file-send action — even though the core tool ... | closed | completed | true | 4 | [
"bug",
"bug:behavior"
] | [] | 2026-04-04T03:01:36Z | 2026-05-05T04:28:26Z | 2026-05-02T06:52:52Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | claw-io | 273,482,092 | U_kgDOEE0BbA | User | false |
openclaw/openclaw | 4,204,352,483 | I_kwDOQb6kR876mVPj | 60,780 | https://github.com/openclaw/openclaw/issues/60780 | https://api.github.com/repos/openclaw/openclaw/issues/60780 | Discord thread-bound ACP sessions (mode=session) deliver each response twice | ## Summary
Thread-bound ACP sessions (`sessions_spawn` with `runtime: "acp"`, `thread: true`, `mode: "session"`) send every response **twice** to the Discord thread, ~300–800ms apart. Both messages are identical and delivered via the same webhook persona.
## Environment
- **OpenClaw:** 2026.4.2 (d74a122)
- **acpx:**... | closed | completed | true | 2 | [] | [] | 2026-04-04T09:17:25Z | 2026-05-05T04:28:58Z | 2026-05-02T04:39:10Z | CONTRIBUTOR | null | 20260505T060034Z | 2026-05-05T06:00:34Z | solavrc | 145,330,217 | U_kgDOCKmQKQ | User | false |
openclaw/openclaw | 4,219,166,614 | I_kwDOQb6kR877e1-W | 62,595 | https://github.com/openclaw/openclaw/issues/62595 | https://api.github.com/repos/openclaw/openclaw/issues/62595 | [Bug]: pnpm check fails with TS2883 inferred type portability errors in tsgo | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Running pnpm check on OpenClaw 2026.4.6 fails during pnpm tsgo with TS2883 portability errors that require explicit type annotations in several files.
### Steps to reproduce
1. Check out OpenClaw version 2026... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-07T16:33:56Z | 2026-05-03T04:34:24Z | 2026-05-01T01:27:20Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | WadydX | 65,117,428 | MDQ6VXNlcjY1MTE3NDI4 | User | false |
openclaw/openclaw | 4,205,139,157 | I_kwDOQb6kR876pVTV | 60,917 | https://github.com/openclaw/openclaw/issues/60917 | https://api.github.com/repos/openclaw/openclaw/issues/60917 | Dashboard save fails on redacted gateway.remote.token sentinel | When saving settings from the dashboard (including unrelated sections like AI & Agents), the save can fail with:
`GatewayRequestError: Sentinel value "__OPENCLAW_REDACTED__" in key gateway.remote.token is not valid as real data`
It looks like the dashboard is submitting the redaction sentinel as literal config data i... | open | null | false | 2 | [] | [] | 2026-04-04T15:09:43Z | 2026-05-02T09:29:57Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | sachser66 | 67,191,800 | MDQ6VXNlcjY3MTkxODAw | User | false |
openclaw/openclaw | 4,205,946,680 | I_kwDOQb6kR876sac4 | 61,058 | https://github.com/openclaw/openclaw/issues/61058 | https://api.github.com/repos/openclaw/openclaw/issues/61058 | secrets audit --check flags $VAR env references as PLAINTEXT_FOUND | **Observed behavior:** `openclaw secrets audit --check` flags `${VAR}` env variable references in `openclaw.json` (e.g. `"apiKey": "${OPENAI_API_KEY}"`) at the same severity level as hardcoded plaintext secrets. This produces noisy, unreliable output and makes it impossible to distinguish real findings from safe env re... | open | null | false | 2 | [] | [] | 2026-04-04T22:40:28Z | 2026-05-02T09:29:59Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | interchainlive | 46,060,789 | MDQ6VXNlcjQ2MDYwNzg5 | User | false |
openclaw/openclaw | 4,212,535,470 | I_kwDOQb6kR877FjCu | 61,970 | https://github.com/openclaw/openclaw/issues/61970 | https://api.github.com/repos/openclaw/openclaw/issues/61970 | Bug: WhatsApp outbound video sends succeed but media never appears in chat | ## Summary
WhatsApp outbound video delivery appears to succeed through the `message` tool / gateway, but the recipient never receives any visible video in chat.
## Environment
- OpenClaw: 2026.4.5
- Host: Ubuntu 24.04.3 LTS
- Channel: WhatsApp
- Session type: direct chat
- Gateway healthy, WhatsApp linked
## Repro
1.... | open | null | false | 2 | [] | [] | 2026-04-06T15:37:01Z | 2026-05-02T11:16:30Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | ZenGardenDubai | 399,985 | MDQ6VXNlcjM5OTk4NQ== | User | false |
openclaw/openclaw | 4,216,190,900 | I_kwDOQb6kR877Tfe0 | 62,367 | https://github.com/openclaw/openclaw/issues/62367 | https://api.github.com/repos/openclaw/openclaw/issues/62367 | [Bug] GPT-5.4 nano no longer survives as nano in Codex runtime (2026.4.5) | ## Summary
On OpenClaw 2026.4.5 in our installation, `openai-codex/gpt-5.4-nano` no longer behaves as a stable/real nano route in the Codex runtime.
We see two distinct behaviors depending on the call path:
1. In the current runtime, requesting nano can resolve to `provider=openai-codex` but the runtime reports `mod... | open | null | false | 2 | [] | [] | 2026-04-07T07:36:42Z | 2026-05-02T10:36:42Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | PabloDaVa | 145,587,051 | U_kgDOCK17aw | User | false |
openclaw/openclaw | 4,217,639,421 | I_kwDOQb6kR877ZBH9 | 62,484 | https://github.com/openclaw/openclaw/issues/62484 | https://api.github.com/repos/openclaw/openclaw/issues/62484 | openclaw tui -h causes 100% CPU hang and never exits | ## Bug
`openclaw tui -h` hangs indefinitely with 100% CPU. Ctrl+C does not work; only `kill -9` terminates it.
## Root Cause
`readCliBannerTaglineMode()` in `src/cli/banner-config-lite.ts` statically imports `createConfigIO` from the heavy config module (~19K lines). When Commander generates help text synchronously ... | open | null | false | 1 | [] | [] | 2026-04-07T12:28:11Z | 2026-05-02T10:22:13Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | VibeSparkingAI | 218,262,388 | U_kgDODQJrdA | User | false |
openclaw/openclaw | 4,218,138,271 | I_kwDOQb6kR877a66f | 62,529 | https://github.com/openclaw/openclaw/issues/62529 | https://api.github.com/repos/openclaw/openclaw/issues/62529 | [Feature]: Anthropic is still defualt in docs/start/openclaw.md example | ### Summary
docs/start/openclaw.md:120 uses anthropic/claude-opus-4-6 in the getting-started config snippet. Same issue as the README — new users get pointed to a hostile provider on first setup.
Since we decided to be "model-agnostic" I think this policy decision should be put throughout the repo.
### Problem to sol... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-07T13:52:41Z | 2026-05-02T11:16:32Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | MadanChaollaPark | 164,869,903 | U_kgDOCdO3Dw | User | false |
openclaw/openclaw | 4,219,394,801 | I_kwDOQb6kR877ftrx | 62,615 | https://github.com/openclaw/openclaw/issues/62615 | https://api.github.com/repos/openclaw/openclaw/issues/62615 | [Feature]: Add gateway-side circuit breaker for unhealthy sessions after consecutive failures | ### Summary
Please add a gateway-side circuit breaker for unhealthy sessions so OpenClaw stops retrying the same failing session after a bounded number of consecutive failures.
### Problem to solve
Persistent sessions are great until a session becomes unhealthy because of:
- oversized accumulated context
- repeated ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-04-07T17:14:02Z | 2026-05-02T11:08:33Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | SurlyPlatypus | 74,913,489 | MDQ6VXNlcjc0OTEzNDg5 | User | false |
openclaw/openclaw | 4,220,594,136 | I_kwDOQb6kR877kSfY | 62,708 | https://github.com/openclaw/openclaw/issues/62708 | https://api.github.com/repos/openclaw/openclaw/issues/62708 | refactor: remove dead claudeCliSessionId legacy field from SessionEntry | ## Summary
`claudeCliSessionId` on `SessionEntry` is a dead legacy field. It was introduced in `9a8fe4d68` (`feat(agent): add claude cli runner`, Jan 9 2026) when Claude CLI was the only CLI backend. The next day, `d8f1124d5` (`feat: add CLI backend fallback`) introduced `cliSessionIds: Record<string, string>` as the ... | open | null | false | 1 | [] | [] | 2026-04-07T21:06:11Z | 2026-05-02T10:22:15Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | alexey-pelykh | 345,102 | MDQ6VXNlcjM0NTEwMg== | User | false |
openclaw/openclaw | 4,221,118,523 | I_kwDOQb6kR877mSg7 | 62,751 | https://github.com/openclaw/openclaw/issues/62751 | https://api.github.com/repos/openclaw/openclaw/issues/62751 | session:patch hook event not fired on model changes via /model or agent tool | ## Summary
When a session model is changed via the `/model` slash command or any agent-internal path (e.g. `session_status` tool), **no `session:patch` hook event fires** in the gateway. Workspace hooks subscribed to `session:patch` never trigger during normal model switches.
## Steps to reproduce
1. Register a work... | open | null | false | 1 | [] | [] | 2026-04-07T23:21:27Z | 2026-05-02T09:30:14Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | alphalphi101-cell | 269,280,451 | U_kgDOEAzkww | User | false |
openclaw/openclaw | 4,221,214,968 | I_kwDOQb6kR877mqD4 | 62,765 | https://github.com/openclaw/openclaw/issues/62765 | https://api.github.com/repos/openclaw/openclaw/issues/62765 | msteams dmPolicy=pairing silently drops unpaired senders with HTTP 200, no log line, no auto-reply | ## Bug: msteams `dmPolicy=pairing` silently drops unpaired senders with HTTP 200, no log line, no auto-reply
### Environment
- OpenClaw `2026.4.5 (3e72c03)`
- Ubuntu 24.04.4 LTS, Node v24.14.1 via nvm
- `@microsoft/teams.apps` SDK (bundled)
- single-tenant Bot Framework registration in Entra (`MsaAppType: SingleTenan... | open | null | false | 2 | [] | [] | 2026-04-07T23:49:33Z | 2026-05-02T11:08:35Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | tmote | 21,976,980 | MDQ6VXNlcjIxOTc2OTgw | User | false |
openclaw/openclaw | 4,221,221,728 | I_kwDOQb6kR877mrtg | 62,769 | https://github.com/openclaw/openclaw/issues/62769 | https://api.github.com/repos/openclaw/openclaw/issues/62769 | [Feature]: Support ACP configured bindings for Telegram DMs (not just groups/topics) | ## Summary
Telegram ACP configured bindings (`bindings[]` with `type: "acp"`) only work for group chats and forum topics, not for direct messages (DMs). This prevents using persistent ACP sessions in Telegram DM conversations with dedicated agent bots.
## Root Cause
In `normalizeTelegramAcpConversationId()`, the fun... | open | null | false | 1 | [] | [] | 2026-04-07T23:51:55Z | 2026-05-02T11:08:36Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | richardmqq | 14,178,267 | MDQ6VXNlcjE0MTc4MjY3 | User | false |
openclaw/openclaw | 4,221,345,340 | I_kwDOQb6kR877nJ48 | 62,780 | https://github.com/openclaw/openclaw/issues/62780 | https://api.github.com/repos/openclaw/openclaw/issues/62780 | Feature: message:before_send hook to enable content-quality fallback gating | ## Problem
OpenClaw's fallback chain (`agents.defaults.model.fallbacks`) only triggers on HTTP-level errors (timeouts, 4xx/5xx, rate limits). When the primary model returns `200 OK` with empty or near-empty content (`finish_reason: "stop"`, body = `""` or whitespace), the response is delivered as-is with no fallback a... | open | null | false | 1 | [] | [] | 2026-04-08T00:30:28Z | 2026-05-02T10:36:45Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | jeremyf327 | 23,111,483 | MDQ6VXNlcjIzMTExNDgz | User | false |
openclaw/openclaw | 4,221,485,604 | I_kwDOQb6kR877nsIk | 62,809 | https://github.com/openclaw/openclaw/issues/62809 | https://api.github.com/repos/openclaw/openclaw/issues/62809 | Recurring isolated cron jobs have unstable first-turn OpenAI prompt-cache hits across runs | ## Summary
Prompt caching on `openai/*` / `openai-codex/*` appears to work well *within* a single run, but recurring `sessionTarget: "isolated"` cron jobs still show unstable first-turn cache hits across runs, even when the job prompt and static file inputs are effectively unchanged and `cacheRetention: "long"` is ena... | open | null | false | 1 | [] | [] | 2026-04-08T01:19:30Z | 2026-05-02T11:16:36Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | lubao515 | 14,576,356 | MDQ6VXNlcjE0NTc2MzU2 | User | false |
openclaw/openclaw | 4,221,535,812 | I_kwDOQb6kR877n4ZE | 62,820 | https://github.com/openclaw/openclaw/issues/62820 | https://api.github.com/repos/openclaw/openclaw/issues/62820 | [Bug]: Models generating text-based tool calls send raw XML to messaging channels | ## Bug Description
When models generate tool calls as TEXT (XML format like `<minimax:tool_call><invoke name="exec">...` or MiniMax's `<invoke>...</invoke>`) instead of proper structured tool calls (`tool_calls[]`), the raw XML syntax is sent directly to users on messaging channels like WhatsApp.
## Steps to Reproduc... | closed | completed | true | 3 | [] | [] | 2026-04-08T01:39:06Z | 2026-05-05T04:28:09Z | 2026-05-02T08:16:30Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | canh0chua | 252,370,775 | U_kgDODwrfVw | User | false |
openclaw/openclaw | 4,221,557,670 | I_kwDOQb6kR877n9um | 62,827 | https://github.com/openclaw/openclaw/issues/62827 | https://api.github.com/repos/openclaw/openclaw/issues/62827 | /activate mention command not recognized (only /activation works) | ## Bug Description
The `/activate mention` command (and `/activate always`) doesn't work in WhatsApp group chats because the command parser only recognizes `/activation` but not the shorter alias `/activate`.
## Steps to Reproduce
1. Use WhatsApp group chat with OpenClaw onboarded
2. Send the message `/activate ment... | open | null | false | 2 | [] | [] | 2026-04-08T01:46:58Z | 2026-05-02T10:36:47Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | canh0chua | 252,370,775 | U_kgDODwrfVw | User | false |
openclaw/openclaw | 4,221,687,645 | I_kwDOQb6kR877oddd | 62,851 | https://github.com/openclaw/openclaw/issues/62851 | https://api.github.com/repos/openclaw/openclaw/issues/62851 | [Bug] arcee/trinity-large-thinking works as subagent but fails as main session, suggesting lane/path mismatch | ## Summary
`arcee/trinity-large-thinking` appears to work in a **subagent** context but fails when used as the **main session** model. Please fix the lane mismatch.
This is the core product bug:
- same configured model/provider,
- same installation,
- same machine,
- but **subagent survives while main session bricks**... | open | null | false | 2 | [] | [] | 2026-04-08T02:29:52Z | 2026-05-02T10:22:17Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | Adam-Researchh | 261,827,313 | U_kgDOD5sq8Q | User | false |
openclaw/openclaw | 4,362,274,024 | I_kwDOQb6kR88AAAABBAME6A | 75,334 | https://github.com/openclaw/openclaw/issues/75334 | https://api.github.com/repos/openclaw/openclaw/issues/75334 | [Bug]: commitments extractor uses direct OpenAI instead of configured openai-codex model, causing repeated background auth failures | ### Bug type
Behavior bug (background task uses wrong model/provider and creates repeated auth/model errors)
### Beta release blocker
Possibly — this can silently create repeated background LLM failures after enabling the commitments feature.
### Summary
After enabling the commitments feature on an OpenClaw instal... | closed | completed | true | 2 | [] | [] | 2026-05-01T01:22:24Z | 2026-05-03T04:34:21Z | 2026-05-01T01:57:23Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | sene1337 | 261,265,124 | U_kgDOD5KW5A | User | false |
openclaw/openclaw | 4,362,373,332 | I_kwDOQb6kR88AAAABBASI1A | 75,355 | https://github.com/openclaw/openclaw/issues/75355 | https://api.github.com/repos/openclaw/openclaw/issues/75355 | Bug: Circuit breaker blocks tool but not agent turn; exec loop only warns, never blocks | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
When the global circuit breaker triggers, it blocks the tool execution but does not terminate the agent turn. The agent continues to retry indefinitely. Additionally, `exec` tool loops only produce warnings and never reach c... | closed | not_planned | true | 1 | [
"bug",
"regression"
] | [] | 2026-05-01T01:58:10Z | 2026-05-03T04:34:20Z | 2026-05-01T02:02:26Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | hellofreud | 17,267,947 | MDQ6VXNlcjE3MjY3OTQ3 | User | false |
openclaw/openclaw | 4,217,071,523 | I_kwDOQb6kR877W2ej | 62,436 | https://github.com/openclaw/openclaw/issues/62436 | https://api.github.com/repos/openclaw/openclaw/issues/62436 | Feature: Lightweight LLM passthrough mode for /v1/chat/completions — skip session persistence entirely | ## Summary
When using OpenClaw as a **unified LLM gateway** (model routing + fallback across multiple providers), the `/v1/chat/completions` endpoint unconditionally creates session files on every request — even when the caller explicitly wants stateless, one-shot LLM calls. This makes OpenClaw unsuitable as a lightwe... | open | null | false | 1 | [] | [] | 2026-04-07T10:33:16Z | 2026-05-02T11:36:03Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | gucasbrg | 3,182,658 | MDQ6VXNlcjMxODI2NTg= | User | false |
openclaw/openclaw | 4,217,604,150 | I_kwDOQb6kR877Y4g2 | 62,475 | https://github.com/openclaw/openclaw/issues/62475 | https://api.github.com/repos/openclaw/openclaw/issues/62475 | feat: prompt cache keep-warm pings to prevent TTL expiry between turns | ## Problem
Anthropic prompt cache expires after ~5 min (`cacheRetention: "short"`) or ~1 hour (`cacheRetention: "long"`). Between user messages, when no API call occurs, the cache goes cold. The next turn re-caches the entire system prompt prefix + conversation history at **write pricing** ($3.75/MTok) instead of read... | open | null | false | 1 | [] | [] | 2026-04-07T12:21:22Z | 2026-05-02T12:37:02Z | null | CONTRIBUTOR | null | 20260502T150030Z | 2026-05-02T15:00:30Z | 100yenadmin | 239,388,517 | U_kgDODkTHZQ | User | false |
openclaw/openclaw | 4,218,555,619 | I_kwDOQb6kR877cgzj | 62,553 | https://github.com/openclaw/openclaw/issues/62553 | https://api.github.com/repos/openclaw/openclaw/issues/62553 | Bug: Spontaneous message delivered without user trigger after cache-ttl event | # Bug Report: Spontaneous "test" Message Delivered Without User Trigger
**Date observed:** 2026-04-07 08:15 CDT (13:15:24 UTC)
**OpenClaw version:** 2026.4.2 (d74a122)
**Platform:** Linux 6.8.0-106-generic (x64), Node v25.6.0
**Channel:** Discord
**Agent:** main
---
## Summary
A Discord channel received a... | open | null | false | 1 | [] | [] | 2026-04-07T14:55:48Z | 2026-05-02T12:12:51Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | adampaulsen | 1,534,261 | MDQ6VXNlcjE1MzQyNjE= | User | false |
openclaw/openclaw | 4,218,664,195 | I_kwDOQb6kR877c7UD | 62,560 | https://github.com/openclaw/openclaw/issues/62560 | https://api.github.com/repos/openclaw/openclaw/issues/62560 | Bug: SSRF guard blocks ALL external URLs when using local proxy (affects Telegram media/WeChat/web_fetch) | ## Bug: SSRF guard blocks ALL external URLs when using local proxy (Clash/Tunnel)
### Issue Description
When OpenClaw is configured with a local proxy (e.g., Clash on port 7890), the SSRF guard blocks nearly ALL external URL fetches, making core functionality unusable.
Root cause: Clash's fake-IP DNS returns private... | open | null | false | 3 | [] | [] | 2026-04-07T15:12:12Z | 2026-05-03T20:27:15Z | null | NONE | null | 20260503T210034Z | 2026-05-03T21:00:34Z | Augusta-waveryu | 255,800,970 | U_kgDODz82ig | User | false |
openclaw/openclaw | 4,219,232,073 | I_kwDOQb6kR877fF9J | 62,599 | https://github.com/openclaw/openclaw/issues/62599 | https://api.github.com/repos/openclaw/openclaw/issues/62599 | [Bug]: openclaw status loads memory plugins locally and can report false vector state | ## Bug type
Behavior bug (incorrect output/state without crash)
## Beta release blocker
No
## Summary
`openclaw status` on 2026.4.2 initializes the active memory plugin in the CLI process, which can trigger plugin startup side effects and report false memory/vector state that disagrees with the live launchd-managed g... | open | null | false | 2 | [] | [] | 2026-04-07T16:46:34Z | 2026-05-02T12:53:46Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | shawnpetros | 9,226,624 | MDQ6VXNlcjkyMjY2MjQ= | User | false |
openclaw/openclaw | 4,219,469,555 | I_kwDOQb6kR877f_7z | 62,629 | https://github.com/openclaw/openclaw/issues/62629 | https://api.github.com/repos/openclaw/openclaw/issues/62629 | [Feature]: Strip custom suffix from model key before sending to provider API | ## Problem
In `agents.defaults.models`, users need to define multiple entries for the same model with different parameters (e.g. different `reasoning.effort` levels). Since JSON requires unique keys, users must add a suffix to the model key (e.g. `openai-codex/gpt-5.3-codex:med`). However, the gateway sends the full k... | open | null | false | 1 | [] | [] | 2026-04-07T17:28:08Z | 2026-05-02T11:36:05Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | markbamforth-dev1 | 263,048,425 | U_kgDOD63M6Q | User | false |
openclaw/openclaw | 4,220,835,259 | I_kwDOQb6kR877lNW7 | 62,725 | https://github.com/openclaw/openclaw/issues/62725 | https://api.github.com/repos/openclaw/openclaw/issues/62725 | Per-agent typingMode support | ## Feature Request
Allow `typingMode` (and `typingIntervalSeconds`) to be configured per-agent via `agents.list[]`, not just globally at `agents.defaults`.
### Use case
I have a "buddy" agent that responds on behalf of my user on WhatsApp to a coworker. This agent should have `typingMode: "message"` to avoid immediat... | open | null | false | 1 | [] | [] | 2026-04-07T22:02:49Z | 2026-05-02T12:12:54Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | Jonathan-Asher | 23,320,490 | MDQ6VXNlcjIzMzIwNDkw | User | false |
openclaw/openclaw | 4,221,213,722 | I_kwDOQb6kR877mpwa | 62,763 | https://github.com/openclaw/openclaw/issues/62763 | https://api.github.com/repos/openclaw/openclaw/issues/62763 | openclaw doctor --fix corrupts signal multi-account config by inventing a phantom accounts.default block | ## Bug: `openclaw doctor --fix` corrupts signal multi-account config by inventing a phantom `accounts.default` block
### Environment
- OpenClaw `2026.4.5 (3e72c03)`
- Ubuntu 24.04.4 LTS, Node v24.14.1 (via nvm)
- signal-cli 0.14.2 (native build) at `/usr/local/bin/signal-cli`
- Two registered Signal accounts in multi... | open | null | false | 2 | [] | [] | 2026-04-07T23:49:13Z | 2026-05-02T11:36:07Z | null | NONE | null | 20260502T120029Z | 2026-05-02T12:00:29Z | tmote | 21,976,980 | MDQ6VXNlcjIxOTc2OTgw | User | false |
openclaw/openclaw | 4,221,395,543 | I_kwDOQb6kR877nWJX | 62,784 | https://github.com/openclaw/openclaw/issues/62784 | https://api.github.com/repos/openclaw/openclaw/issues/62784 | [Bug] Multi-agent Slack: concurrent Socket Mode connections cause pong timeout storm and gateway SIGKILL | ## Environment
- OpenClaw version: 2026.4.5
- Node: v25.6.1
- OS: macOS Darwin 23.6.0 (x64)
- Slack plugin: Socket Mode, one app per agent
## Problem
When running multiple agents (8+) each with their own Slack app and Socket Mode WebSocket connection, the gateway crashes repeatedly (20+ times/day) with SIGKILL (exit=... | open | null | false | 1 | [] | [] | 2026-04-08T00:44:18Z | 2026-05-02T12:37:03Z | null | NONE | null | 20260502T150030Z | 2026-05-02T15:00:30Z | CHE10X | 67,395,160 | MDQ6VXNlcjY3Mzk1MTYw | User | false |
openclaw/openclaw | 4,362,416,204 | I_kwDOQb6kR88AAAABBAUwTA | 75,359 | https://github.com/openclaw/openclaw/issues/75359 | https://api.github.com/repos/openclaw/openclaw/issues/75359 | [Bug]: Discord DM heartbeats erroneously resolve `user:<snowflake>` to `channel:<snowflake>` | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Heartbeat messages targeting discord DMs fail permanently. Attached is a full bug report generated by kimi-code (who found the bug as well).
[openclaw-discord-dm-heartbeat-bug-report.md](https://github.com/us... | closed | completed | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-05-01T02:10:57Z | 2026-05-03T04:34:17Z | 2026-05-01T02:17:58Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | luna-system | 183,767 | MDQ6VXNlcjE4Mzc2Nw== | User | false |
openclaw/openclaw | 4,362,471,965 | I_kwDOQb6kR88AAAABBAYKHQ | 75,365 | https://github.com/openclaw/openclaw/issues/75365 | https://api.github.com/repos/openclaw/openclaw/issues/75365 | Test issue from xiaobai-khhome | Testing write access | closed | completed | true | 0 | [] | [] | 2026-05-01T02:30:30Z | 2026-05-03T04:34:12Z | 2026-05-01T02:31:00Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | xiaobai-khhome | 262,472,254 | U_kgDOD6UCPg | User | false |
openclaw/openclaw | 4,362,472,890 | I_kwDOQb6kR88AAAABBAYNug | 75,366 | https://github.com/openclaw/openclaw/issues/75366 | https://api.github.com/repos/openclaw/openclaw/issues/75366 | [Linux] Browser profile locked after Z2 Mini Chrome crash | ## Issue Description
After migrating from Z2 Mini to Z8G4, the Chrome profile locks when Chrome crashes on Z2 Mini, causing "profile in use" errors on Z8G4.
## Steps to Reproduce
1. Run OpenClaw on Z2 Mini with Chrome open
2. Machine loses network or Chrome crashes
3. After restarting on Z8G4, Chrome refuses to launch... | closed | completed | true | 1 | [] | [] | 2026-05-01T02:30:55Z | 2026-05-03T04:34:12Z | 2026-05-01T02:34:41Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | xiaobai-khhome | 262,472,254 | U_kgDOD6UCPg | User | false |
openclaw/openclaw | 4,362,491,711 | I_kwDOQb6kR88AAAABBAZXPw | 75,373 | https://github.com/openclaw/openclaw/issues/75373 | https://api.github.com/repos/openclaw/openclaw/issues/75373 | [Bug]: MiniMax fetch blocks Node.js event loop for ~135s → watchdog SIGTERM (regression in v4.27) | ## Environment
- OpenClaw version: v2026.4.27 (regression; v2026.4.26 unaffected)
- OS: Ubuntu 24.04 LTS, Linux kernel 6.17
- Node.js: default bundled
- Primary model: minimax/MiniMax-M2.7
- Deployment: systemd service, loopback-only, Telegram channel
## Symptom
After upgrading from v4.26 → v4.27, the gateway enters... | closed | not_planned | true | 3 | [] | [] | 2026-05-01T02:38:36Z | 2026-05-03T04:34:11Z | 2026-05-01T02:42:38Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | xiaobai-khhome | 262,472,254 | U_kgDOD6UCPg | User | false |
openclaw/openclaw | 4,362,604,201 | I_kwDOQb6kR88AAAABBAgOqQ | 75,386 | https://github.com/openclaw/openclaw/issues/75386 | https://api.github.com/repos/openclaw/openclaw/issues/75386 | [bug] sessions.json write lock held 26-28s under multi-subagent fan-out; stuck sessions never reclaimed | ## Summary
Under multi-subagent fan-out, the `sessions.json` write lock becomes a serialization bottleneck. Lock holds repeatedly exceed the configured 15s ceiling (observed 26-28 seconds), which serializes every `node.list` / `sessions.list` call behind a single writer. Concurrently, subagent sessions transition to `... | closed | not_planned | true | 1 | [] | [] | 2026-05-01T03:18:07Z | 2026-05-03T04:34:07Z | 2026-05-01T03:22:25Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Jason-Vaughan | 95,194,903 | U_kgDOBayPFw | User | false |
openclaw/openclaw | 4,362,603,635 | I_kwDOQb6kR88AAAABBAgMcw | 75,385 | https://github.com/openclaw/openclaw/issues/75385 | https://api.github.com/repos/openclaw/openclaw/issues/75385 | Bug: Messages appear twice in webchat transcript | ## Bug Description
Messages appear duplicated in the webchat transcript when sending a message.
## Steps to Reproduce
1. Send a message in webchat
2. Observe the message appears twice in the dialog
3. Check transcript file - message is written twice
## Expected Behavior
Each message should appear only once in the ... | closed | completed | true | 1 | [] | [] | 2026-05-01T03:17:56Z | 2026-05-03T04:34:06Z | 2026-05-01T03:23:04Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | luangteng | 101,879,109 | U_kgDOBhKNRQ | User | false |
openclaw/openclaw | 4,362,659,532 | I_kwDOQb6kR88AAAABBAjmzA | 75,405 | https://github.com/openclaw/openclaw/issues/75405 | https://api.github.com/repos/openclaw/openclaw/issues/75405 | [Bug]: 2026.4.27 regression: slash commands silently dropped in Telegram groups bound to non-main agents | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading from 2026.4.23 → 2026.4.27, native slash commands (`/status@<bot>`, `/new@<bot>`, `/model@<bot>`) are silently dropped when sent in Telegram groups bound to
non-`main` agents via `bindings`. The same... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-05-01T03:38:05Z | 2026-05-05T04:27:28Z | 2026-05-02T11:47:26Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | ziptbm | 46,136,563 | MDQ6VXNlcjQ2MTM2NTYz | User | false |
openclaw/openclaw | 4,362,668,640 | I_kwDOQb6kR88AAAABBAkKYA | 75,408 | https://github.com/openclaw/openclaw/issues/75408 | https://api.github.com/repos/openclaw/openclaw/issues/75408 | core-plugin-tools 15s per-turn cost on high-RTT hosts traced to per-turn Discord/Telegram describe-tool probes (#75290 + #75341 + #73428) | ## Summary
On openclaw 2026.4.29, users on geographically-distant VPS hosts (high RTT to Discord/Telegram API endpoints) experience a consistent 17–37s pre-stream latency on **every** user turn. Stage-level instrumentation pinpoints the cost to the `core-plugin-tools` prep stage (~15s) plus `model-resolution` (~10s) a... | closed | not_planned | true | 1 | [] | [] | 2026-05-01T03:41:13Z | 2026-05-03T04:34:05Z | 2026-05-01T03:46:10Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | jackiedepp | 31,266,295 | MDQ6VXNlcjMxMjY2Mjk1 | User | false |
openclaw/openclaw | 4,362,670,087 | I_kwDOQb6kR88AAAABBAkQBw | 75,409 | https://github.com/openclaw/openclaw/issues/75409 | https://api.github.com/repos/openclaw/openclaw/issues/75409 | [Bug]: [Control UI] New session fails with "messages must not be empty" when using /new command | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Bug Description
When creating a new session via Control UI using the `/new` command, the session immediately fails with:
### Steps to reproduce
This happens **every time** — 100% reproducible on fresh sessions via Cont... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-05-01T03:41:45Z | 2026-05-03T04:34:05Z | 2026-05-01T03:46:26Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | Leo7Guo | 34,794,155 | MDQ6VXNlcjM0Nzk0MTU1 | User | false |
openclaw/openclaw | 4,362,672,539 | I_kwDOQb6kR88AAAABBAkZmw | 75,410 | https://github.com/openclaw/openclaw/issues/75410 | https://api.github.com/repos/openclaw/openclaw/issues/75410 | 2026.4.29: Discord replies take 70-85s despite lean config | # 2026.4.29: Discord replies take 70-85s despite lean config; per-turn startup/prep stalls
## Summary
After upgrading OpenClaw from 2026.4.26 to 2026.4.29, Jarvis/main Discord-visible replies are consistently slow, typically around 70-85 seconds for a simple "hi" or "ping". This persists after narrowing the runtime t... | closed | completed | true | 1 | [] | [] | 2026-05-01T03:42:50Z | 2026-05-03T04:34:04Z | 2026-05-01T03:47:03Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | ramitrkar-hash | 280,062,852 | U_kgDOELFrhA | User | false |
openclaw/openclaw | 4,362,623,764 | I_kwDOQb6kR88AAAABBAhbFA | 75,394 | https://github.com/openclaw/openclaw/issues/75394 | https://api.github.com/repos/openclaw/openclaw/issues/75394 | [Bug]: Failed to render WhatsApp QR - Cannot find package ‘qrcode’ in 2026.4.29 | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Failed to render the WhatsApp QR: Cannot find package ‘qrcode’ imported from /data/.openclaw/plugin-runtime-deps/openclaw-2026.4.29-c27ae31043c7/dist/qr-terminal-DI6mj_pL.js
////////
<img width="638" height="388" alt="Imag... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-05-01T03:25:49Z | 2026-05-04T04:36:42Z | 2026-05-01T11:37:51Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | FelipeX2001 | 146,594,144 | U_kgDOCLzZYA | User | false |
openclaw/openclaw | 4,359,349,901 | I_kwDOQb6kR88AAAABA9ZmjQ | 75,131 | https://github.com/openclaw/openclaw/issues/75131 | https://api.github.com/repos/openclaw/openclaw/issues/75131 | [Bug]: delivery-queue: send-retry creates fresh UUIDs (not idempotent); recovery should fail-permanent on Telegram 400 'too long'; send-message should pre-split at 4096 chars | ## Summary
Three related defects in the Telegram delivery path that compound into a persistent queue jam and measurable gateway-wide event-loop saturation:
1. **Send-retry is not idempotent at the queue layer.** Each retry creates a fresh queue UUID instead of incrementing `retryCount` on the existing one.
2. **`deli... | open | null | false | 1 | [] | [] | 2026-04-30T15:35:58Z | 2026-05-03T08:26:31Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | solosage1 | 158,075,162 | U_kgDOCWwJGg | User | false |
openclaw/openclaw | 4,359,483,659 | I_kwDOQb6kR88AAAABA9hxCw | 75,151 | https://github.com/openclaw/openclaw/issues/75151 | https://api.github.com/repos/openclaw/openclaw/issues/75151 | [Bug]: Context overflow reset can map sessionFile to nonexistent transcript, orphaning real session history | ## Summary
A long-lived Telegram direct agent session hit context-overflow / compaction failure, then the active `sessions.json` mapping was rotated to a new `sessionId` whose transcript file was never created. As a result, `chat.history` / session lookup returned an empty history even though the real prior transcript... | open | null | false | 1 | [] | [] | 2026-04-30T15:57:19Z | 2026-05-03T11:50:33Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | MickeMG | 24,614,114 | MDQ6VXNlcjI0NjE0MTE0 | User | false |
openclaw/openclaw | 4,359,503,277 | I_kwDOQb6kR88AAAABA9i9rQ | 75,153 | https://github.com/openclaw/openclaw/issues/75153 | https://api.github.com/repos/openclaw/openclaw/issues/75153 | [Feature]: expose channels.start / stop / restart via CLI so a wedged channel can recover without container restart or QR re-pair | ## TL;DR
When a channel transiently fails (e.g., DNS blip → `EAI_AGAIN web.whatsapp.com` → 408/428), it lands in `stopped, disconnected` state and **does not auto-recover** even after the network is healthy again. The gateway implements `channels.start` as a JSON-RPC method, and that method works perfectly — but it is... | open | null | false | 1 | [] | [] | 2026-04-30T16:00:30Z | 2026-05-03T01:23:28Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | Jason-Vaughan | 95,194,903 | U_kgDOBayPFw | User | false |
openclaw/openclaw | 4,359,521,991 | I_kwDOQb6kR88AAAABA9kGxw | 75,154 | https://github.com/openclaw/openclaw/issues/75154 | https://api.github.com/repos/openclaw/openclaw/issues/75154 | Question: should openai-codex-responses be in shouldRepairMalformedToolCallArguments? | ## Context
I'm running GPT-5.5 via the `openai-codex` provider (ChatGPT Plus OAuth, `openai-codex-responses` API) on 2026.4.27 and tracing what tool-call self-repair is active for that path. The current gate excludes my transport:
```ts
// src/agents/pi-embedded-runner/run/attempt.tool-call-argument-repair.ts:297-306... | closed | completed | true | 1 | [] | [] | 2026-04-30T16:03:43Z | 2026-05-05T04:27:40Z | 2026-05-02T10:42:33Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | keane-751892 | 252,737,230 | U_kgDODxB2zg | User | false |
openclaw/openclaw | 4,359,538,657 | I_kwDOQb6kR88AAAABA9lH4Q | 75,155 | https://github.com/openclaw/openclaw/issues/75155 | https://api.github.com/repos/openclaw/openclaw/issues/75155 | Talk Mode: user voice transcripts don't render in WebChat/Control UI thread | ## Summary
When using Talk Mode on macOS, my spoken messages are correctly transcribed and sent to the model (the assistant replies appropriately, both in text and TTS). However, the **user side** of the conversation never appears in the WebChat / Control UI chat pane — only assistant replies show up.
## Expected
Us... | closed | completed | true | 1 | [] | [] | 2026-04-30T16:06:36Z | 2026-05-04T04:34:13Z | 2026-05-02T03:29:58Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | SledderBling | 279,591,126 | U_kgDOEKo41g | User | false |
openclaw/openclaw | 4,360,061,872 | I_kwDOQb6kR88AAAABA-FDsA | 75,184 | https://github.com/openclaw/openclaw/issues/75184 | https://api.github.com/repos/openclaw/openclaw/issues/75184 | [Bug]: `agents.defaults.skipBootstrap: true` is a no-op for workspace bootstrap files | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Setting `agents.defaults.skipBootstrap: true` in `openclaw.json` does not skip workspace bootstrap file injection; the systemPromptReport for subsequent agent runs still reports all bootstrap files (AGENTS.md, ... | open | null | false | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-04-30T17:32:50Z | 2026-05-03T04:44:30Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | camerono | 1,108,491 | MDQ6VXNlcjExMDg0OTE= | User | false |
openclaw/openclaw | 4,360,349,195 | I_kwDOQb6kR88AAAABA-WmCw | 75,203 | https://github.com/openclaw/openclaw/issues/75203 | https://api.github.com/repos/openclaw/openclaw/issues/75203 | [Bug]: Control UI webchat routes user input into :heartbeat-suffixed session when main session is pruned | ## Summary
When aggressive session pruning removes `agent:main:main` from sessions, the Control UI session picker no longer shows the primary session. Heartbeat ticks keep `agent:main:main:heartbeat` alive, so that becomes the only visible session in the picker. The webchat input then routes user messages directly int... | open | null | false | 1 | [] | [] | 2026-04-30T18:21:17Z | 2026-05-03T10:41:52Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | JakeSc | 275,063 | MDQ6VXNlcjI3NTA2Mw== | User | false |
openclaw/openclaw | 4,360,536,050 | I_kwDOQb6kR88AAAABA-h_8g | 75,211 | https://github.com/openclaw/openclaw/issues/75211 | https://api.github.com/repos/openclaw/openclaw/issues/75211 | WhatsApp plugin silently ignores route bindings (security-relevant) | ## Bug: WhatsApp plugin silently ignores `bindings` route configuration
**Version:** OpenClaw 2026.4.14
**Plugin:** `whatsapp`
**Severity:** Security-relevant — silently routes ALL WA inbound to default agent regardless of configured bindings
## Summary
The WhatsApp plugin does not implement `compileConfiguredBindin... | open | null | false | 1 | [] | [] | 2026-04-30T18:51:32Z | 2026-05-03T10:20:10Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | juanez2024 | 187,393,810 | U_kgDOCytnEg | User | false |
openclaw/openclaw | 4,361,351,224 | I_kwDOQb6kR88AAAABA_TwOA | 75,268 | https://github.com/openclaw/openclaw/issues/75268 | https://api.github.com/repos/openclaw/openclaw/issues/75268 | [Bug]: [object Object][object Object]... in agents messages and in memory with mistral thinking | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
I get results with several [object Object] in the reply with Mistral, in session files and memory.
I'm sorry to bother you with this issue.
This is exactly issue is the same as https://github.com/openclaw/ope... | open | null | false | 2 | [
"bug",
"bug:behavior"
] | [] | 2026-04-30T21:25:38Z | 2026-05-03T11:50:35Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | fmatray | 8,267,716 | MDQ6VXNlcjgyNjc3MTY= | User | false |
openclaw/openclaw | 4,361,508,676 | I_kwDOQb6kR88AAAABA_dXRA | 75,279 | https://github.com/openclaw/openclaw/issues/75279 | https://api.github.com/repos/openclaw/openclaw/issues/75279 | [Performance]: 3-7 minute agent response latency on Windows — 66 plugins loaded, 49 disabled, 48 skills missing requirements | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
[Performance]: 3-7 minute agent response latency on Windows — 66 plugins loaded, 49 disabled, 48 skills missing requirements
### Steps to reproduce
1. Fresh OpenClaw v2026.4.26 install on Windows 11 via npm install -g opencla... | closed | duplicate | true | 2 | [
"bug",
"bug:crash"
] | [] | 2026-04-30T21:57:59Z | 2026-05-04T04:35:07Z | 2026-05-01T22:26:28Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | Stellarjester315 | 254,353,553 | U_kgDODykgkQ | User | false |
openclaw/openclaw | 4,361,806,071 | I_kwDOQb6kR88AAAABA_vg9w | 75,298 | https://github.com/openclaw/openclaw/issues/75298 | https://api.github.com/repos/openclaw/openclaw/issues/75298 | Webhook plugin re-registers repeatedly whenever event loop saturates (dreaming, cron catchup, Control UI polling) | **Version:** 2026.4.27 (also reproduced on 2026.4.26)
**Setup:** Docker container, Windows 11 host, Compose stack, single webhook route (ai-news)
## Summary
The webhooks plugin re-registers its route repeatedly (5–10x per gateway session) whenever the Node.js event loop saturates. This is not limited to cold start — ... | open | null | false | 1 | [] | [] | 2026-04-30T23:06:40Z | 2026-05-03T08:26:35Z | null | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | balaji1968-kingler | 268,807,278 | U_kgDOEAWsbg | User | false |
openclaw/openclaw | 4,361,848,165 | I_kwDOQb6kR88AAAABA_yFZQ | 75,301 | https://github.com/openclaw/openclaw/issues/75301 | https://api.github.com/repos/openclaw/openclaw/issues/75301 | [Feature]: `openclaw caches` command to inspect and prune unbounded `~/.openclaw/` cache dirs (plugin-runtime-deps, browser, tools, orphan transcripts) | ### Summary
Add an `openclaw caches` command (or `openclaw doctor --prune-caches`) that lists cache-class directories under `~/.openclaw/` with their sizes and what populates them, and lets the user prune safely (stale plugin-runtime-deps version dirs, orphan agent transcripts, unused browser/tools downloads).
### Pr... | open | null | false | 2 | [] | [] | 2026-04-30T23:16:20Z | 2026-05-03T09:49:59Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | mogglemoss | 10,966,195 | MDQ6VXNlcjEwOTY2MTk1 | User | false |
openclaw/openclaw | 4,362,626,247 | I_kwDOQb6kR88AAAABBAhkxw | 75,395 | https://github.com/openclaw/openclaw/issues/75395 | https://api.github.com/repos/openclaw/openclaw/issues/75395 | [Bug]: Claude Opus 4.6 via OpenRouter — "assistant message prefill" Error | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
When using Anthropic Claude Opus 4.6 through **OpenRouter** with the
`openai-completions` API adapter, every agent turn fails with:
> **"This model does not support assistant message prefill. The conversation must end with ... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-05-01T03:26:53Z | 2026-05-05T04:28:37Z | 2026-05-02T05:48:42Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | sbmilburn | 11,755,869 | MDQ6VXNlcjExNzU1ODY5 | User | false |
openclaw/openclaw | 4,362,708,402 | I_kwDOQb6kR88AAAABBAmlsg | 75,414 | https://github.com/openclaw/openclaw/issues/75414 | https://api.github.com/repos/openclaw/openclaw/issues/75414 | Event loop blocking causes WebSocket disconnections and cron failures | ## Bug Description
Synchronous event loop blocking (up to 2581 seconds / 43 minutes) causes QQBot WebSocket disconnections and cron job failures.
## Steps to Reproduce
1. Run OpenClaw gateway with QQBot channel enabled
2. After extended session with large trajectory files (20MB+), observe:
- `eventLoopDelayMaxMs`... | closed | completed | true | 1 | [] | [] | 2026-05-01T03:55:59Z | 2026-05-04T04:36:22Z | 2026-05-01T13:14:38Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | tangda18 | 264,406,045 | U_kgDOD8KEHQ | User | false |
openclaw/openclaw | 4,362,712,795 | I_kwDOQb6kR88AAAABBAm22w | 75,416 | https://github.com/openclaw/openclaw/issues/75416 | https://api.github.com/repos/openclaw/openclaw/issues/75416 | Bug: QQBot streaming + stuck session recovery deadlock causes permanent session freeze | ## Bug Report
### Summary
QQBot streaming replies can cause **permanent session deadlock** due to the interaction between:
1. QQ platform API rate limiting (streaming chunks arrive very slowly)
2. Stuck session detection firing after 120s
3. Recovery being skipped because `active_embedded_run` is still running
4. Sub... | open | null | false | 2 | [] | [] | 2026-05-01T03:57:46Z | 2026-05-03T05:12:07Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | subaochen | 1,390,258 | MDQ6VXNlcjEzOTAyNTg= | User | false |
openclaw/openclaw | 4,362,748,060 | I_kwDOQb6kR88AAAABBApAnA | 75,419 | https://github.com/openclaw/openclaw/issues/75419 | https://api.github.com/repos/openclaw/openclaw/issues/75419 | [Bug]: [2026.4.29] openclaw cannot accept for any input | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
When I upgrade the newest version(2026.4.29), My openclaw becomes unavailable for tui reject any of my input.
### Steps to reproduce
just run `openclaw tui`
### Expected behavior
I can input my prompt for ... | closed | not_planned | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-05-01T04:11:25Z | 2026-05-05T04:28:11Z | 2026-05-02T07:57:52Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | oyoungs | 15,865,312 | MDQ6VXNlcjE1ODY1MzEy | User | false |
openclaw/openclaw | 4,361,008,774 | I_kwDOQb6kR88AAAABA--2hg | 75,251 | https://github.com/openclaw/openclaw/issues/75251 | https://api.github.com/repos/openclaw/openclaw/issues/75251 | [Bug]: Discord-awaiting gateway readiness | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Discord channel: bot stays "awaiting gateway readiness" after gateway restart
### Environment
- OpenClaw version: 2026.4.27
- Platform: macOS (Apple Silicon, M2 Ultra Mac Studio)
- Node: 22.22.0
- Channel: Discord (bundl... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-30T20:16:46Z | 2026-05-03T04:34:02Z | 2026-04-30T20:21:19Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | iamoutthere-sudo | 280,801,321 | U_kgDOELywKQ | User | false |
openclaw/openclaw | 4,362,762,607 | I_kwDOQb6kR88AAAABBAp5bw | 75,420 | https://github.com/openclaw/openclaw/issues/75420 | https://api.github.com/repos/openclaw/openclaw/issues/75420 | Bug: web_search can receive unresolved SecretRef for Gemini provider after 2026.4.29 | ## Summary
After upgrading to OpenClaw `2026.4.29`, managed `web_search` calls can fail with an unresolved SecretRef error for Gemini web search, even though the secret exists and `openclaw secrets audit` reports `unresolved=0`.
This appears to be a runtime/config plumbing issue: at least one web_search execution pat... | closed | completed | true | 2 | [] | [] | 2026-05-01T04:16:59Z | 2026-05-04T04:34:27Z | 2026-05-02T02:29:31Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | richardmqq | 14,178,267 | MDQ6VXNlcjE0MTc4MjY3 | User | false |
openclaw/openclaw | 4,362,771,040 | I_kwDOQb6kR88AAAABBAqaYA | 75,421 | https://github.com/openclaw/openclaw/issues/75421 | https://api.github.com/repos/openclaw/openclaw/issues/75421 | Fallback inserts blank user messages, rejected by providers (ZAI GLM 400 error) | ## Bug Description
When context overflow occurs and auto-compaction doesn't free enough space, OpenClaw falls back to alternate models. During this fallback process, blank user messages are inserted into the message history, which causes some providers (e.g., ZAI/GLM) to reject the request with `400 messages 参数非法`.
#... | closed | completed | true | 1 | [] | [] | 2026-05-01T04:19:51Z | 2026-05-03T04:33:58Z | 2026-05-01T04:24:35Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | king-bot-king | 267,250,520 | U_kgDOD-3rWA | User | false |
openclaw/openclaw | 4,359,498,138 | I_kwDOQb6kR88AAAABA9ipmg | 75,152 | https://github.com/openclaw/openclaw/issues/75152 | https://api.github.com/repos/openclaw/openclaw/issues/75152 | [Bug]: Edit faild, cant correctly use Edit tool | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
Yes
### Summary
The edit tool consistently rejects the edits parameter. Passing it as a JSON string yields edits.0: must be object. Passing it as a native JSON array yields edits.0: must have required properties oldText, newT... | open | null | false | 1 | [
"bug"
] | [] | 2026-04-30T15:59:41Z | 2026-05-03T04:44:29Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | velove52 | 50,912,588 | MDQ6VXNlcjUwOTEyNTg4 | User | false |
openclaw/openclaw | 4,361,086,453 | I_kwDOQb6kR88AAAABA_Dl9Q | 75,255 | https://github.com/openclaw/openclaw/issues/75255 | https://api.github.com/repos/openclaw/openclaw/issues/75255 | Support Xiaomi MiMo native web_search tool type | ## Feature Request
### Summary
Xiaomi MiMo API provides a native `web_search` tool type that performs web searches at the provider level. Currently, OpenClaw does not pass this tool type when making API calls to the Xiaomi provider, so the feature is unused.
### Xiaomi web_search API
The Xiaomi MiMo API supports a sp... | open | null | false | 2 | [] | [] | 2026-04-30T20:31:15Z | 2026-05-03T03:38:09Z | null | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | sosiman | 192,899,825 | U_kgDOC39q8Q | User | false |
openclaw/openclaw | 4,361,911,066 | I_kwDOQb6kR88AAAABA_17Gg | 75,305 | https://github.com/openclaw/openclaw/issues/75305 | https://api.github.com/repos/openclaw/openclaw/issues/75305 | [Bug]: Pre- compaction memory flush sends empty user message to Anthropic API | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
When a session approaches the context limit and auto-compaction triggers, OpenClaw's pre-compaction memory flush constructs a user message with empty string content (""). The Anthropic API rejects this with a 400 error, caus... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-30T23:35:00Z | 2026-05-05T04:28:35Z | 2026-05-02T06:14:28Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | sableassistant3777-source | 278,228,683 | U_kgDOEJVuyw | User | false |
openclaw/openclaw | 4,362,679,140 | I_kwDOQb6kR88AAAABBAkzZA | 75,411 | https://github.com/openclaw/openclaw/issues/75411 | https://api.github.com/repos/openclaw/openclaw/issues/75411 | [Bug]: Copilot 429 fail-fast fix is bypassed by session-custom stream path on 2026.4.29 | ### Bug type
Regression / incomplete fix after #71120
### Beta release blocker
No
### Summary
The `v2026.4.29` fix for #71120 is present and works when OpenClaw routes a model call through `createOpenAIResponsesTransportStreamFn`, but live `github-copilot/gpt-5.4` traffic still bypasses that path.
On `v2026.4.29`... | open | null | false | 1 | [] | [] | 2026-05-01T03:45:32Z | 2026-05-03T10:41:57Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | zzl360 | 100,331,525 | U_kgDOBfrwBQ | User | false |
openclaw/openclaw | 4,362,663,137 | I_kwDOQb6kR88AAAABBAj04Q | 75,406 | https://github.com/openclaw/openclaw/issues/75406 | https://api.github.com/repos/openclaw/openclaw/issues/75406 | v2026.4.29 - synchronous WASM tokenizer init blocks event loop 20-79s on first LLM call (system-prompt / runtime-plugins stage) | On v2026.4.29, the embedded agent's startup trace shows the event loop blocking for 20-79 seconds during the first LLM call after boot. This manifests as eventLoopUtilization=1.0 and causes channel drops (WhatsApp 408, Telegram WebSocket handshake timeouts).
Observed stage timings (first call after boot):
- runtime-pl... | closed | completed | true | 4 | [] | [] | 2026-05-01T03:39:18Z | 2026-05-04T04:36:33Z | 2026-05-01T11:48:08Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | balaji1968-kingler | 268,807,278 | U_kgDOEAWsbg | User | false |
openclaw/openclaw | 4,362,823,603 | I_kwDOQb6kR88AAAABBAtnsw | 75,426 | https://github.com/openclaw/openclaw/issues/75426 | https://api.github.com/repos/openclaw/openclaw/issues/75426 | [Bug]: Signal inbound/SSE listener appears wedged: outbound Signal works, channel probe OK, but inbound DMs do not reliably trigger agent responses and `channels status` times out | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
## Summary
OpenClaw Signal outbound messaging works, but inbound Signal messages do not reliably trigger agent responses.
The Signal channel is configured, the Signal account is registered, `signal-cli` works... | open | null | false | 2 | [
"bug"
] | [] | 2026-05-01T04:37:23Z | 2026-05-03T10:41:58Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | willcharlton | 6,548,926 | MDQ6VXNlcjY1NDg5MjY= | User | false |
openclaw/openclaw | 4,193,723,396 | I_kwDOQb6kR8759yQE | 59,629 | https://github.com/openclaw/openclaw/issues/59629 | https://api.github.com/repos/openclaw/openclaw/issues/59629 | [Bug]: Auth-profiles.json parsing broken in v2026.3.31 (regression from 2026.3.8) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Bug description
**Regression in v2026.3.31:** Agent cannot read `auth-profiles.json`. Same configuration works perfectly in v2026.3.8.
### Steps to reproduce
## Steps to reproduce
```bash
# Create auth file
printf '... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-04-02T10:52:27Z | 2026-05-01T04:45:09Z | 2026-04-28T05:54:50Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Xsanders555 | 123,403,444 | U_kgDOB1r8tA | User | false |
openclaw/openclaw | 4,079,390,685 | I_kwDOQb6kR87zJo_d | 47,674 | https://github.com/openclaw/openclaw/issues/47674 | https://api.github.com/repos/openclaw/openclaw/issues/47674 | [Bug]: Telegram channel shows Running briefly then disconnects — deleteWebhook 401 Unauthorized (valid token confirmed) | ### Bug type
Regression (worked before, now fails)
### Summary
Telegram channel starts (Running = true momentarily), then immediately disconnects. The gateway calls deleteWebhook and Telegram returns 401 Unauthorized, despite the bot token being verified as valid in BotFather multiple times.
### Steps to reproduce
... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-03-15T22:57:52Z | 2026-05-01T04:45:10Z | 2026-04-28T05:51:16Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | samaedan-arch | 268,460,770 | U_kgDOEABi4g | User | false |
openclaw/openclaw | 4,340,500,200 | I_kwDOQb6kR88AAAABArbG6A | 73,242 | https://github.com/openclaw/openclaw/issues/73242 | https://api.github.com/repos/openclaw/openclaw/issues/73242 | 2026.4.26: openai-codex/gpt-5.4-mini listed as available but cron fails with ChatGPT account unsupported | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
**Body:**
```md
OpenClaw 2026.4.26 (be8c246)
After updating to .26, isolated cron jobs using `openai-codex/gpt-5.4-mini` started failing.
`openclaw models list --provider openai-codex` shows:
- openai-codex/gpt-5.4
- open... | closed | completed | true | 1 | [
"bug"
] | [] | 2026-04-28T04:01:42Z | 2026-05-01T04:45:11Z | 2026-04-28T05:44:26Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | 0xCyda | 116,336,463 | U_kgDOBu8nTw | User | false |
openclaw/openclaw | 4,340,303,806 | I_kwDOQb6kR88AAAABArPHvg | 73,198 | https://github.com/openclaw/openclaw/issues/73198 | https://api.github.com/repos/openclaw/openclaw/issues/73198 | Bug: Dreaming narrative spawns unbounded concurrent subagent sessions across workspaces | ## Problem
The dreaming narrative system spawns unbounded concurrent subagent sessions when processing multiple workspaces, causing resource exhaustion, session lock contention, and cascading timeouts.
## Root Cause
In `dreaming.ts`, when `trigger === "cron"`, `detachNarratives` is set to `true`. This causes narrati... | closed | completed | true | 0 | [] | [] | 2026-04-28T03:07:01Z | 2026-05-01T04:45:11Z | 2026-04-28T05:42:08Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | stayif | 7,867,389 | MDQ6VXNlcjc4NjczODk= | User | false |
openclaw/openclaw | 4,147,583,968 | I_kwDOQb6kR873Nxvg | 55,387 | https://github.com/openclaw/openclaw/issues/55387 | https://api.github.com/repos/openclaw/openclaw/issues/55387 | [Bug]:Telegram startup fails with 404 (deleteWebhook / deleteMyCommands / setMyCommands) even though direct Bot API calls succeed | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
### Version / environment
- OpenClaw: 2026.3.24
- OS: macOS 26.2 (arm64)
- Node: 22.22.2
### Summary
Telegram fails during OpenClaw startup with 404 errors and never becomes healthy, even though the configured... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-03-26T21:55:51Z | 2026-05-01T04:45:12Z | 2026-04-28T05:39:07Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | brendanmatthewjones-cmyk | 268,941,102 | U_kgDOEAe3Lg | User | false |
openclaw/openclaw | 4,340,510,013 | I_kwDOQb6kR88AAAABArbtPQ | 73,246 | https://github.com/openclaw/openclaw/issues/73246 | https://api.github.com/repos/openclaw/openclaw/issues/73246 | [Bug]: Control UI: Thinking icon disappears after ~1 minute and conversation goes blank (webchat channel) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
## Bug Description
In QClaw's openclaw-control-ui (desktop client), when using the webchat channel, the thinking icon appears after sending a message, but disappears after approximately 1 minute without any response being r... | closed | completed | true | 1 | [
"bug",
"r: support"
] | [] | 2026-04-28T04:04:50Z | 2026-05-01T04:45:13Z | 2026-04-28T05:36:55Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | keersys | 275,093,933 | U_kgDOEGWZrQ | User | false |
openclaw/openclaw | 4,340,820,861 | I_kwDOQb6kR88AAAABArurfQ | 73,301 | https://github.com/openclaw/openclaw/issues/73301 | https://api.github.com/repos/openclaw/openclaw/issues/73301 | [Bug]: Bonjour plugin crashes gateway with "CIAO PROBING CANCELLED" on cloud/headless environments | ### Bug type
Crash (process/app exits or hangs)
### Beta release blocker
No
### Summary
The Bonjour plugin (used for mDNS/service discovery) generates unhandled promise rejections that crash the entire gateway process. On affected systems, the gateway enters a restart loop, making it unusable until the plugin is m... | closed | completed | true | 1 | [
"bug",
"bug:crash"
] | [] | 2026-04-28T05:30:01Z | 2026-05-01T04:45:14Z | 2026-04-28T05:35:23Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | fabiantcc1 | 101,777,416 | U_kgDOBhEACA | User | false |
openclaw/openclaw | 4,362,892,882 | I_kwDOQb6kR88AAAABBAx2Ug | 75,432 | https://github.com/openclaw/openclaw/issues/75432 | https://api.github.com/repos/openclaw/openclaw/issues/75432 | [Bug]: openai-codex/gpt-5.4 rejects xhigh thinking despite provider xhigh allowlist | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
`openai-codex/gpt-5.4` rejects `thinking: "xhigh"` at runtime even though the provider source lists `gpt-5.4` in `OPENAI_CODEX_XHIGH_MODEL_IDS`.
### Steps to reproduce
1. Start OpenClaw 2026.4.27 with `openai... | closed | not_planned | true | 1 | [
"bug",
"bug:behavior"
] | [] | 2026-05-01T04:57:49Z | 2026-05-04T04:37:27Z | 2026-05-01T05:02:15Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | bukik07 | 25,246,816 | MDQ6VXNlcjI1MjQ2ODE2 | User | false |
openclaw/openclaw | 4,362,894,145 | I_kwDOQb6kR88AAAABBAx7QQ | 75,433 | https://github.com/openclaw/openclaw/issues/75433 | https://api.github.com/repos/openclaw/openclaw/issues/75433 | BUG: Embedded channel reply runs crash on SecretRef-backed Telegram/Discord credentials | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to OpenClaw `2026.4.29`, inbound Telegram messages reach the gateway and Telegram can still send bot replies, but embedded agent reply runs fail before model output when channel credentials are configured thr... | open | null | false | 1 | [
"bug"
] | [] | 2026-05-01T04:58:06Z | 2026-05-02T23:13:26Z | null | NONE | null | 20260503T000641Z | 2026-05-03T00:06:41Z | kAIborg24 | 261,855,322 | U_kgDOD5uYWg | User | false |
openclaw/openclaw | 4,362,911,805 | I_kwDOQb6kR88AAAABBAzAPQ | 75,436 | https://github.com/openclaw/openclaw/issues/75436 | https://api.github.com/repos/openclaw/openclaw/issues/75436 | Heartbeat exec-event interval-bypass regression (re: #17797): runaway runs every ~10s instead of configured every:30m | ## Summary
#17797 ("Heartbeat flood: exec-event bypasses interval check, causing runaway heartbeat runs") is regressed on `2026.4.30` (or never actually shipped — that issue is closed as completed but has no closing PR). The exact same code path described in #17797 is still live, with the same self-feeding loop and th... | closed | not_planned | true | 1 | [] | [] | 2026-05-01T05:03:22Z | 2026-05-04T04:37:27Z | 2026-05-01T05:07:12Z | CONTRIBUTOR | null | 20260504T060031Z | 2026-05-04T06:00:31Z | hexsprite | 23,538 | MDQ6VXNlcjIzNTM4 | User | false |
openclaw/openclaw | 4,360,162,005 | I_kwDOQb6kR88AAAABA-LK1Q | 75,197 | https://github.com/openclaw/openclaw/issues/75197 | https://api.github.com/repos/openclaw/openclaw/issues/75197 | [Bug]: voice-call OpenAI realtime transcription times out during Twilio media stream while direct WebSocket succeeds | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
Twilio inbound voice-call media streams connect and initial TTS plays, but OpenAI realtime transcription times out during the live call, so caller speech is never transcribed or routed to the agent.
### Steps ... | closed | completed | true | 2 | [
"bug"
] | [] | 2026-04-30T17:50:09Z | 2026-05-04T04:37:26Z | 2026-05-01T05:25:38Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | donkeykong91 | 15,494,499 | MDQ6VXNlcjE1NDk0NDk5 | User | false |
openclaw/openclaw | 4,362,984,031 | I_kwDOQb6kR88AAAABBA3aXw | 75,437 | https://github.com/openclaw/openclaw/issues/75437 | https://api.github.com/repos/openclaw/openclaw/issues/75437 | Gateway blocks event loop 30s+ per message; bundled runtime deps re-stage every startup, manifest never persists | # Gateway blocks event loop 30s+ per message; bundled runtime deps re-stage every startup, manifest never persists
## Summary
On a clean Linux install (Ubuntu / Linux 6.8, Node 22.22.2), `openclaw gateway run` blocks the Node event loop for 30+ seconds per inbound message. Reproduced across **2026.4.15, 2026.4.24, 20... | closed | completed | true | 3 | [] | [] | 2026-05-01T05:23:52Z | 2026-05-04T04:33:25Z | 2026-05-01T05:28:03Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | alvaro-raise | 270,761,089 | U_kgDOECN8gQ | User | false |
openclaw/openclaw | 4,362,902,718 | I_kwDOQb6kR88AAAABBAycvg | 75,434 | https://github.com/openclaw/openclaw/issues/75434 | https://api.github.com/repos/openclaw/openclaw/issues/75434 | [Bug]: Feishu multi-agent accounts broken - session resolves to default agent only | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
Multi-agent Feishu routing resolves correct agent but session file resolves to default 'main' agent, breaking non-main agent message handling
### Steps to reproduce
1. Configure multiple Feishu bot accounts with per-agent ... | closed | completed | true | 1 | [
"bug",
"regression"
] | [] | 2026-05-01T05:00:15Z | 2026-05-04T04:37:23Z | 2026-05-01T05:28:25Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | huihongze | 267,369,254 | U_kgDOD--7Jg | User | false |
openclaw/openclaw | 4,363,032,611 | I_kwDOQb6kR88AAAABBA6YIw | 75,444 | https://github.com/openclaw/openclaw/issues/75444 | https://api.github.com/repos/openclaw/openclaw/issues/75444 | Fireworks provider: kimi-k2p5-turbo rejects tool schemas containing 'not' keyword (400 JSON Schema not supported) | ## Summary
The Fireworks API rejects any tool schema containing the JSON Schema `not` keyword with:
```
400 JSON Schema not supported: could not understand the instance `{'not': {}}`.
```
This causes every run using `fireworks/accounts/fireworks/routers/kimi-k2p5-turbo` to fail at the first tool call with a format e... | closed | not_planned | true | 1 | [] | [] | 2026-05-01T05:38:39Z | 2026-05-04T04:34:51Z | 2026-05-02T01:04:05Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | bobrenze-bot | 258,452,996 | U_kgDOD2euBA | User | false |
openclaw/openclaw | 4,362,668,380 | I_kwDOQb6kR88AAAABBAkJXA | 75,407 | https://github.com/openclaw/openclaw/issues/75407 | https://api.github.com/repos/openclaw/openclaw/issues/75407 | feat: Telegram channel Fallback IP transport for restricted networks (GFW bypass) | ## Problem
The Telegram channel fails to connect to `api.telegram.org` in networks where DNS resolution or direct TCP connections are blocked (e.g., behind the Great Firewall of China). The gateway logs show repeated failures:
```
fetch timeout reached; aborting operation (url: api.telegram.org, timeoutMs=2500, elaps... | open | null | false | 2 | [] | [] | 2026-05-01T03:41:10Z | 2026-05-03T11:25:55Z | null | NONE | null | 20260503T120038Z | 2026-05-03T12:00:38Z | huangdijia | 8,337,659 | MDQ6VXNlcjgzMzc2NTk= | User | false |
openclaw/openclaw | 4,363,057,534 | I_kwDOQb6kR88AAAABBA75fg | 75,449 | https://github.com/openclaw/openclaw/issues/75449 | https://api.github.com/repos/openclaw/openclaw/issues/75449 | Telegram group topic responses silently dropped — model generates text but does not auto-deliver | ## Bug Summary
In Telegram **group topics** (forum-style group chats with thread IDs), the assistant generates text responses but they are **silently dropped** and never delivered to the chat. The model produces the text expecting auto-delivery (as happens in DMs), but the OpenClaw Telegram provider does not auto-send... | closed | completed | true | 1 | [] | [] | 2026-05-01T05:48:54Z | 2026-05-04T04:37:22Z | 2026-05-01T05:54:18Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | dustinprojectcoordinator-alt | 268,295,773 | U_kgDOD_3eXQ | User | false |
openclaw/openclaw | 4,363,070,391 | I_kwDOQb6kR88AAAABBA8rtw | 75,452 | https://github.com/openclaw/openclaw/issues/75452 | https://api.github.com/repos/openclaw/openclaw/issues/75452 | Heartbeat per-turn model override persists after turn completes (2026.4.29) | ## Summary
The per-turn `agents.defaults.heartbeat.model` override is applied as a session-level persistence rather than a scoped single-turn override. After the heartbeat turn completes, the override is not cleared, so subsequent user turns silently run on the heartbeat model instead of the configured primary.
## Env... | closed | completed | true | 1 | [] | [] | 2026-05-01T05:54:02Z | 2026-05-05T04:27:01Z | 2026-05-02T15:14:32Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | suifatt7799-oss | 241,606,991 | U_kgDODmahTw | User | false |
openclaw/openclaw | 4,362,690,230 | I_kwDOQb6kR88AAAABBAletg | 75,412 | https://github.com/openclaw/openclaw/issues/75412 | https://api.github.com/repos/openclaw/openclaw/issues/75412 | active-memory and other memory-only embedded runs incur 16-18s of fixed setup overhead per invocation; propose `fastPath` opt-in | ## Summary
The `active-memory` plugin (`extensions/active-memory/index.js`) calls `params.api.runtime.agent.runEmbeddedPiAgent({ toolsAllow: ["memory_recall","memory_search","memory_get"], bootstrapContextMode: "lightweight", ... })` before every eligible reply. On a moderately-sized workspace (≤20 active skills, defa... | closed | duplicate | true | 9 | [] | [] | 2026-05-01T03:49:18Z | 2026-05-04T04:35:07Z | 2026-05-01T22:26:26Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | Catteres | 264,463,750 | U_kgDOD8Nlhg | User | false |
openclaw/openclaw | 4,363,092,374 | I_kwDOQb6kR88AAAABBA-Blg | 75,458 | https://github.com/openclaw/openclaw/issues/75458 | https://api.github.com/repos/openclaw/openclaw/issues/75458 | Discord status reactions can leave permanent thinking emoji after completion | ## Summary
Discord status reactions can leave the final/permanent reaction in a non-terminal lifecycle state. In a live Discord channel, after a turn completed successfully, the only remaining reaction on the triggering message was `🤔` (`thinking`) instead of the terminal `👍` (`done`) or no reaction.
This is confus... | closed | completed | true | 2 | [] | [] | 2026-05-01T06:01:30Z | 2026-05-04T04:35:25Z | 2026-05-01T21:36:55Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | davelutztx | 48,866,029 | MDQ6VXNlcjQ4ODY2MDI5 | User | false |
openclaw/openclaw | 4,363,088,385 | I_kwDOQb6kR88AAAABBA9yAQ | 75,457 | https://github.com/openclaw/openclaw/issues/75457 | https://api.github.com/repos/openclaw/openclaw/issues/75457 | Bug: tweakcn theme import fails with "failed to fetch" due to CSP connect-src restriction | ## Problem
The Control UI's tweakcn theme import feature fails with a `failed to fetch` error when pasting a valid tweakcn theme share link.
## Root Cause
The Control UI is served with a restrictive Content-Security-Policy header:
```
connect-src 'self' ws: wss:
```
This blocks the browser from making fetch reques... | closed | not_planned | false | 1 | [] | [] | 2026-05-01T06:00:20Z | 2026-05-03T07:05:56Z | 2026-05-03T07:05:56Z | NONE | null | 20260503T090142Z | 2026-05-03T09:01:42Z | Lunareith | 262,595,192 | U_kgDOD6bieA | User | false |
openclaw/openclaw | 4,363,085,900 | I_kwDOQb6kR88AAAABBA9oTA | 75,456 | https://github.com/openclaw/openclaw/issues/75456 | https://api.github.com/repos/openclaw/openclaw/issues/75456 | security audit --json can hang while staging plugin runtime deps and scanning backup extension dirs | ## Summary
`openclaw security audit --json` can hang / run for a long time and leave an `openclaw-security` process behind. The current evidence points to two related audit-boundary issues:
1. The security audit plugin collector discovery path loads plugin metadata through the plugin runtime loader, which can trigger... | closed | completed | true | 3 | [] | [] | 2026-05-01T05:59:31Z | 2026-05-05T04:27:21Z | 2026-05-02T12:13:09Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | liushuaiiu | 42,984,729 | MDQ6VXNlcjQyOTg0NzI5 | User | false |
openclaw/openclaw | 4,362,743,026 | I_kwDOQb6kR88AAAABBAos8g | 75,418 | https://github.com/openclaw/openclaw/issues/75418 | https://api.github.com/repos/openclaw/openclaw/issues/75418 | [Bug] Embedded agent startup regressed to ~70s after updating to 2026.4.29 (Linux/Node 24) | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After updating from a previous version to 2026.4.29, embedded agent startup time regressed from near-instant to ~70 seconds on every triggered run. Prior to the update, Milo responded almost immediately via Telegram.
### St... | closed | completed | true | 2 | [
"bug",
"regression"
] | [] | 2026-05-01T04:09:36Z | 2026-05-03T04:33:46Z | 2026-05-01T04:14:56Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | sotongking | 107,119,150 | U_kgDOBmKCLg | User | false |
openclaw/openclaw | 4,362,221,586 | I_kwDOQb6kR88AAAABBAI4Eg | 75,329 | https://github.com/openclaw/openclaw/issues/75329 | https://api.github.com/repos/openclaw/openclaw/issues/75329 | Agent prep takes >40s: system-prompt building regression (30-35s spent in resolvePromptBuildSystemPrompt) | ## Version
OpenClaw 2026.4.29 (a448042)
## Environment
- OS: CachyOS (Arch Linux), Kernel 6.19.12-1
- Node: v25.9.0
- Install: npm global
- Hardware: Dell Latitude 5300
## Symptoms
Every agent start takes over 40 seconds. The `system-prompt` stage alone consumes 30-35 seconds (74-78% of total prep time). This was <5 ... | closed | not_planned | true | 5 | [] | [] | 2026-05-01T01:04:43Z | 2026-05-04T04:36:00Z | 2026-05-01T15:51:19Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | erforschtbot-cmyk | 272,903,176 | U_kgDOEEQsCA | User | false |
openclaw/openclaw | 4,363,384,709 | I_kwDOQb6kR88AAAABBBP3hQ | 75,508 | https://github.com/openclaw/openclaw/issues/75508 | https://api.github.com/repos/openclaw/openclaw/issues/75508 | [Bug]: `openclaw security audit` can hang with no output while the installed `runSecurityAudit(...)` path completes | ## Summary
On OpenClaw `2026.4.29` installed via global npm on Linux, both:
```bash
openclaw security audit
openclaw security audit --deep
```
can appear to hang silently with no output for a long time.
This does **not** look like an audit-core failure. In the same install, direct invocation of the installed `runSe... | closed | not_planned | true | 1 | [] | [] | 2026-05-01T07:32:40Z | 2026-05-04T04:36:44Z | 2026-05-01T11:34:01Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | evva122evva-droid | 267,462,071 | U_kgDOD_Eltw | User | false |
openclaw/openclaw | 4,364,083,934 | I_kwDOQb6kR88AAAABBB6i3g | 75,611 | https://github.com/openclaw/openclaw/issues/75611 | https://api.github.com/repos/openclaw/openclaw/issues/75611 | Skill Contribution: agent-external-memory — Persistent Cross-Session Memory for AI Agents | ## Skill Submission
**Name:** agent-external-memory
**Repo:** https://github.com/binbinyu99-crypto/agent-external-memory
**Type:** New skill contribution
## What It Does
Gives OpenClaw agents persistent, structured memory that survives session boundaries using static JSON files on any web server.
### The Problem
O... | closed | not_planned | true | 1 | [] | [] | 2026-05-01T11:08:05Z | 2026-05-04T04:36:43Z | 2026-05-01T11:34:14Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | binbinyu99-crypto | 279,754,879 | U_kgDOEKy4fw | User | false |
openclaw/openclaw | 4,362,608,321 | I_kwDOQb6kR88AAAABBAgewQ | 75,387 | https://github.com/openclaw/openclaw/issues/75387 | https://api.github.com/repos/openclaw/openclaw/issues/75387 | [Bug]: 2026.4.29 regression: high Telegram response latency and embedded-agent event loop stalls | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
After upgrading to OpenClaw 2026.4.29, Telegram message handling became much slower, with embedded-agent traces showing 34-57s before `stream-ready` and gateway liveness warnings showing event loop delays up to ~32s.
### S... | closed | completed | true | 3 | [
"bug"
] | [] | 2026-05-01T03:19:25Z | 2026-05-03T04:33:45Z | 2026-05-01T03:24:07Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | LDLoeb | 216,241,987 | U_kgDODOOXQw | User | false |
openclaw/openclaw | 4,362,994,855 | I_kwDOQb6kR88AAAABBA4Epw | 75,438 | https://github.com/openclaw/openclaw/issues/75438 | https://api.github.com/repos/openclaw/openclaw/issues/75438 | `IMessageRpcClient.request()` does not handle async stdin write errors → uncaughtException → gateway crash |
### Summary
In `dist/probe-*.js` (the `IMessageRpcClient` class, an `imsg rpc` JSON-RPC client over a child process's stdin), `request()` writes to `this.child.stdin.write(line)` without an error callback or stream `error` listener. When the child closes its stdin (rate limit, child crash, network hiccup), Node emits... | closed | completed | true | 1 | [] | [] | 2026-05-01T05:26:59Z | 2026-05-04T04:36:40Z | 2026-05-01T11:45:15Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | SL4N | 272,086,617 | U_kgDOEDe2WQ | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.