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 | 3,911,283,568 | I_kwDOQb6kR87pIXNw | 11,473 | https://github.com/openclaw/openclaw/issues/11473 | https://api.github.com/repos/openclaw/openclaw/issues/11473 | Feature: Static message cron payload (no AI tokens) | ## Problem
Currently, all cron payloads require an AI call to process:
- `systemEvent` → injects into session → Claude processes → output
- `agentTurn` → spawns isolated session → Claude processes → output
This means even simple, predictable messages (e.g., "Good morning!", "Don't forget to drink water!") cost tokens... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-07T21:26:07Z | 2026-05-05T01:18:27Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | daniella-11ways | 259,312,847 | U_kgDOD3TMzw | User | false |
openclaw/openclaw | 3,911,346,386 | I_kwDOQb6kR87pImjS | 11,487 | https://github.com/openclaw/openclaw/issues/11487 | https://api.github.com/repos/openclaw/openclaw/issues/11487 | Add optional `name`/`label` field to group config entries | ## Problem
Telegram (and other channel) group entries in `openclaw.json` are keyed by numeric group IDs, e.g.:
```json
"groups": {
"-1001234567890": { "requireMention": true },
"-1009876543210": { "requireMention": false },
"-100555666777": {}
}
```
These are unreadable for humans and **also for agents** — whe... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-07T21:54:34Z | 2026-05-05T01:18:28Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | geilt | 1,961,259 | MDQ6VXNlcjE5NjEyNTk= | User | false |
openclaw/openclaw | 4,329,829,413 | I_kwDOQb6kR88AAAABAhP0JQ | 71,895 | https://github.com/openclaw/openclaw/issues/71895 | https://api.github.com/repos/openclaw/openclaw/issues/71895 | [Bug]: Device-less operator can self-declare arbitrary scopes including admin via WS connect params | ## Severity Assessment
### CVSS Assessment
| Metric | v3.1 | v4.0 |
|--------|------|------|
| **Score** | 9.6 / 10.0 | 9.3 / 10.0 |
| **Severity** | Critical | Critical |
| **Vector** | CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:H/A:N | CVSS:4.0/AV:N/AC:L/AT:N/PR:L/UI:N/VC:H/VI:H/VA:N/SC:H/SI:H/SA:N |
| **Calculator** |... | closed | completed | true | 1 | [] | [] | 2026-04-26T02:23:23Z | 2026-04-28T04:42:50Z | 2026-04-26T02:29:59Z | CONTRIBUTOR | null | 20260428T060020Z | 2026-04-28T06:00:20Z | coygeek | 65,363,919 | MDQ6VXNlcjY1MzYzOTE5 | User | false |
openclaw/openclaw | 3,911,351,095 | I_kwDOQb6kR87pIns3 | 11,489 | https://github.com/openclaw/openclaw/issues/11489 | https://api.github.com/repos/openclaw/openclaw/issues/11489 | [Feature Request] Group pairing flow: notify owner + read-only mode for unconfigured groups | ## Feature Request
### Problem
When a bot is added to a new group (Telegram, WhatsApp, etc.), there is no pairing flow like there is for DMs. The bot either:
- Ignores the group entirely (if not in allowlist)
- Processes all messages (if `groups: "*"`)
Neither option gives the owner visibility or control over new gro... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-07T21:56:29Z | 2026-05-05T03:09:45Z | null | CONTRIBUTOR | null | 20260505T060034Z | 2026-05-05T06:00:34Z | napetrov | 18,015,221 | MDQ6VXNlcjE4MDE1MjIx | User | false |
openclaw/openclaw | 3,911,367,781 | I_kwDOQb6kR87pIrxl | 11,492 | https://github.com/openclaw/openclaw/issues/11492 | https://api.github.com/repos/openclaw/openclaw/issues/11492 | [Feature]: Allow agents to resolve Slack channel IDs to names via channel-info/channel-list | ### Summary
The message tool supports `channel-info` and `channel-list` for Discord but not Slack. Agents have no way to resolve Slack channel IDs to human-readable names.
### Problem
When agents receive Slack events, they only see channel IDs (e.g. `C0123ABC`). There is no built-in way to resolve these to channel n... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-07T22:04:04Z | 2026-05-05T03:09:47Z | null | CONTRIBUTOR | null | 20260505T060034Z | 2026-05-05T06:00:34Z | Lukavyi | 1,013,690 | MDQ6VXNlcjEwMTM2OTA= | User | false |
openclaw/openclaw | 3,911,496,378 | I_kwDOQb6kR87pJLK6 | 11,517 | https://github.com/openclaw/openclaw/issues/11517 | https://api.github.com/repos/openclaw/openclaw/issues/11517 | Architecture: reduce import coupling in pi-embedded-runner/attempt.ts (50+ imports) and discord monitor (30 imports) | ## Summary
Several critical files have very deep import chains that create tight coupling across subsystems. This makes refactoring risky and increases the blast radius of changes.
## Worst Offenders
### `src/agents/pi-embedded-runner/run/attempt.ts` (915 lines, 50+ imports)
This file imports from nearly every subsy... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-07T22:58:31Z | 2026-05-05T01:18:29Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | Jdharris06 | 150,635,654 | U_kgDOCPqEhg | User | false |
openclaw/openclaw | 3,911,800,961 | I_kwDOQb6kR87pKViB | 11,623 | https://github.com/openclaw/openclaw/issues/11623 | https://api.github.com/repos/openclaw/openclaw/issues/11623 | Feature: Floating agent bubbles (Clawi) for macOS | ## Proposal
**Floating agent indicator bubbles** — each running agent gets a visible orb on screen.
### UX
- Each agent = one floating bubble (lobster red orb with smooth animation)
- Bubble shows agent name (configurable: "🦞 Johnny", "🤖 Chitti", etc.)
- Double-click toggles mouse-follow mode (with toast notificati... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-08T02:45:19Z | 2026-05-05T01:18:31Z | null | CONTRIBUTOR | null | 20260505T030055Z | 2026-05-05T03:00:55Z | soumikbhatta | 29,822,748 | MDQ6VXNlcjI5ODIyNzQ4 | User | false |
openclaw/openclaw | 3,911,903,854 | I_kwDOQb6kR87pKupu | 11,655 | https://github.com/openclaw/openclaw/issues/11655 | https://api.github.com/repos/openclaw/openclaw/issues/11655 | [Feature]: Slack App Home tab support (views.publish) | ## Summary
Add support for the Slack App Home tab by listening for `app_home_opened` events and calling `views.publish` to render custom Block Kit content.
## Motivation
The Slack plugin currently handles messages, reactions, pins, and member info, but the App Home tab is unused. The manifest already enables `app_ho... | closed | completed | true | 5 | [
"enhancement"
] | [] | 2026-02-08T03:45:36Z | 2026-05-04T04:36:21Z | 2026-05-01T13:20:47Z | CONTRIBUTOR | null | 20260504T060031Z | 2026-05-04T06:00:31Z | TinyTb | 5,957,298 | MDQ6VXNlcjU5NTcyOTg= | User | false |
openclaw/openclaw | 3,911,931,519 | I_kwDOQb6kR87pK1Z_ | 11,665 | https://github.com/openclaw/openclaw/issues/11665 | https://api.github.com/repos/openclaw/openclaw/issues/11665 | [Feature]: Webhook hook sessions should reuse existing session when sessionKey is consistent (multi-turn support) | ## Problem
The docs for `/hooks/agent` state:
> `sessionKey` optional (string): Using a consistent key allows for a **multi-turn conversation** within the hook context.
However, this does not work as documented. `resolveCronSession()` in `src/cron/isolated-agent/run.ts` **always generates a new `crypto.randomUUID()`... | open | null | false | 6 | [
"enhancement"
] | [] | 2026-02-08T04:05:02Z | 2026-05-05T01:18:32Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | marieldejesus12 | 38,159,708 | MDQ6VXNlcjM4MTU5NzA4 | User | false |
openclaw/openclaw | 3,911,974,120 | I_kwDOQb6kR87pK_zo | 11,676 | https://github.com/openclaw/openclaw/issues/11676 | https://api.github.com/repos/openclaw/openclaw/issues/11676 | [Feature]: Support running OpenClaw on devices without Linux / Node.js (inspired by MimiClaw) | I’d like to propose adding support for running **OpenClaw on devices that do not require Linux or Node.js** — such as embedded systems, microcontrollers, and ultra-low-resource hardware.
refer to: https://github.com/memovai/mimiclaw
Today, OpenClaw relies on the Node.js runtime and typical OS environments (Linux/macOS... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-08T04:33:36Z | 2026-05-05T01:18:33Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | crispyberry | 73,692,049 | MDQ6VXNlcjczNjkyMDQ5 | User | false |
openclaw/openclaw | 3,912,081,541 | I_kwDOQb6kR87pLaCF | 11,703 | https://github.com/openclaw/openclaw/issues/11703 | https://api.github.com/repos/openclaw/openclaw/issues/11703 | [Feature]: Batch WhatsApp messages by messageTimestamp instead of receive time | ## Summary
When multiple WhatsApp messages are sent simultaneously (e.g., forward + comment, or rapid-fire messages), they arrive as separate agent turns instead of being batched together, even with `debounceMs` configured. This is because baileys delivers messages 5-10+ seconds apart despite having the same WhatsApp ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-08T06:04:33Z | 2026-05-05T01:18:35Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | VsevolodUstinov | 10,168,615 | MDQ6VXNlcjEwMTY4NjE1 | User | false |
openclaw/openclaw | 3,912,247,079 | I_kwDOQb6kR87pMCcn | 11,747 | https://github.com/openclaw/openclaw/issues/11747 | https://api.github.com/repos/openclaw/openclaw/issues/11747 | [Feature]: Auto-Response for Discord Reaction Events | **Current Behavior**
When reactionNotifications is enabled in Discord config ("own", "all", or "allowlist"), reaction events are:
✅ Received by the gateway
✅ Logged as system events
✅ Visible in agent context
❌ NOT triggering an agent response
The agent only processes reactions when already active in a conversation. I... | closed | not_planned | true | 5 | [
"enhancement"
] | [] | 2026-02-08T07:53:31Z | 2026-05-01T04:40:42Z | 2026-04-29T03:20:14Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | anthonyharley32 | 73,502,817 | MDQ6VXNlcjczNTAyODE3 | User | false |
openclaw/openclaw | 3,912,784,218 | I_kwDOQb6kR87pOFla | 11,894 | https://github.com/openclaw/openclaw/issues/11894 | https://api.github.com/repos/openclaw/openclaw/issues/11894 | Feature: Per-call cost tracking & auth profile attribution | ## Feature Request
### Problem
When using multiple auth profiles (e.g. `anthropic:manual` as primary and `anthropic:payg` as fallback), there's currently no way to see:
1. **Which auth profile** was actually used for a given API call
2. **Per-call cost** (input/output tokens × model pricing)
### Use Case
I have a fr... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-08T13:57:11Z | 2026-05-05T01:18:36Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | aolinger | 9,626,304 | MDQ6VXNlcjk2MjYzMDQ= | User | false |
openclaw/openclaw | 3,913,021,516 | I_kwDOQb6kR87pO_hM | 11,955 | https://github.com/openclaw/openclaw/issues/11955 | https://api.github.com/repos/openclaw/openclaw/issues/11955 | [Feature]: Memory/Context Improvements (metrics + global semantic search + conversation chaining + preload on restart) | ## Summary
This issue groups several related improvements around **agent memory, cross-session context, and observability**. They came up while evolving a local-first personal assistant setup.
## Proposed solution
### 1) Agent self-evaluation / metrics API
Provide session-level or rolling metrics that an agent (and us... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-08T16:13:37Z | 2026-05-05T02:00:19Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | bzhz2023 | 213,139,146 | U_kgDODLQ-yg | User | false |
openclaw/openclaw | 3,913,028,299 | I_kwDOQb6kR87pPBLL | 11,960 | https://github.com/openclaw/openclaw/issues/11960 | https://api.github.com/repos/openclaw/openclaw/issues/11960 | Feature request: session:reset hook event for memory persistence on auto-resets | ## Problem
The `session-memory` bundled hook only fires on `command:new` events. When sessions auto-reset via:
- Daily reset (`session.reset.mode: "daily"`, `atHour: 4`)
- Idle timeout (`resetByType.dm.idleMinutes: 480`, `resetByType.group.idleMinutes: 120`)
...the conversation context is silently discarded without t... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-08T16:17:47Z | 2026-05-05T01:18:37Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | acce1erated | 127,651,273 | U_kgDOB5vNyQ | User | false |
openclaw/openclaw | 3,913,106,835 | I_kwDOQb6kR87pPUWT | 11,977 | https://github.com/openclaw/openclaw/issues/11977 | https://api.github.com/repos/openclaw/openclaw/issues/11977 | feat: add linux/riscv64 Docker image to multi-platform build | ## Component
Docker release pipeline (`docker-release.yml`)
## Problem statement
OpenClaw currently ships Docker images for `linux/amd64` and `linux/arm64`. Users running RISC-V 64-bit hardware — developer boards (BananaPi F3, StarFive VisionFive 2, Milk-V Pioneer), server platforms (SiFive HiFive), and cloud instan... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-08T17:04:25Z | 2026-05-05T02:00:20Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | gounthar | 116,569 | MDQ6VXNlcjExNjU2OQ== | User | false |
openclaw/openclaw | 3,913,239,958 | I_kwDOQb6kR87pP02W | 12,008 | https://github.com/openclaw/openclaw/issues/12008 | https://api.github.com/repos/openclaw/openclaw/issues/12008 | [Feature] Allow configuring Gemini safety settings in OpenClaw | ## Summary:
When using Gemini-based providers/models, OpenClaw doesn’t expose Gemini “Safety Settings” (harassment, hate, sexually explicit, dangerous content). I’d like to adjust these thresholds per provider/model to reduce unnecessary blocked/empty responses.
## Proposed solution:
- Add optional config fields for G... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-08T18:13:17Z | 2026-05-05T01:18:39Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | ALPERDURUKAN | 1,380,193 | MDQ6VXNlcjEzODAxOTM= | User | false |
openclaw/openclaw | 3,913,380,186 | I_kwDOQb6kR87pQXFa | 12,047 | https://github.com/openclaw/openclaw/issues/12047 | https://api.github.com/repos/openclaw/openclaw/issues/12047 | Feature Request: Extract text from document attachments in Telegram channel | ## Summary
When users send document attachments (e.g., .docx, .pdf, .txt files) via Telegram, OpenClaw downloads the file but only shows a `<media:document>` placeholder to the agent. The document contents are not extracted or made available.
## Current Behavior
- Documents ARE downloaded and saved to the media store
... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-08T19:26:56Z | 2026-05-05T01:23:52Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | barry-p5cc | 258,459,530 | U_kgDOD2fHig | User | false |
openclaw/openclaw | 3,913,762,832 | I_kwDOQb6kR87pR0gQ | 12,163 | https://github.com/openclaw/openclaw/issues/12163 | https://api.github.com/repos/openclaw/openclaw/issues/12163 | Feature Request: Add speed parameter support for OpenAI TTS | ## Summary
The OpenAI TTS API supports a `speed` parameter (0.25 to 4.0), but OpenClaw doesn't pass it through. This means users can't control the speech rate when using OpenAI as the TTS provider.
## Current Behavior
The `[[tts:speed=1.5]]` directive works for ElevenLabs but is ignored for OpenAI TTS.
## Expected ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-08T23:03:01Z | 2026-05-05T02:55:25Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | useramuser | 31,619,963 | MDQ6VXNlcjMxNjE5OTYz | User | false |
openclaw/openclaw | 3,913,894,243 | I_kwDOQb6kR87pSUlj | 12,198 | https://github.com/openclaw/openclaw/issues/12198 | https://api.github.com/repos/openclaw/openclaw/issues/12198 | [Feature]: Markdown validation for prompt files | ## Summary
Markdown files used for system prompts should follow rules to ensure they don't contain any patterns that could lead to prompt injections. While not everyone might agree on what is an appropriate level of severity, there are patterns that could only used for bad purposes.
## Proposed solution
When readin... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T00:26:41Z | 2026-05-05T01:18:41Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | horribleCodes | 7,748,313 | MDQ6VXNlcjc3NDgzMTM= | User | false |
openclaw/openclaw | 3,913,949,898 | I_kwDOQb6kR87pSiLK | 12,208 | https://github.com/openclaw/openclaw/issues/12208 | https://api.github.com/repos/openclaw/openclaw/issues/12208 | All .large bubbles should have scrollable body regardless of truncation | ## Feature
All bubbles with size class `.large` should wrap their body content in a scroll view — not just the ones that hit the truncation threshold.
Currently, inner scroll is only enabled when `shouldTruncate == true` (content exceeds `effectiveTruncationHeight`). But .large bubbles that are tall but don't quite e... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T00:57:08Z | 2026-05-05T01:18:42Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | leftspin | 377,231 | MDQ6VXNlcjM3NzIzMQ== | User | false |
openclaw/openclaw | 3,914,004,965 | I_kwDOQb6kR87pSvnl | 12,219 | https://github.com/openclaw/openclaw/issues/12219 | https://api.github.com/repos/openclaw/openclaw/issues/12219 | [Feature]: Skill Permission Manifest Standard (skill.yaml) | ## Problem
Skills currently run with full trust - there's no standard way for a skill to declare what permissions it needs, and no way for users to review permissions before installation. Recent incidents (credential stealers, undisclosed network access) highlight the need for informed consent.
## Proposal
A `skill.... | open | null | false | 5 | [
"enhancement"
] | [] | 2026-02-09T01:22:10Z | 2026-05-05T01:18:43Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | ellistev | 9,433,198 | MDQ6VXNlcjk0MzMxOTg= | User | false |
openclaw/openclaw | 3,914,741,284 | I_kwDOQb6kR87pVjYk | 12,394 | https://github.com/openclaw/openclaw/issues/12394 | https://api.github.com/repos/openclaw/openclaw/issues/12394 | [Feature]: Per-user heartbeat routing for channel-level heartbeat messages | # Feature Request: Per-user heartbeat routing for channel-level heartbeat messages
**Labels:** `enhancement`, `feature-request`, `channels`
---
## Problem / Use Case
In multi-user family/team setups with shared WhatsApp channel access, the current `channels.whatsapp.heartbeat.showOk` broadcasts `HEARTBEAT_OK` messa... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T06:38:53Z | 2026-05-05T01:18:45Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | khannedy | 70,091 | MDQ6VXNlcjcwMDkx | User | false |
openclaw/openclaw | 3,914,746,402 | I_kwDOQb6kR87pVkoi | 12,398 | https://github.com/openclaw/openclaw/issues/12398 | https://api.github.com/repos/openclaw/openclaw/issues/12398 | Feature: Postcondition checks for cron jobs (deterministic verification of agent outcomes) | ## Problem
Cron jobs with `sessionTarget: "isolated"` and `payload.kind: "agentTurn"` are inherently non-deterministic. The agent may stop prematurely, hallucinate completion, or silently underperform — and the framework reports `status: "ok"` because the LLM session terminated cleanly.
**There is currently no mechan... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T06:40:45Z | 2026-05-05T01:18:46Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | Sojemand | 214,938,160 | U_kgDODM-yMA | User | false |
openclaw/openclaw | 3,914,923,969 | I_kwDOQb6kR87pWP_B | 12,429 | https://github.com/openclaw/openclaw/issues/12429 | https://api.github.com/repos/openclaw/openclaw/issues/12429 | [Feature]: Add a secure way to pair dynamically autoscaled worked nodes in K8s | # Summary
When running OpenClaw in Kubernetes, gateway and worker nodes need to run as separate pods for independent scaling and fault isolation. Currently, the only way to pair workers to a gateway is through interactive device pairing (PIN-based), which doesn't work in headless container environments and blocks auto... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T07:37:01Z | 2026-05-05T02:55:26Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | NTurakulov | 7,259,932 | MDQ6VXNlcjcyNTk5MzI= | User | false |
openclaw/openclaw | 3,914,989,539 | I_kwDOQb6kR87pWf_j | 12,441 | https://github.com/openclaw/openclaw/issues/12441 | https://api.github.com/repos/openclaw/openclaw/issues/12441 | [Feature] Control UI should accept gateway token from Authorization header | ## Problem
When OpenClaw Control UI is placed behind a reverse proxy, injecting `Authorization: Bearer <token>` at the proxy is not enough for UI auth.
The UI currently expects token bootstrap via URL/query/hash (`?token=` / `#token=`) or manual input in UI settings, then sends token in WS connect payload.
This makes... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T07:56:02Z | 2026-05-05T01:18:48Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | geilt | 1,961,259 | MDQ6VXNlcjE5NjEyNTk= | User | false |
openclaw/openclaw | 3,915,515,001 | I_kwDOQb6kR87pYgR5 | 12,505 | https://github.com/openclaw/openclaw/issues/12505 | https://api.github.com/repos/openclaw/openclaw/issues/12505 | [Feature Request] Unified Built-in Sandbox with Multi-Platform Support and Tiered Presets | **Type**: Feature Request
**Category**: feature-request
**Severity**: Critical
## Problem Statement
OpenClaw's core architecture includes partial security mechanisms (network binding, Token/password auth, Scope authorization, tool policy allowlists, path sandbox), but these are fragmented, inconsistently applied, and... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T09:59:48Z | 2026-05-05T01:18:49Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | fr33d3m0n | 127,989,854 | U_kgDOB6D4Xg | User | false |
openclaw/openclaw | 3,915,515,493 | I_kwDOQb6kR87pYgZl | 12,507 | https://github.com/openclaw/openclaw/issues/12507 | https://api.github.com/repos/openclaw/openclaw/issues/12507 | [Feature Request] Code Verification and Signature for Skills | **Type**: Feature Request
**Category**: feature-request
**Severity**: Critical
## Problem Statement
SKILL.md files (the primary skill definition format) have no integrity verification mechanism. Content can be modified after distribution without detection. There is no code signing, no hash verification, and no conten... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T09:59:55Z | 2026-05-05T03:09:48Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | fr33d3m0n | 127,989,854 | U_kgDOB6D4Xg | User | false |
openclaw/openclaw | 3,915,515,703 | I_kwDOQb6kR87pYgc3 | 12,508 | https://github.com/openclaw/openclaw/issues/12508 | https://api.github.com/repos/openclaw/openclaw/issues/12508 | [Feature Request] Hook Chain Integrity Protection Against Message Tampering | **Type**: Feature Request
**Category**: feature-request
**Severity**: High
## Problem Statement
The hook system (9 hook types: before_agent_start, agent_end, PreToolUse, PostToolUse, etc.) allows sequential message modification by any registered plugin. Hooks execute in registration order with no integrity protection... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-09T09:59:58Z | 2026-05-05T01:18:50Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | fr33d3m0n | 127,989,854 | U_kgDOB6D4Xg | User | false |
openclaw/openclaw | 3,915,516,647 | I_kwDOQb6kR87pYgrn | 12,512 | https://github.com/openclaw/openclaw/issues/12512 | https://api.github.com/repos/openclaw/openclaw/issues/12512 | [Feature Request] SKILL.md Instruction Isolation from Agent Context | **Type**: Feature Request
**Category**: feature-request
**Severity**: High
## Problem Statement
SKILL.md content is injected directly into the AI agent's system prompt/context without any instruction isolation or permission boundary. Malicious skill instructions can override the agent's role, bypass safety constraint... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T10:00:11Z | 2026-05-05T02:00:22Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | fr33d3m0n | 127,989,854 | U_kgDOB6D4Xg | User | false |
openclaw/openclaw | 3,915,526,074 | I_kwDOQb6kR87pYi-6 | 12,554 | https://github.com/openclaw/openclaw/issues/12554 | https://api.github.com/repos/openclaw/openclaw/issues/12554 | [Feature]: sessions_kill should terminate all child exec processes spawned by a session |
## Is your feature request related to a problem?
Yes. When a sub-agent session spawns multiple `exec` commands (e.g., shell processes via `system.run`), there is no way to terminate those running processes when the session is stopped or aborted.
Currently, `sessions_send("STOP")` only queues a text message for the a... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T10:02:23Z | 2026-05-05T01:18:51Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | jmart987 | 140,838,358 | U_kgDOCGUF1g | User | false |
openclaw/openclaw | 3,915,660,370 | I_kwDOQb6kR87pZDxS | 12,578 | https://github.com/openclaw/openclaw/issues/12578 | https://api.github.com/repos/openclaw/openclaw/issues/12578 | Feature Request: Add usageDefault config option for per-session token display | ## Summary
Currently, `/usage off|tokens|full` sets a per-session preference for displaying token usage after each response. However, there is no way to set this as a default for all new sessions via configuration.
## Current Behavior
- `/usage tokens` must be run manually in each new session to enable per-message t... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T10:30:04Z | 2026-05-05T02:55:27Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | the-vec | 259,683,303 | U_kgDOD3pz5w | User | false |
openclaw/openclaw | 3,915,786,030 | I_kwDOQb6kR87pZicu | 12,601 | https://github.com/openclaw/openclaw/issues/12601 | https://api.github.com/repos/openclaw/openclaw/issues/12601 | [Feature Request] Slack Canvas content read support | ## Summary
Add support for reading Slack Canvas content via the message tool or a dedicated canvases tool.
## Current Behavior
- The Slack message tool can receive messages containing Canvas links
- File metadata (ID, title, type) can be retrieved
- **Canvas content (markdown) cannot be read** even with `files:read` a... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-09T10:59:48Z | 2026-05-05T01:18:54Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | mokuda-sx | 218,289,083 | U_kgDODQLTuw | User | false |
openclaw/openclaw | 3,915,796,688 | I_kwDOQb6kR87pZlDQ | 12,602 | https://github.com/openclaw/openclaw/issues/12602 | https://api.github.com/repos/openclaw/openclaw/issues/12602 | [Feature]: Slack Block Kit support for agent messages | Feature Description
Allow agents to compose and send Slack Block Kit messages for richer, more interactive responses.
Use Case
When an AI agent responds in Slack, it currently sends plain text (with markdown). For many use cases — CRM summaries, daily briefings, database query results, action confirmations — Block K... | open | null | false | 13 | [
"enhancement"
] | [
"vincentkoc"
] | 2026-02-09T11:02:08Z | 2026-05-05T01:18:55Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | JonasBoury | 15,573,175 | MDQ6VXNlcjE1NTczMTc1 | User | false |
openclaw/openclaw | 3,916,426,930 | I_kwDOQb6kR87pb-6y | 12,678 | https://github.com/openclaw/openclaw/issues/12678 | https://api.github.com/repos/openclaw/openclaw/issues/12678 | [Feature]: Capability-based permissions for skills/tools (default-deny high-risk actions) | ## Summary
OpenClaw skills and tool integrations are intentionally powerful, but today there is no explicit, enforced permission model at the gateway/tool layer that defines what a given skill is allowed to do.
This means that if a skill is malicious, compromised, or influenced by untrusted input, it may be able to p... | open | null | false | 5 | [
"enhancement"
] | [] | 2026-02-09T13:35:34Z | 2026-05-05T02:55:29Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | ramtinz | 46,963,585 | MDQ6VXNlcjQ2OTYzNTg1 | User | false |
openclaw/openclaw | 3,916,996,629 | I_kwDOQb6kR87peKAV | 12,736 | https://github.com/openclaw/openclaw/issues/12736 | https://api.github.com/repos/openclaw/openclaw/issues/12736 | Feature Request: tools.cron.tools.deny config option for cron job tool restrictions | ## Summary
Cron jobs (isolated sessions with `cron:<jobId>` session keys) are not covered by the existing `tools.subagents.tools.deny` config option, which only applies to sessions matching the `subagent:` prefix.
## Current Behavior
- `tools.subagents.tools.deny: ["message", "tts"]` blocks tools for sub-agent session... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T15:37:31Z | 2026-05-05T02:12:52Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | Koul91 | 107,915,825 | U_kgDOBm6qMQ | User | false |
openclaw/openclaw | 3,917,154,902 | I_kwDOQb6kR87pewpW | 12,756 | https://github.com/openclaw/openclaw/issues/12756 | https://api.github.com/repos/openclaw/openclaw/issues/12756 | Feature: support message read action for Telegram | ## Summary
The `message` tool's `read` action currently supports Discord and Slack channels but not Telegram. This limits the ability to build proactive monitoring and notification systems for Telegram groups.
## Proposed solution
Extend the `message` tool to support Telegram `read` action with parameters similar t... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-09T16:13:29Z | 2026-05-05T01:18:57Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | jibzus | 28,382,334 | MDQ6VXNlcjI4MzgyMzM0 | User | false |
openclaw/openclaw | 3,917,789,057 | I_kwDOQb6kR87phLeB | 12,831 | https://github.com/openclaw/openclaw/issues/12831 | https://api.github.com/repos/openclaw/openclaw/issues/12831 | Feature: Activity panel / tool call observer mode | ## Summary
Se... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T19:27:18Z | 2026-05-05T01:18:58Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | michaellatham | 44,771,437 | MDQ6VXNlcjQ0NzcxNDM3 | User | false |
openclaw/openclaw | 3,917,919,925 | I_kwDOQb6kR87phra1 | 12,850 | https://github.com/openclaw/openclaw/issues/12850 | https://api.github.com/repos/openclaw/openclaw/issues/12850 | Feature: optionally preserve newlines in system-event previews (avoid flattening multi-line Markdown) | ### Summary
Today the Mattermost channel (and possibly other channels) emits an inbound activity `systemEvent` preview computed as:
`preview = bodyText.replace(/\s+/g, " ").slice(0, 160)`
This collapses *all* whitespace including `\n`, so multi-line Markdown appears as a single line in the system-event preview.
Impo... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-09T20:10:11Z | 2026-05-04T04:37:18Z | 2026-05-01T07:37:24Z | CONTRIBUTOR | null | 20260504T060031Z | 2026-05-04T06:00:31Z | echo931 | 259,437,483 | U_kgDOD3azqw | User | false |
openclaw/openclaw | 3,917,968,378 | I_kwDOQb6kR87ph3P6 | 12,855 | https://github.com/openclaw/openclaw/issues/12855 | https://api.github.com/repos/openclaw/openclaw/issues/12855 | [Feature]: Built-in auto-update with configurable schedule, confirmation, and post-update notification | ## Problem
Currently, OpenClaw has basic update primitives (`update.checkOnStart`, `update.channel`, and the `gateway.update.run` tool action), but no built-in auto-update workflow. Users who want scheduled updates, confirmation prompts, or post-update notifications have to wire this up manually using cron jobs, BOOT.... | open | null | false | 5 | [
"enhancement"
] | [] | 2026-02-09T20:21:10Z | 2026-05-05T01:19:00Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | jarvislanou | 258,907,730 | U_kgDOD26eUg | User | false |
openclaw/openclaw | 3,918,598,890 | I_kwDOQb6kR87pkRLq | 12,931 | https://github.com/openclaw/openclaw/issues/12931 | https://api.github.com/repos/openclaw/openclaw/issues/12931 | Feature: configurable /new greeting model (or option to suppress greeting) | ## Problem
When `/new` or `/reset` is sent alone (no trailing message), OpenClaw runs a model turn to generate a greeting. This uses the session's default model, which is often an expensive one (e.g. Opus 4.6 with high thinking).
There's no way to:
1. Use a cheaper model for the greeting turn
2. Suppress the greeting... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-09T23:12:55Z | 2026-05-05T01:19:02Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | sherly-techtree | 258,044,850 | U_kgDOD2Fzsg | User | false |
openclaw/openclaw | 3,919,571,266 | I_kwDOQb6kR87pn-lC | 13,175 | https://github.com/openclaw/openclaw/issues/13175 | https://api.github.com/repos/openclaw/openclaw/issues/13175 | Feature: Feishu interactive card buttons (inline buttons support) | ## Summary
Add support for interactive card buttons in the Feishu channel, similar to the existing Telegram inline buttons feature.
lobster-biscuit
## Problem
Currently, OpenClaw supports `inlineButtons` for Telegram but not for Feishu. The config key `channels.feishu.capabilities.inlineButtons` can be set but has ... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T05:48:03Z | 2026-05-01T04:40:41Z | 2026-04-29T03:20:21Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | julius2016 | 148,059,117 | U_kgDOCNMz7Q | User | false |
openclaw/openclaw | 3,919,765,316 | I_kwDOQb6kR87pot9E | 13,219 | https://github.com/openclaw/openclaw/issues/13219 | https://api.github.com/repos/openclaw/openclaw/issues/13219 | Feature Request: Per-model usage logging for cost tracking | ## Summary
Request native per-model usage logging to enable cost tracking and model-mix optimization.
## Current Behavior
- Session JSONL files contain usage data (input/output tokens, cache, cost) per assistant message
- No aggregated view or log file dedicated to usage tracking
- Users must parse all session files m... | open | null | false | 5 | [
"enhancement"
] | [] | 2026-02-10T06:49:37Z | 2026-05-05T01:19:04Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | duckshrug | 259,036,244 | U_kgDOD3CUVA | User | false |
openclaw/openclaw | 3,919,805,122 | I_kwDOQb6kR87po3rC | 13,225 | https://github.com/openclaw/openclaw/issues/13225 | https://api.github.com/repos/openclaw/openclaw/issues/13225 | Feature Request: Pre-tool-call routing hook for model delegation | ## Summary
Request a hook that fires BEFORE tool execution, with the ability to route the call to a different model/session based on task classification.
## Use Case
Cost optimization via automatic delegation. Simple tasks (file reads, greps, status checks) should route to cheaper models (Haiku) while complex reasonin... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-10T07:03:09Z | 2026-05-05T03:09:50Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | duckshrug | 259,036,244 | U_kgDOD3CUVA | User | false |
openclaw/openclaw | 3,919,845,007 | I_kwDOQb6kR87ppBaP | 13,239 | https://github.com/openclaw/openclaw/issues/13239 | https://api.github.com/repos/openclaw/openclaw/issues/13239 | Proposal: Publish a provider compatibility matrix with deprecation policy | ## Summary
OpenClaw supports many providers and rapidly changing upstream APIs, but users currently discover breakages only after runtime failures. We should define a **public compatibility matrix + deprecation policy**.
## Problem
- Provider API changes break behavior unexpectedly.
- Users can't easily know which pro... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-10T07:15:46Z | 2026-05-05T01:19:05Z | null | CONTRIBUTOR | null | 20260505T030055Z | 2026-05-05T03:00:55Z | 0xRaini | 190,923,101 | U_kgDOC2FBXQ | User | false |
openclaw/openclaw | 3,919,845,699 | I_kwDOQb6kR87ppBlD | 13,241 | https://github.com/openclaw/openclaw/issues/13241 | https://api.github.com/repos/openclaw/openclaw/issues/13241 | Proposal: Add lightweight issue triage and priority framework | ## Summary
With the current issue volume, important regressions and strategic proposals can get buried quickly. I propose a lightweight **issue triage and prioritization framework**.
## Problem
- High-volume reports make it hard to separate urgent reliability work from long-tail requests.
- Users lack visibility into ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T07:16:00Z | 2026-05-05T01:19:06Z | null | CONTRIBUTOR | null | 20260505T030055Z | 2026-05-05T03:00:55Z | 0xRaini | 190,923,101 | U_kgDOC2FBXQ | User | false |
openclaw/openclaw | 3,920,128,810 | I_kwDOQb6kR87pqGsq | 13,271 | https://github.com/openclaw/openclaw/issues/13271 | https://api.github.com/repos/openclaw/openclaw/issues/13271 | Feature request: Per-channel tool call budget with auto-delegation | ## Use Case
Running an agent across multiple channels (Discord, Telegram, etc.) with different resource constraints per channel. Discord sessions need to stay lightweight (≤2 tool calls per turn) to avoid context bloat and WebSocket stability issues, while Telegram sessions can handle full tool access.
Currently ther... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T08:29:04Z | 2026-05-05T01:19:08Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | denys-prog | 238,574,357 | U_kgDODjhbFQ | User | false |
openclaw/openclaw | 3,920,498,441 | I_kwDOQb6kR87prg8J | 13,304 | https://github.com/openclaw/openclaw/issues/13304 | https://api.github.com/repos/openclaw/openclaw/issues/13304 | [Feature]: Content-based agent routing in bindings | ## Summary
Allow `bindings` to match on **message content patterns** (regex or prefix), not just channel/peer/account. This enables routing messages to different agents based on what the user says, without a middleman agent acting as dispatcher.
## Use Case
In a multi-agent setup, a user wants messages in a group ch... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T09:51:21Z | 2026-05-05T01:19:09Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | dan-dekel | 22,286,397 | MDQ6VXNlcjIyMjg2Mzk3 | User | false |
openclaw/openclaw | 3,921,046,591 | I_kwDOQb6kR87ptmw_ | 13,362 | https://github.com/openclaw/openclaw/issues/13362 | https://api.github.com/repos/openclaw/openclaw/issues/13362 | [Feature]: Add openclaw doctor --dry-run for safer config validation | ## Summary
Over the past few issues (for example #9627, #11268, #4654), there seems to be a recurring pattern where openclaw doctor --fix or other config-writing commands modify user configuration in ways that are difficult to recover from. In some cases, environment variables are resolved into plaintext, or redacted ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T11:57:05Z | 2026-05-05T03:09:51Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | Riya-chandra | 136,116,667 | U_kgDOCBz5uw | User | false |
openclaw/openclaw | 3,921,068,509 | I_kwDOQb6kR87ptsHd | 13,364 | https://github.com/openclaw/openclaw/issues/13364 | https://api.github.com/repos/openclaw/openclaw/issues/13364 | [Feature]: Expose before_tool_call/after_tool_call in internal hooks system | ## Summary
Expose the existing `before_tool_call` and `after_tool_call` plugin hooks through the internal hooks discovery system (`HOOK.md` + `handler.ts`), enabling managed hooks to intercept and block tool operations before execution.
## Background
OpenClaw has two hook systems:
1. **Plugin Hooks** (`src/plugins/... | open | null | false | 6 | [
"enhancement"
] | [] | 2026-02-10T12:02:08Z | 2026-05-05T01:19:12Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | similar-bespoke | 125,871,494 | U_kgDOB4Clhg | User | false |
openclaw/openclaw | 3,921,318,344 | I_kwDOQb6kR87pupHI | 13,387 | https://github.com/openclaw/openclaw/issues/13387 | https://api.github.com/repos/openclaw/openclaw/issues/13387 | feat(session): Auto-prune consecutive NO_REPLY messages from transcripts | ## Problem
Accumulated `NO_REPLY` assistant messages in session transcripts cause the model to learn that "NO_REPLY is always correct" — a cascade of silence effect. This is particularly problematic after:
- Extended periods of heartbeat-only activity
- Tool failures or rate limits
- Discussing NO_REPLY itself (#5191)... | closed | not_planned | false | 3 | [
"enhancement"
] | [] | 2026-02-10T12:55:49Z | 2026-05-03T21:24:04Z | 2026-05-03T21:24:04Z | NONE | null | 20260504T000036Z | 2026-05-04T00:00:36Z | Masha-L | 36,271,341 | MDQ6VXNlcjM2MjcxMzQx | User | false |
openclaw/openclaw | 3,921,599,881 | I_kwDOQb6kR87pvt2J | 13,417 | https://github.com/openclaw/openclaw/issues/13417 | https://api.github.com/repos/openclaw/openclaw/issues/13417 | WhatsApp Newsletter/Channel support in message tool | ## Feature Request
### Problem
The `message` tool currently only accepts E.164 phone numbers and group JIDs as targets for WhatsApp. WhatsApp Channels (Newsletters) use a different JID format (`xxxxx@newsletter`) which is rejected by the target validation with:
```
Error: Delivering to WhatsApp requires target <E.164... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-10T13:54:41Z | 2026-05-05T04:26:25Z | 2026-05-02T20:09:33Z | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | agentz-manfred | 258,416,989 | U_kgDOD2chXQ | User | false |
openclaw/openclaw | 3,921,792,102 | I_kwDOQb6kR87pwcxm | 13,440 | https://github.com/openclaw/openclaw/issues/13440 | https://api.github.com/repos/openclaw/openclaw/issues/13440 | Feature request: Pluggable session store backends (Convex, SQLite, Redis, etc.) | ## Summary
The session store is currently hardcoded to file-based JSON (`sessions.json`) with a file lock mechanism (`.lock` via `open('wx')`). This works well for single-instance deployments but has limitations:
1. **Lock contention** under concurrent access (related: #13436)
2. **No multi-instance support** — can't... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-10T14:30:13Z | 2026-05-05T01:19:14Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | realsamrat | 59,194,708 | MDQ6VXNlcjU5MTk0NzA4 | User | false |
openclaw/openclaw | 3,922,120,460 | I_kwDOQb6kR87pxs8M | 13,479 | https://github.com/openclaw/openclaw/issues/13479 | https://api.github.com/repos/openclaw/openclaw/issues/13479 | Feature Request: Show cron job details (prompt/summary) in Web UI | ## Feature Request
### Problem
Currently there's no quick way to view what a specific cron job does without digging into the config. When viewing the cron dashboard in the web UI, clicking on a cron job should reveal its details.
### Proposed Solution
When clicking on a cron job in the dashboard, show:
- **Prompt/Tas... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T15:28:41Z | 2026-05-05T01:19:16Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | vokaplok | 41,402,825 | MDQ6VXNlcjQxNDAyODI1 | User | false |
openclaw/openclaw | 3,922,150,836 | I_kwDOQb6kR87px0W0 | 13,487 | https://github.com/openclaw/openclaw/issues/13487 | https://api.github.com/repos/openclaw/openclaw/issues/13487 | Discord routing: mention > reply-target > default owner (suppress default owner when targeted) | Goal
- Deterministic single-target routing in Discord channels for multi-bot OpenClaw setups.
Desired behavior (strict precedence)
1) If inbound message contains explicit Discord mention token(s) for one or more of our bot accounts (e.g. <@...>), ONLY those mentioned bot(s) are eligible responders.
2) Else if inbound ... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-10T15:35:09Z | 2026-05-05T01:19:17Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | sebastion-warbinder | 258,520,316 | U_kgDOD2i0_A | User | false |
openclaw/openclaw | 3,922,774,708 | I_kwDOQb6kR87p0Mq0 | 13,543 | https://github.com/openclaw/openclaw/issues/13543 | https://api.github.com/repos/openclaw/openclaw/issues/13543 | Feature: Tool-level sandbox mode for selective isolation | ## Summary
Add the ability to configure **per-tool sandbox execution** independent of session sandbox mode. This would allow specific tools (especially those handling external/untrusted data) to always run in a Docker sandbox while keeping other tools on the host for performance and compatibility.
## Use Case
**Prob... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T17:44:50Z | 2026-05-05T01:19:18Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | mpeay-caylent | 95,248,081 | U_kgDOBa1e0Q | User | false |
openclaw/openclaw | 3,922,949,330 | I_kwDOQb6kR87p03TS | 13,570 | https://github.com/openclaw/openclaw/issues/13570 | https://api.github.com/repos/openclaw/openclaw/issues/13570 | context-pruning: forcePruneRatio to bypass TTL at high usage | In long sessions, cache TTL keeps tool results even when the context window is nearly full. Add optional contextPruning.forcePruneRatio so pruning can bypass the TTL when estimated context usage meets the threshold. Update schema and add a test for TTL bypass. | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T18:27:44Z | 2026-05-05T01:19:20Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | 1kuna | 47,035,140 | MDQ6VXNlcjQ3MDM1MTQw | User | false |
openclaw/openclaw | 3,923,048,763 | I_kwDOQb6kR87p1Pk7 | 13,593 | https://github.com/openclaw/openclaw/issues/13593 | https://api.github.com/repos/openclaw/openclaw/issues/13593 | Add logging and warnings for cron job execution failures | ## Problem
The cron scheduler currently suffers from silent failures where `nextRunAtMs` advances without actual job execution. When a scheduled job fails to run, there's no indication in logs or status output, making it difficult to detect and diagnose issues.
**Related:** This is a follow-up to the root cause fix i... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-10T18:53:42Z | 2026-05-05T01:19:22Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,053,106 | I_kwDOQb6kR87p1Qoy | 13,597 | https://github.com/openclaw/openclaw/issues/13597 | https://api.github.com/repos/openclaw/openclaw/issues/13597 | Add comprehensive AWS deployment guide (EC2, ECS, Lambda) | ## Overview
Request documentation for deploying OpenClaw on AWS infrastructure. Current documentation lacks cloud deployment specifics, which would help with adoption and reduce support burden.
## Requested Coverage
### EC2 Deployment
- EC2 instance setup and configuration
- systemd service configuration
- Security ... | open | null | false | 6 | [
"enhancement"
] | [] | 2026-02-10T18:54:35Z | 2026-05-05T01:19:23Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,058,145 | I_kwDOQb6kR87p1R3h | 13,601 | https://github.com/openclaw/openclaw/issues/13601 | https://api.github.com/repos/openclaw/openclaw/issues/13601 | Add persistent cron job execution history with filtering | ## Problem
Cron job execution history is currently stored in-memory only, which means:
- History is lost on gateway restart
- No way to audit job executions over time
- Can't filter by failure/success status
- No historical analysis of job reliability
- Debugging issues requires catching them in real-time
## Why Thi... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T18:55:51Z | 2026-05-05T01:19:24Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,066,469 | I_kwDOQb6kR87p1T5l | 13,607 | https://github.com/openclaw/openclaw/issues/13607 | https://api.github.com/repos/openclaw/openclaw/issues/13607 | Feature: global default for reasoning visiblity | ## Problem
Currently, `reasoningLevel` is stored per-session. New sessions default to `off`, requiring users to manually configure each channel/session.
## Proposed Solution
Add global defaults in agent configuration:
```yaml
agents:
defaults:
reasoningLevel: stream # off | on | stream
```
New sessions would i... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-10T18:58:03Z | 2026-05-05T01:19:26Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | jtatum | 340,001 | MDQ6VXNlcjM0MDAwMQ== | User | false |
openclaw/openclaw | 3,923,067,099 | I_kwDOQb6kR87p1UDb | 13,610 | https://github.com/openclaw/openclaw/issues/13610 | https://api.github.com/repos/openclaw/openclaw/issues/13610 | Add native secrets management integration (AWS Secrets Manager, Vault, etc.) | ## Problem
Currently, API keys and tokens are stored directly in `openclaw.json` config files. This has several issues:
- **Security risk:** Secrets in plaintext config files
- **Version control:** Can't safely commit configs to git
- **Rotation difficulty:** Changing secrets requires config edits and restarts
- **Co... | open | null | false | 7 | [
"enhancement"
] | [
"vincentkoc"
] | 2026-02-10T18:58:14Z | 2026-05-05T01:19:27Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,072,718 | I_kwDOQb6kR87p1VbO | 13,613 | https://github.com/openclaw/openclaw/issues/13613 | https://api.github.com/repos/openclaw/openclaw/issues/13613 | Add multi-gateway cron job coordination for HA deployments | ## Problem
When running multiple OpenClaw gateways (high availability, load balancing, blue/green deployments), there's no coordination mechanism for cron jobs. This causes:
- **Duplicate execution:** Multiple gateways fire the same job simultaneously
- **Race conditions:** Jobs that modify state (file writes, API ca... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T18:59:47Z | 2026-05-05T01:19:29Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,072,949 | I_kwDOQb6kR87p1Ve1 | 13,615 | https://github.com/openclaw/openclaw/issues/13615 | https://api.github.com/repos/openclaw/openclaw/issues/13615 | Add rate limiting and throttling for external API calls | ## Problem
OpenClaw makes many external API calls (LLM providers, web search, Slack, Discord, etc.) but lacks:
- **Rate limit protection:** Hits provider limits unexpectedly
- **Cost control:** Runaway API usage from bugs or misconfigurations
- **Backpressure handling:** No automatic retry when rate limited
- **Visib... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T18:59:51Z | 2026-05-05T01:19:30Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,072,996 | I_kwDOQb6kR87p1Vfk | 13,616 | https://github.com/openclaw/openclaw/issues/13616 | https://api.github.com/repos/openclaw/openclaw/issues/13616 | Add backup/restore utility for config, cron jobs, and session history | ## Problem
Currently, there's no standardized way to:
- Backup OpenClaw configuration and state
- Migrate between environments (dev → staging → prod)
- Recover from catastrophic failures
- Export/import cron jobs
- Archive session history
This makes disaster recovery difficult and environment setup manual.
## Propos... | open | null | false | 8 | [
"enhancement"
] | [] | 2026-02-10T18:59:51Z | 2026-05-05T01:19:31Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,072,997 | I_kwDOQb6kR87p1Vfl | 13,617 | https://github.com/openclaw/openclaw/issues/13617 | https://api.github.com/repos/openclaw/openclaw/issues/13617 | Add testing framework for cron jobs and skills (dry-run mode) | ## Problem
Currently, testing cron jobs and skills requires:
- Running them in production (risky)
- Manually triggering with `--force` (but still executes for real)
- No way to validate behavior without side effects
This makes it difficult to:
- Test new jobs before deploying
- Validate changes to existing jobs
- Run... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T18:59:52Z | 2026-05-05T01:19:33Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,101,512 | I_kwDOQb6kR87p1cdI | 13,620 | https://github.com/openclaw/openclaw/issues/13620 | https://api.github.com/repos/openclaw/openclaw/issues/13620 | Feature: Support Telegram media groups (albums) in message tool | ## Feature Request
**Problem:**
Currently the `message` tool can only send media files one at a time. Telegram supports `sendMediaGroup` API which allows sending 2-10 photos/videos as an album with a shared caption.
**Use case:**
When sending multiple related videos/images (e.g., a curated movie list with trailers), ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T19:07:33Z | 2026-05-05T02:12:54Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | dmnksss | 138,099,646 | U_kgDOCDs7vg | User | false |
openclaw/openclaw | 3,923,174,505 | I_kwDOQb6kR87p1uRp | 13,627 | https://github.com/openclaw/openclaw/issues/13627 | https://api.github.com/repos/openclaw/openclaw/issues/13627 | Feature: outbound message throttle for Slack (timestamp ordering) | ## Problem
Slack sorts messages by timestamp with ~1 second resolution. When an agent sends multiple messages in rapid succession (within the same second), Slack can display them in arbitrary order. This is a Slack-specific issue — other platforms have finer-grained timestamps or sort by message ID.
## Proposed Solut... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T19:28:20Z | 2026-05-05T01:19:34Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | doomclaw | 259,108,889 | U_kgDOD3GwGQ | User | false |
openclaw/openclaw | 3,923,226,485 | I_kwDOQb6kR87p1691 | 13,634 | https://github.com/openclaw/openclaw/issues/13634 | https://api.github.com/repos/openclaw/openclaw/issues/13634 | Feature: Slack workspace file injection + global systemPrompt (parity with Telegram) | ## Problem
Slack sessions start with **zero context** — no AGENTS.md, SOUL.md, MEMORY.md, USER.md injected. This is different from Telegram which gets full workspace file injection automatically.
**Current workaround:** Use `systemPrompt` with `file:path/to/file.md` per-channel. But this requires:
1. Manually creatin... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T19:41:41Z | 2026-05-05T02:12:55Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | visshehi | 162,566,191 | U_kgDOCbCQLw | User | false |
openclaw/openclaw | 3,923,487,735 | I_kwDOQb6kR87p26v3 | 13,670 | https://github.com/openclaw/openclaw/issues/13670 | https://api.github.com/repos/openclaw/openclaw/issues/13670 | Add Model and Token Usage Status Display to Control UI Topbar | ## Feature Request: Model and Token Usage Display in Control UI
**Summary**: Add a real-time status display showing current model and token usage in the OpenClaw Control UI header/topbar.
**Motivation**:
- Users need quick visibility into which model is currently active
- Token usage awareness helps with cost manage... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-10T20:46:39Z | 2026-05-05T01:19:35Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | RobbyMo81 | 3,741,507 | MDQ6VXNlcjM3NDE1MDc= | User | false |
openclaw/openclaw | 3,923,568,220 | I_kwDOQb6kR87p3OZc | 13,676 | https://github.com/openclaw/openclaw/issues/13676 | https://api.github.com/repos/openclaw/openclaw/issues/13676 | Feature Request: First-class Projects in Gateway Dashboard (workspace + skill scoping + runtime constraints) | ## Summary
Add a first-class **Project** concept to the Gateway/Control UI dashboard so users can manage real work as projects (not just ad-hoc sessions), with clear isolation boundaries and fast context switching.
Today, users approximate this with folders + multiple sessions + manual conventions (SDD/TASKS/RUNBOOK) ... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-10T21:05:34Z | 2026-05-05T01:19:37Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | kundeng | 89,032 | MDQ6VXNlcjg5MDMy | User | false |
openclaw/openclaw | 3,923,851,621 | I_kwDOQb6kR87p4Tll | 13,700 | https://github.com/openclaw/openclaw/issues/13700 | https://api.github.com/repos/openclaw/openclaw/issues/13700 | [Feature Request]: Session snapshots — save and load context checkpoints (/session save|load) | ## Problem
During long development sessions, users sometimes want to:
- Test alternative approaches without losing current progress
- Roll back to a known-good state after mistakes
- A/B test different prompts or models in parallel
- Branch conversations without manual transcripts
Currently, OpenClaw saves sessions a... | open | null | false | 6 | [
"enhancement"
] | [] | 2026-02-10T22:30:35Z | 2026-05-05T01:19:38Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | gbacskai | 16,809,418 | MDQ6VXNlcjE2ODA5NDE4 | User | false |
openclaw/openclaw | 3,924,023,695 | I_kwDOQb6kR87p49mP | 13,736 | https://github.com/openclaw/openclaw/issues/13736 | https://api.github.com/repos/openclaw/openclaw/issues/13736 | [Feature]: Support per-agent compaction settings (mode, reserveTokensFloor, memoryFlush) in agents.list[] | ## Summary
Currently, compaction settings in OpenClaw (such as compaction mode, reserveTokensFloor, and memoryFlush configuration) are only configurable in `agents.defaults.compaction`. There is no way to override compaction settings for individual agents listed in `agents.list[]`, unlike other features such as heartbe... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-02-10T23:35:24Z | 2026-05-01T04:40:40Z | 2026-04-29T03:20:29Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | TeemuSailynoja | 33,978,952 | MDQ6VXNlcjMzOTc4OTUy | User | false |
openclaw/openclaw | 3,924,072,513 | I_kwDOQb6kR87p5JhB | 13,744 | https://github.com/openclaw/openclaw/issues/13744 | https://api.github.com/repos/openclaw/openclaw/issues/13744 | Make session write lock configurable + narrow lock scope (avoid timeout=All models failed) | ### Problem
OpenClaw uses a file lock `${sessionFile}.lock` for session writes. In current build (npm `openclaw@2026.2.9`), the lock defaults are:
- `timeoutMs = 10000` (10s)
- `staleMs = 30min`
However, call sites invoke it without overrides:
- `acquireSessionWriteLock({ sessionFile })`
so `timeoutMs/staleMs` are eff... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-10T23:56:35Z | 2026-05-05T01:19:39Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | ruslanavapro | 213,179,132 | U_kgDODLTa_A | User | false |
openclaw/openclaw | 3,924,125,932 | I_kwDOQb6kR87p5Wjs | 13,751 | https://github.com/openclaw/openclaw/issues/13751 | https://api.github.com/repos/openclaw/openclaw/issues/13751 | Feishu plugin: remove dependency on contact:contact.base:readonly for sender name resolution | ## Problem
The Feishu plugin currently requires `contact:contact.base:readonly` permission to resolve sender display names. This is a highly sensitive permission that grants read access to the entire organization's contact directory — far too broad for the simple need of knowing "who sent this message".
## Where it's... | open | null | false | 6 | [
"enhancement"
] | [
"vincentkoc"
] | 2026-02-11T00:16:45Z | 2026-05-05T02:12:57Z | null | CONTRIBUTOR | null | 20260505T030055Z | 2026-05-05T03:00:55Z | echoVic | 16,428,813 | MDQ6VXNlcjE2NDI4ODEz | User | false |
openclaw/openclaw | 3,924,658,321 | I_kwDOQb6kR87p7YiR | 13,870 | https://github.com/openclaw/openclaw/issues/13870 | https://api.github.com/repos/openclaw/openclaw/issues/13870 | Feature Request: Human-friendly device names in paired devices list | ## Problem
The paired devices list in the web UI (and CLI via `openclaw devices list`) displays devices using long cryptographic hash IDs (e.g., `c5f3ba123c4b008a9dd3b29cd21b2c21ae2e3f...`). With multiple devices paired (phones, laptops, desktops, web UI sessions), it's nearly impossible to tell which device is which ... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-11T04:21:17Z | 2026-05-05T02:00:29Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | koome23 | 2,000,526 | MDQ6VXNlcjIwMDA1MjY= | User | false |
openclaw/openclaw | 3,924,966,161 | I_kwDOQb6kR87p8jsR | 13,911 | https://github.com/openclaw/openclaw/issues/13911 | https://api.github.com/repos/openclaw/openclaw/issues/13911 | Feature: per-channel announce suppression for sub-agents | ## Problem
When the main agent spawns a sub-agent from a Telegram session, the completion announcement always routes back to Telegram. There's no config option to suppress or reroute sub-agent announces per channel.
Current workaround: the main agent responds with `NO_REPLY` behaviorally, but this depends on the agen... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-11T06:29:11Z | 2026-05-05T02:12:58Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | nimrodlehavi | 333,174 | MDQ6VXNlcjMzMzE3NA== | User | false |
openclaw/openclaw | 3,925,237,543 | I_kwDOQb6kR87p9l8n | 13,948 | https://github.com/openclaw/openclaw/issues/13948 | https://api.github.com/repos/openclaw/openclaw/issues/13948 | Feature Request: Action-Level Tool Deny | Feature Request: Action-Level Tool Deny
Problem
tools.deny: ["message"] blocks the entire message tool — sends, reads, reacts, search, member-info — all or nothing. There's no way to deny message with action=send while keeping action=read available.
This makes it impossible to build a safety architecture where an ag... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-11T07:59:28Z | 2026-05-05T03:09:53Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | iranitehrani | 217,807,139 | U_kgDODPt5Iw | User | false |
openclaw/openclaw | 3,925,279,542 | I_kwDOQb6kR87p9wM2 | 13,962 | https://github.com/openclaw/openclaw/issues/13962 | https://api.github.com/repos/openclaw/openclaw/issues/13962 | Feature: Per-mention model routing + context window for group mentions | ## Feature Request
### 1. Per-mention model routing in groups
When `requireMention: true`, allow configuring different models for mentioned vs non-mentioned messages.
**Use case:** In group chats, non-mentioned messages could be processed by a local/cheap model (e.g. Ollama) for context awareness, while @-mentioned... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-11T08:12:26Z | 2026-05-05T02:13:00Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | dantaik | 99,078,276 | U_kgDOBefQhA | User | false |
openclaw/openclaw | 3,925,337,281 | I_kwDOQb6kR87p9-TB | 13,968 | https://github.com/openclaw/openclaw/issues/13968 | https://api.github.com/repos/openclaw/openclaw/issues/13968 | On-demand HEARTBEAT.md loading | ### Description
HEARTBEAT.md is currently loaded into context on every message (~380 tokens), but it's only used during heartbeat polls (~every 30 minutes).
### Current behavior
`loadWorkspaceBootstrapFiles()` unconditionally loads all context files on every message:
- AGENTS.md
- SOUL.md
- TOOLS.md
- USER.md
- **HE... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-11T08:28:41Z | 2026-05-05T02:00:33Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | buildz-ops | 185,640,833 | U_kgDOCxCngQ | User | false |
openclaw/openclaw | 3,925,988,082 | I_kwDOQb6kR87qAdLy | 14,051 | https://github.com/openclaw/openclaw/issues/14051 | https://api.github.com/repos/openclaw/openclaw/issues/14051 | Feature: Activity-based heartbeat with idle timeout | ## Summary
Add an idle timeout option for heartbeats so they automatically pause after a period of inactivity.
## Use Case
Currently heartbeats fire on a fixed interval (e.g. every 30m), burning tokens even when the user is asleep or away. Users want heartbeats to:
1. **Start** when they begin a conversation
2. **Stop... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-11T11:15:01Z | 2026-05-05T03:09:54Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | Keiracom | 212,117,713 | U_kgDODKSo0Q | User | false |
openclaw/openclaw | 3,926,420,685 | I_kwDOQb6kR87qCGzN | 14,079 | https://github.com/openclaw/openclaw/issues/14079 | https://api.github.com/repos/openclaw/openclaw/issues/14079 | Linux: No desktop notifications support | ## Problem
macOS gets native desktop notifications via `osascript`. Linux has no notification support at all — no `notify-send`, no libnotify, no D-Bus notification integration.
## Impact
Linux desktop users never receive system notifications from OpenClaw. On macOS, users get notified of important events; on Linux,... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-11T12:56:44Z | 2026-05-05T03:09:55Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | holistic-medical-supply-base | 249,014,699 | U_kgDODtepqw | User | false |
openclaw/openclaw | 3,927,050,948 | I_kwDOQb6kR87qEgrE | 14,133 | https://github.com/openclaw/openclaw/issues/14133 | https://api.github.com/repos/openclaw/openclaw/issues/14133 | Feature: Reminders tab in Mission Control | ## Feature Request
### Summary
Reminders (one-shot `at` cron jobs created by the reminders skill) currently show up in the Cron tab alongside recurring system jobs. They should have their own dedicated **Reminders tab** in Mission Control.
### Desired Behavior
1. **New "Reminders" tab** in Mission Control nav
2. **C... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-11T15:04:20Z | 2026-05-05T02:55:31Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | patrickaha | 209,368,222 | U_kgDODHq0ng | User | false |
openclaw/openclaw | 3,927,313,320 | I_kwDOQb6kR87qFguo | 14,151 | https://github.com/openclaw/openclaw/issues/14151 | https://api.github.com/repos/openclaw/openclaw/issues/14151 | [Feature]: Option to force-prepend an opening <think> tag for local reasoning models | ## Summary
Some local “reasoning” models (notably **GLM-4.7-Flash** and **MiniMax-M2.1**) do **not** reliably emit `<think>...</think>` tags at all, even when they are producing reasoning-style outputs. In practice, it’s **expected that the frontend/orchestrator** (OpenClaw) will handle any required reasoning delimite... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-11T15:54:00Z | 2026-05-05T02:00:35Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | LisaMacintosh | 51,481,552 | MDQ6VXNlcjUxNDgxNTUy | User | false |
openclaw/openclaw | 3,927,798,669 | I_kwDOQb6kR87qHXON | 14,206 | https://github.com/openclaw/openclaw/issues/14206 | https://api.github.com/repos/openclaw/openclaw/issues/14206 | Add Message Rate Limiting for Feishu Channel | ## Summary
Feishu server will temporarily ban api requests if messages are sent with high frequency. It seems that openclaw separates responses from llm into a sequence of shorter messages, and send them one by one, sometimes triggering Feishu server limit.
## Proposed solution
Add Message Rate Limiting for Feis... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-11T17:41:08Z | 2026-05-05T03:14:57Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | immars | 1,144,653 | MDQ6VXNlcjExNDQ2NTM= | User | false |
openclaw/openclaw | 3,927,804,534 | I_kwDOQb6kR87qHYp2 | 14,207 | https://github.com/openclaw/openclaw/issues/14207 | https://api.github.com/repos/openclaw/openclaw/issues/14207 | Feature Request: Configurable Memory Indexing Concurrency | # Feature Request: Configurable Memory Indexing Concurrency
## Problem
During `openclaw memory index`, the indexer processes files with limited concurrency (~4 workers observed), resulting in low GPU utilization when using remote embedding providers with high-capacity GPUs.
**Observed behavior:**
- 4 concurrent requ... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-11T17:42:26Z | 2026-05-01T04:40:40Z | 2026-04-29T03:20:35Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | gsalingu | 1,230,368 | MDQ6VXNlcjEyMzAzNjg= | User | false |
openclaw/openclaw | 3,928,501,869 | I_kwDOQb6kR87qKC5t | 14,248 | https://github.com/openclaw/openclaw/issues/14248 | https://api.github.com/repos/openclaw/openclaw/issues/14248 | Expose clientPingTimeout for Slack Socket Mode configuration | ## Problem
The Slack Socket Mode integration uses `@slack/socket-mode` under the hood, which defaults `clientPingTimeout` to **5000ms**. In production, Slack's WebSocket servers frequently take longer than 5 seconds to respond to pings, causing a flood of pong timeout warnings (~70/hour) that eventually cascade into f... | closed | completed | true | 7 | [
"enhancement"
] | [] | 2026-02-11T20:14:44Z | 2026-05-03T04:37:10Z | 2026-04-30T14:04:30Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | shivasymbl | 216,878,638 | U_kgDODO1OLg | User | false |
openclaw/openclaw | 3,928,526,772 | I_kwDOQb6kR87qKI-0 | 14,251 | https://github.com/openclaw/openclaw/issues/14251 | https://api.github.com/repos/openclaw/openclaw/issues/14251 | [Feature]: Improve gateway stop/uninstall to handle system-level LaunchDaemons | ## Summary
When gateway services are installed as system-level LaunchDaemons (`/Library/LaunchDaemons/`), the `openclaw gateway stop` and `openclaw gateway uninstall` commands cannot detect or manage them. These commands only check user-level LaunchAgents
(`~/Library/LaunchAgents/`), leading to confusing situati... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-11T20:20:41Z | 2026-05-05T02:13:01Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | xinbenlv | 640,325 | MDQ6VXNlcjY0MDMyNQ== | User | false |
openclaw/openclaw | 3,929,129,004 | I_kwDOQb6kR87qMcAs | 14,317 | https://github.com/openclaw/openclaw/issues/14317 | https://api.github.com/repos/openclaw/openclaw/issues/14317 | Feature: Add 'private' groupPolicy option for Slack channels | ## Summary
Add a new `groupPolicy` value `"private"` for Slack that allows the bot to respond in **private channels only**, ignoring public channels unless explicitly allowlisted.
## Motivation
Currently, `groupPolicy` for Slack offers three options:
- `"open"` — respond in any channel (default when no `channels.sla... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-11T23:25:54Z | 2026-05-05T02:00:37Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | debedb | 619,943 | MDQ6VXNlcjYxOTk0Mw== | User | false |
openclaw/openclaw | 3,929,254,876 | I_kwDOQb6kR87qM6vc | 14,344 | https://github.com/openclaw/openclaw/issues/14344 | https://api.github.com/repos/openclaw/openclaw/issues/14344 | Feature request: message delete (revoke) action for WhatsApp channel | ## Feature Request
**What:** Add a `delete` action to the `message` tool for the WhatsApp channel, enabling agents to delete (revoke) their own sent messages.
**Why:** Agents sometimes send messages that need to be retracted (wrong group, leaked info, typos in important contexts). Currently `message delete` returns `... | open | null | false | 5 | [
"enhancement"
] | [] | 2026-02-12T00:25:00Z | 2026-05-05T02:13:03Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | ariconnect | 61,789,560 | MDQ6VXNlcjYxNzg5NTYw | User | false |
openclaw/openclaw | 3,929,396,154 | I_kwDOQb6kR87qNdO6 | 14,376 | https://github.com/openclaw/openclaw/issues/14376 | https://api.github.com/repos/openclaw/openclaw/issues/14376 | Feature: Reason-aware cron guardrails (quota/auth/rate-limit aware backoff + circuit breaker) | ## Problem
OpenClaw cron jobs already apply an exponential backoff based on consecutive errors, but the backoff is *reason-agnostic*. In practice, different failure reasons should trigger different safety behavior:
- **billing/quota exhausted (402 / insufficient_quota)**: continuing to retry is wasteful; jobs should ... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-12T01:27:13Z | 2026-05-05T02:13:05Z | null | CONTRIBUTOR | null | 20260505T030055Z | 2026-05-05T03:00:55Z | futuremind2026 | 258,860,756 | U_kgDOD23m1A | User | false |
openclaw/openclaw | 3,929,580,873 | I_kwDOQb6kR87qOKVJ | 14,407 | https://github.com/openclaw/openclaw/issues/14407 | https://api.github.com/repos/openclaw/openclaw/issues/14407 | [Feature]: Integrate password toggle functionality across all platforms to let users see password in clear text | Summary Section:
- ✅ Clear problem statement: "Add password toggle/show functionality across all platforms"
- ✅ Explains the feature: Users can control password visibility with eye icon
- ✅ Platforms listed: Web, Desktop (macOS/Windows), iOS, Android
Proposed Solution Section:
- ✅ Specific implementation d... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-12T02:47:20Z | 2026-05-05T02:13:06Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | gibtang | 1,892,674 | MDQ6VXNlcjE4OTI2NzQ= | User | false |
openclaw/openclaw | 3,930,342,808 | I_kwDOQb6kR87qREWY | 14,526 | https://github.com/openclaw/openclaw/issues/14526 | https://api.github.com/repos/openclaw/openclaw/issues/14526 | Proposal: safer self-update with pre-update backup + health check + auto-rollback/restore | ## Problem
On production deployments (e.g., VPS), updates can occasionally introduce regressions or configuration/compat changes. Operators want a safer, more automated upgrade path that reduces downtime and prevents “broken after update” situations.
## Proposal
Add an *optional* "safe update" mode/workflow that:
1) ... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-12T07:48:10Z | 2026-05-05T02:13:08Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | sovushik | 30,425,777 | MDQ6VXNlcjMwNDI1Nzc3 | User | false |
openclaw/openclaw | 3,931,225,822 | I_kwDOQb6kR87qUb7e | 14,601 | https://github.com/openclaw/openclaw/issues/14601 | https://api.github.com/repos/openclaw/openclaw/issues/14601 | [Feature]: : Add MoonshotAI Kimi K2.5 (via OpenRouter) to Model Selector in onboard | the Model is not in the onboard for OpenRouter | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-12T11:07:08Z | 2026-05-05T02:00:45Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | Tomer2006 | 90,401,167 | MDQ6VXNlcjkwNDAxMTY3 | User | false |
openclaw/openclaw | 3,931,505,565 | I_kwDOQb6kR87qVgOd | 14,619 | https://github.com/openclaw/openclaw/issues/14619 | https://api.github.com/repos/openclaw/openclaw/issues/14619 | [Feature]: Option to skip tool list text in system prompt (reduce duplication) | ## Summary
The system prompt currently includes a "Tooling" section with a human-readable list of tool names and descriptions. However, the full tool schemas (which include descriptions) are also sent via the API for function calling. This creates duplication and unnecessary token overhead.
## Current behavior
1. **... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-12T12:07:03Z | 2026-05-05T02:00:46Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | JasperMcGarry | 259,993,619 | U_kgDOD38wEw | User | false |
openclaw/openclaw | 3,931,582,710 | I_kwDOQb6kR87qVzD2 | 14,629 | https://github.com/openclaw/openclaw/issues/14629 | https://api.github.com/repos/openclaw/openclaw/issues/14629 | Output sanitizer: improve duplicate detection for same-line and partial duplicates | ## Problem
The `sanitizeUserFacingText()` pipeline in `pi-embedded-helpers` includes `collapseConsecutiveDuplicateBlocks()`, which deduplicates identical paragraphs separated by `\n\n`. However, it misses several common model-generated duplicate patterns:
### What's caught ✅
- `<final>` / `</final>` tags (via `stripF... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-12T12:22:26Z | 2026-05-05T02:13:09Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | communityorg-discord | 255,771,849 | U_kgDODz7EyQ | User | false |
openclaw/openclaw | 3,932,532,530 | I_kwDOQb6kR87qZa8y | 14,747 | https://github.com/openclaw/openclaw/issues/14747 | https://api.github.com/repos/openclaw/openclaw/issues/14747 | Feature request: configurable lane wait diagnostic threshold | ## Summary
The lane scheduler emits [diagnostic] lane wait exceeded warnings with a hardcoded threshold of 2 seconds (warnAfterMs ?? 2e3 in extensionAPI.js:748). For setups with cron jobs that legitimately take 60-120s to complete (e.g. Gmail polling, calendar checks via external CLI tools), this fires on every single... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-12T15:41:14Z | 2026-05-05T02:00:49Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | Stache73 | 7,196,384 | MDQ6VXNlcjcxOTYzODQ= | User | false |
openclaw/openclaw | 3,932,969,293 | I_kwDOQb6kR87qbFlN | 14,804 | https://github.com/openclaw/openclaw/issues/14804 | https://api.github.com/repos/openclaw/openclaw/issues/14804 | Enhancement: Add ephemeral file system support for isolated sessions | **Is your feature request related to a problem? Please describe.**
Currently, isolated sessions persist file changes in the workspace unless cleaned up. This can lead to clutter or side effects across runs.
**Describe the solution you'd like**
Add an option to use an ephemeral/tmpfs-backed workspace for sessions, ens... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-12T17:10:25Z | 2026-05-05T02:00:51Z | null | CONTRIBUTOR | null | 20260505T030055Z | 2026-05-05T03:00:55Z | Protocol-zero-0 | 257,158,451 | U_kgDOD1PtMw | User | false |
openclaw/openclaw | 4,069,891,043 | I_kwDOQb6kR87ylZvj | 44,908 | https://github.com/openclaw/openclaw/issues/44908 | https://api.github.com/repos/openclaw/openclaw/issues/44908 | [Bug]: WhatsApp audio-only messages in groups do not trigger preflight transcription or mention check | ### Bug type
Regression / Missing feature (doc says it works, it doesn't)
### Summary
When `requireMention: true` is set for a WhatsApp group, audio-only voice messages are not preflight-transcribed before the mention check. The gateway immediately marks `wasMentioned: false` and archives the message without ever ca... | closed | completed | true | 3 | [] | [] | 2026-03-13T09:12:25Z | 2026-04-30T04:42:08Z | 2026-04-26T04:49:09Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | miguelvidex | 26,462,830 | MDQ6VXNlcjI2NDYyODMw | User | false |
openclaw/openclaw | 3,933,301,414 | I_kwDOQb6kR87qcWqm | 14,850 | https://github.com/openclaw/openclaw/issues/14850 | https://api.github.com/repos/openclaw/openclaw/issues/14850 | [Feature]: Session key override on bindings (cross-channel session unification) | ## Problem
In a multi-agent setup where each agent has both a Telegram bot and a Discord guild channel, there is no way to unify sessions across these surfaces. A message sent in a Discord guild channel creates a `channel:<channelId>` session, while a Telegram DM creates a DM session — even though both are routed to t... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-12T18:19:00Z | 2026-05-05T02:00:52Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | rita-aga | 36,133,358 | MDQ6VXNlcjM2MTMzMzU4 | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.