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 | 3,887,700,676 | I_kwDOQb6kR87nuZrE | 7,433 | https://github.com/openclaw/openclaw/issues/7433 | https://api.github.com/repos/openclaw/openclaw/issues/7433 | Improve WhatsApp/Baileys group message reliability | # Improve WhatsApp/Baileys group message reliability
## Summary
The current WhatsApp/Baileys integration has a solid foundation but is missing several recommended patterns from the Baileys documentation that improve group message reliability, especially for high-volume groups and message retry scenarios.
## Current ... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-02T20:28:07Z | 2026-05-05T00:08:02Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | octopuslabs-fl | 168,146,581 | U_kgDOCgW2lQ | User | false |
openclaw/openclaw | 3,887,814,932 | I_kwDOQb6kR87nu1kU | 7,456 | https://github.com/openclaw/openclaw/issues/7456 | https://api.github.com/repos/openclaw/openclaw/issues/7456 | [Feature]: Add "go back" navigation during onboarding wizard | ## Summary
During the onboarding wizard (`openclaw onboard`), users cannot navigate back to a previous step if they make a mistake or want to change an earlier selection. The only options are to continue forward or cancel and restart the entire process.
This is frustrating when:
- You accidentally skip a step or sele... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-02T21:15:51Z | 2026-05-04T23:09:53Z | null | NONE | null | 20260505T000034Z | 2026-05-05T00:00:34Z | celsomilne | 24,833,767 | MDQ6VXNlcjI0ODMzNzY3 | User | false |
openclaw/openclaw | 3,887,861,154 | I_kwDOQb6kR87nvA2i | 7,476 | https://github.com/openclaw/openclaw/issues/7476 | https://api.github.com/repos/openclaw/openclaw/issues/7476 | Feature Request: WhatsApp sticker send support | ## Description
The WhatsApp plugin currently has no way to send sticker messages. Sending a `.webp` file (512x512, <100KB) via the message tool delivers it as a regular image attachment, not as a sticker.
## Current behavior
- Inbound stickers are received correctly (as `.webp` files with `<media:sticker>` tag)
- `m... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-02T21:36:36Z | 2026-05-05T00:08:04Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | giannisanni | 115,853,836 | U_kgDOBufKDA | User | false |
openclaw/openclaw | 3,887,910,696 | I_kwDOQb6kR87nvM8o | 7,490 | https://github.com/openclaw/openclaw/issues/7490 | https://api.github.com/repos/openclaw/openclaw/issues/7490 | Feature Request: Add 'description' field to agent config for dynamic agent discovery | ## Problem
Currently, the agent configuration schema supports these fields:
- `id`, `name`, `workspace`, `agentDir`, `model`, `identity`, `tools`, `heartbeat`, `subagents`
However, there's **no `description` field** to describe what each agent specializes in.
## Use Case: Dynamic Agent Discovery
When orchestrating ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-02T21:59:05Z | 2026-05-05T03:09:32Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | shannon0430 | 258,282,406 | U_kgDOD2UTpg | User | false |
openclaw/openclaw | 3,888,046,297 | I_kwDOQb6kR87nvuDZ | 7,524 | https://github.com/openclaw/openclaw/issues/7524 | https://api.github.com/repos/openclaw/openclaw/issues/7524 | Feature: groupScope option to consolidate group sessions into main | ## Problem
Currently, all group chats are **always isolated** into their own sessions via the session key format: `agent:${agentId}:${channel}:${peerKind}:${peerId}` (session-key.js line 130). There is no equivalent of `dmScope: "main"` for groups.
This means users cannot have a group chat consolidated into their mai... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-02T22:58:56Z | 2026-05-05T11:23:31Z | null | NONE | null | 20260505T120033Z | 2026-05-05T12:00:33Z | ThatGuySizemore | 27,755,958 | MDQ6VXNlcjI3NzU1OTU4 | User | false |
openclaw/openclaw | 3,888,294,413 | I_kwDOQb6kR87nwqoN | 7,575 | https://github.com/openclaw/openclaw/issues/7575 | https://api.github.com/repos/openclaw/openclaw/issues/7575 | feat: Sysbox Docker Runtime for Secure Container Isolation (Host Maintenance Required) | ## Summary
**Status:** 🟡 **HOLD** - Requires host-level changes, deferred until maintenance window
Implement Sysbox runtime for secure Docker-in-Docker isolation without privileged mode. This upgrade enhances container security by moving from Docker Socket Sharing (DooD) to true isolated sandbox containers.
---
##... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-03T00:32:47Z | 2026-05-05T09:10:56Z | null | NONE | null | 20260505T120033Z | 2026-05-05T12:00:33Z | iavankda | 219,703,262 | U_kgDODRhn3g | User | false |
openclaw/openclaw | 3,888,593,536 | I_kwDOQb6kR87nxzqA | 7,629 | https://github.com/openclaw/openclaw/issues/7629 | https://api.github.com/repos/openclaw/openclaw/issues/7629 | memory-lancedb: Add hybrid search (BM25 + vector) | ## Feature Request
**Plugin:** `memory-lancedb`
### Current Behavior
The memory-lancedb plugin uses pure vector search via `table.vectorSearch()`. This works well for semantic similarity but misses exact keyword matches (IDs, env vars, code symbols, names).
### Proposed Enhancement
Add hybrid search combining:
- **V... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-03T02:12:30Z | 2026-05-05T00:08:08Z | null | CONTRIBUTOR | null | 20260505T030055Z | 2026-05-05T03:00:55Z | joeykrug | 5,925,937 | MDQ6VXNlcjU5MjU5Mzc= | User | false |
openclaw/openclaw | 3,888,705,859 | I_kwDOQb6kR87nyPFD | 7,648 | https://github.com/openclaw/openclaw/issues/7648 | https://api.github.com/repos/openclaw/openclaw/issues/7648 | Feature Request: Option to disable auto-scroll in Control UI / WebChat | ## Problem
When the agent responds to a message in the Control UI or WebChat, the view auto-scrolls to the bottom of the response. For longer responses, users have to scroll back up to read from the beginning.
## Requested Feature
Add a configuration option or UI toggle to:
1. **Disable auto-scroll** — Keep the view... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-03T02:43:24Z | 2026-05-05T02:55:20Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | shawmar12 | 223,550,435 | U_kgDODVMb4w | User | false |
openclaw/openclaw | 3,888,776,668 | I_kwDOQb6kR87nygXc | 7,661 | https://github.com/openclaw/openclaw/issues/7661 | https://api.github.com/repos/openclaw/openclaw/issues/7661 | Feature Request: Add channel creation to message tool | ## Feature Request
**Action needed:** `channel create` or similar for the `message` tool
**Current state:**
- OpenClaw CLI has `openclaw message channel` with only `info` and `list` subcommands
- Users can grant bots `channels:manage` scope in Slack
- No way to programmatically create channels from OpenClaw
**Use ca... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-03T03:04:54Z | 2026-05-05T08:19:42Z | null | CONTRIBUTOR | null | 20260505T090045Z | 2026-05-05T09:00:45Z | bennewton999 | 458,991 | MDQ6VXNlcjQ1ODk5MQ== | User | false |
openclaw/openclaw | 3,888,815,408 | I_kwDOQb6kR87nyp0w | 7,669 | https://github.com/openclaw/openclaw/issues/7669 | https://api.github.com/repos/openclaw/openclaw/issues/7669 | [Feature] Re-enable dev channel distribution tag for npm releases | ## Summary
Re-enable the `dev` channel for npm distribution, as documented at https://docs.openclaw.ai/install/development-channels.
## Background
Development channels allow users to install packages from different release stages:
- **stable**: Tagged releases (vYYYY.M.D or vYYYY.M.D-\<patch\>), npm dist-tag `lates... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-03T03:17:24Z | 2026-05-04T23:10:04Z | null | NONE | null | 20260505T000034Z | 2026-05-05T00:00:34Z | ImBIOS | 41,441,643 | MDQ6VXNlcjQxNDQxNjQz | User | false |
openclaw/openclaw | 3,888,856,419 | I_kwDOQb6kR87nyz1j | 7,679 | https://github.com/openclaw/openclaw/issues/7679 | https://api.github.com/repos/openclaw/openclaw/issues/7679 | Telegram should default to allowlist mode with owner ID | ## Summary
The current default `dmPolicy: pairing` for Telegram allows anyone to message the bot and request access via a pairing code. This creates a security concern where random users could potentially get paired if someone approves them accidentally.
## Suggested Change
Default to `dmPolicy: allowlist` with the ... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-03T03:30:36Z | 2026-05-05T00:08:10Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | jregalbuto | 258,301,628 | U_kgDOD2VevA | User | false |
openclaw/openclaw | 3,889,015,617 | I_kwDOQb6kR87nzatB | 7,707 | https://github.com/openclaw/openclaw/issues/7707 | https://api.github.com/repos/openclaw/openclaw/issues/7707 | Feature Request: Memory Trust Tagging by Source | **Feature:** Tag agent memory entries by trust level based on their origin (user commands, web scrapes, third-party skills).
**Use case:** Prevent memory poisoning attacks where malicious instructions are hidden in untrusted content (web pages, messages, third-party integrations) and later influence agent behavior.
*... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-03T04:13:38Z | 2026-05-05T03:09:33Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | LumenLantern | 258,675,890 | U_kgDOD2sUsg | User | false |
openclaw/openclaw | 3,889,062,154 | I_kwDOQb6kR87nzmEK | 7,717 | https://github.com/openclaw/openclaw/issues/7717 | https://api.github.com/repos/openclaw/openclaw/issues/7717 | Feature Request: Discord role-mention triggers | ## Feature Request
**Description:**
Add support for Discord role mentions (e.g., `@bots`, `@AI`) to trigger bot responses, not just direct @mentions.
**Current Behavior:**
- Direct `@BotName` mentions → bot wakes up and responds ✅
- Role mentions (`@RoleName` where bot has that role) → bot sees the message but doesn'... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-03T04:30:25Z | 2026-05-05T00:08:11Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | jsquared71 | 4,314,939 | MDQ6VXNlcjQzMTQ5Mzk= | User | false |
openclaw/openclaw | 3,889,065,894 | I_kwDOQb6kR87nzm-m | 7,722 | https://github.com/openclaw/openclaw/issues/7722 | https://api.github.com/repos/openclaw/openclaw/issues/7722 | Feature Request: Filesystem Sandboxing Config (tools.fileAccess) | **Feature:** Filesystem access restrictions via configuration
**What happened:**
We attempted to implement filesystem sandboxing with this config:
```javascript
{
"tools": {
"fileAccess": {
"allowedPaths": ["/home/seraph/.openclaw/workspace", "/tmp"],
"denyPaths": ["/etc", "/root", "~/.ssh", "/var/lo... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-03T04:31:56Z | 2026-05-05T07:00:35Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | LumenLantern | 258,675,890 | U_kgDOD2sUsg | User | false |
openclaw/openclaw | 3,889,607,844 | I_kwDOQb6kR87n1rSk | 7,817 | https://github.com/openclaw/openclaw/issues/7817 | https://api.github.com/repos/openclaw/openclaw/issues/7817 | Feature: Project trees + threaded entry points in Control UI | ## Feature Request
Enhance the Control UI webchat with project-based thread organization and ADHD-friendly message triage.
## Current State
- Session dropdown in upper-right shows flat list of sessions
- Single linear conversation per session
- No way to split/reorganize messages between sessions
## Proposed Enhanc... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-03T07:11:22Z | 2026-05-05T11:30:30Z | null | CONTRIBUTOR | null | 20260505T120033Z | 2026-05-05T12:00:33Z | ianderrington | 76,016,868 | MDQ6VXNlcjc2MDE2ODY4 | User | false |
openclaw/openclaw | 3,889,668,419 | I_kwDOQb6kR87n16FD | 7,834 | https://github.com/openclaw/openclaw/issues/7834 | https://api.github.com/repos/openclaw/openclaw/issues/7834 | Add thorough QA tests for observability ingestor | ## Summary
The observability ingestor needs comprehensive QA testing beyond unit tests.
## Current State
- Basic ingestor works (bug fixed in 872de4641)
- Manual testing with real Telegram requires actual bot interaction
- No automated integration tests for the full pipeline
## Needed
- [ ] Integration test that simu... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-03T07:28:25Z | 2026-04-29T04:35:27Z | 2026-04-26T15:54:01Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | ldraney | 107,153,866 | U_kgDOBmMJyg | User | false |
openclaw/openclaw | 4,329,813,098 | I_kwDOQb6kR88AAAABAhO0ag | 71,891 | https://github.com/openclaw/openclaw/issues/71891 | https://api.github.com/repos/openclaw/openclaw/issues/71891 | [Bug]: vLLM/Nemotron response stored as thinking-only; assistantTexts empty despite successful completion | ### Bug type
Regression (worked before, now fails)
### Beta release blocker
No
### Summary
OpenClaw can receive a successful vLLM/Nemotron completion but store the visible answer as a thinking-only content part. The model.completed trace then has assistantTexts=[], so the user sees the generic “Agent couldn’t gene... | closed | completed | true | 3 | [
"bug",
"regression"
] | [] | 2026-04-26T02:12:37Z | 2026-04-28T04:39:04Z | 2026-04-26T03:07:19Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | jmystaki-create | 267,217,950 | U_kgDOD-1sHg | User | false |
openclaw/openclaw | 3,890,257,663 | I_kwDOQb6kR87n4J7_ | 7,909 | https://github.com/openclaw/openclaw/issues/7909 | https://api.github.com/repos/openclaw/openclaw/issues/7909 | [Feature]: Add plain text copy option | Description:
Currently, when clicking the copy button on message content, it only supports copying as Markdown format. Please add an option to copy as plain text (.txt format), similar to how Gemini allows users to copy content in either Markdown or plain text.
Use Case:
Many users need raw text without Markdown form... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-03T09:38:41Z | 2026-05-05T08:56:53Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | rvpoochen | 2,375,305 | MDQ6VXNlcjIzNzUzMDU= | User | false |
openclaw/openclaw | 3,891,582,940 | I_kwDOQb6kR87n9Nfc | 8,061 | https://github.com/openclaw/openclaw/issues/8061 | https://api.github.com/repos/openclaw/openclaw/issues/8061 | [Feature]: per-agent maxConcurrent in agents.list | ## Summary
maxConcurrent for subagents can only be set globally via agents.defaults.subagents.maxConcurrent. There's no way to limit concurrency for a specific subagent while leaving others unrestricted.
Use case: I have a TTS subagent running a local model that should only have 1 instance at a time. My other subagen... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-03T14:34:29Z | 2026-05-05T00:08:14Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | NicolasBataille | 63,408,806 | MDQ6VXNlcjYzNDA4ODA2 | User | false |
openclaw/openclaw | 3,892,087,305 | I_kwDOQb6kR87n_IoJ | 8,122 | https://github.com/openclaw/openclaw/issues/8122 | https://api.github.com/repos/openclaw/openclaw/issues/8122 | Feature Request: Per-group session reset configuration | ## Summary
Support configuring session reset settings at the per-group level, not just globally or by type/channel.
## Use Case
I have dedicated TTS (text-to-speech) groups where each message is an independent translation task with no need for conversation context. I'd like these groups to auto-reset after 1 minute id... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-03T16:25:11Z | 2026-05-05T03:09:35Z | null | CONTRIBUTOR | null | 20260505T060034Z | 2026-05-05T06:00:34Z | tylerliu612 | 179,897,207 | U_kgDOCrkDdw | User | false |
openclaw/openclaw | 4,329,762,708 | I_kwDOQb6kR88AAAABAhLvlA | 71,880 | https://github.com/openclaw/openclaw/issues/71880 | https://api.github.com/repos/openclaw/openclaw/issues/71880 | Claude Opus can produce false-success empty terminal turns (stop=stop, content=[], usage=0) leading to 'Agent couldn't generate a response' | ## Summary
On OpenClaw 2026.4.23, Claude Opus 4.6 / 4.7 can intermittently fail with:
> ⚠️ Agent couldn't generate a response. Please try again.
In the failing cases, the assistant turn is sometimes recorded as a **false-success empty terminal turn** instead of an error:
- `stopReason = "stop"`
- `content = []`
- `... | closed | completed | true | 2 | [] | [] | 2026-04-26T01:41:07Z | 2026-04-28T04:42:52Z | 2026-04-26T02:28:22Z | NONE | null | 20260428T060020Z | 2026-04-28T06:00:20Z | MagnaAI | 265,843,828 | U_kgDOD9h0dA | User | false |
openclaw/openclaw | 3,892,110,763 | I_kwDOQb6kR87n_OWr | 8,130 | https://github.com/openclaw/openclaw/issues/8130 | https://api.github.com/repos/openclaw/openclaw/issues/8130 | Feature Request: Proactive user notification on LLM failover/API errors | ## Summary
When the primary LLM provider (e.g., Anthropic) returns errors (HTTP 500, rate limits, etc.) and Clawdbot fails over to a backup provider, the user currently receives no notification. The system goes silent during retries/failover, leaving the user wondering what happened.
## Requested Features
### 1. Fai... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-03T16:30:33Z | 2026-05-05T03:09:36Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | alvelda | 1,775,986 | MDQ6VXNlcjE3NzU5ODY= | User | false |
openclaw/openclaw | 3,892,247,613 | I_kwDOQb6kR87n_vw9 | 8,147 | https://github.com/openclaw/openclaw/issues/8147 | https://api.github.com/repos/openclaw/openclaw/issues/8147 | [Feature]: Add configurable send hotkey for WebChat (Cmd+Enter instead of Enter) | Description Currently, pressing Enter in the WebChat input box sends the message immediately. This causes issues for users who use IME (Input Method Editor) and are accustomed to pressing Enter to select suggestions, as it accidentally sends incomplete messages.
Proposed Solution Add a configuration option to change t... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-03T17:00:48Z | 2026-05-05T08:05:19Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | moomx | 26,848,368 | MDQ6VXNlcjI2ODQ4MzY4 | User | false |
openclaw/openclaw | 3,892,416,468 | I_kwDOQb6kR87oAY_U | 8,176 | https://github.com/openclaw/openclaw/issues/8176 | https://api.github.com/repos/openclaw/openclaw/issues/8176 | Allow before_agent_start hooks to override prompt | ## Problem
Plugins can currently only prepend context or adjust the system prompt in `before_agent_start`. They cannot change the actual user prompt. This blocks safe mitigation for context-limit errors when the *current* prompt is large, even if history pruning succeeds.
Real-world symptom (Anthropic):
```
LLM reques... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-03T17:42:13Z | 2026-05-05T00:08:16Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | richardfogaca | 63,572,350 | MDQ6VXNlcjYzNTcyMzUw | User | false |
openclaw/openclaw | 3,892,432,271 | I_kwDOQb6kR87oAc2P | 8,181 | https://github.com/openclaw/openclaw/issues/8181 | https://api.github.com/repos/openclaw/openclaw/issues/8181 | Make local embeddings for memory search easier to set up (zero-config) | ## Feature Request
### Problem
Setting up memory search currently requires either:
- A remote API key (OpenAI or Gemini) for embeddings
- A manual local setup with `node-llama-cpp` that requires `pnpm approve-builds` and a rebuild step
For users who want fully private, zero-cost memory search, the local option should... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-03T17:46:02Z | 2026-05-03T04:35:36Z | 2026-04-30T18:44:19Z | NONE | null | 20260503T060030Z | 2026-05-03T06:00:30Z | hashslingers | 129,774,036 | U_kgDOB7wx1A | User | false |
openclaw/openclaw | 3,892,464,057 | I_kwDOQb6kR87oAkm5 | 8,190 | https://github.com/openclaw/openclaw/issues/8190 | https://api.github.com/repos/openclaw/openclaw/issues/8190 | Feature: Global session reset on auth switch | ## Problem
When switching auth backends or API keys (e.g., switching between different Claude API accounts, OpenRouter, or other providers), each session caches its auth at creation time. This means:
1. Config change applies to gateway
2. Gateway restarts with new auth
3. **Existing sessions still use old cached auth... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-03T17:53:55Z | 2026-05-05T00:08:17Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | lancejames221b | 60,149,579 | MDQ6VXNlcjYwMTQ5NTc5 | User | false |
openclaw/openclaw | 3,892,680,101 | I_kwDOQb6kR87oBZWl | 8,209 | https://github.com/openclaw/openclaw/issues/8209 | https://api.github.com/repos/openclaw/openclaw/issues/8209 | Feature Request: Native Cron Job → Mission Control Integration | **Priority:** Medium
**Component:** OpenClaw Gateway - Cron System
---
## Summary
Add native gateway-level hooks to automatically log cron job start/completion events to external task management systems, eliminating reliance on agent compliance for audit trails.
---
## Problem Statement
Currently, cron job exec... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-03T18:48:45Z | 2026-04-29T04:35:52Z | 2026-04-26T14:55:30Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | 100menotu001 | 64,228,916 | MDQ6VXNlcjY0MjI4OTE2 | User | false |
openclaw/openclaw | 3,893,245,193 | I_kwDOQb6kR87oDjUJ | 8,285 | https://github.com/openclaw/openclaw/issues/8285 | https://api.github.com/repos/openclaw/openclaw/issues/8285 | Feature request: Auto-send intent/acknowledgment text message before agent processing | ## Problem
When users send messages, there is currently no way to automatically send a brief text acknowledgment (e.g., "On it, looking into this...") before the agent starts processing. This leaves users uncertain about whether their message was received and is being worked on.
## What we tried
We attempted multipl... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-03T21:07:37Z | 2026-05-05T03:09:38Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | kmgroen | 161,303,532 | U_kgDOCZ1L7A | User | false |
openclaw/openclaw | 3,893,263,890 | I_kwDOQb6kR87oDn4S | 8,287 | https://github.com/openclaw/openclaw/issues/8287 | https://api.github.com/repos/openclaw/openclaw/issues/8287 | Feature Request: Node-Registered Agent Tools | # Feature Request: Node-Registered Agent Tools
## Summary
Allow connected nodes to register agent tools that the Gateway automatically exposes to the LLM. This would enable nodes to extend agent capabilities without requiring a separate plugin.
## Problem
Currently, when building integrations that bridge external s... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-03T21:12:28Z | 2026-05-05T04:07:08Z | null | CONTRIBUTOR | null | 20260505T060034Z | 2026-05-05T06:00:34Z | Oceanswave | 760,674 | MDQ6VXNlcjc2MDY3NA== | User | false |
openclaw/openclaw | 3,893,295,646 | I_kwDOQb6kR87oDvoe | 8,295 | https://github.com/openclaw/openclaw/issues/8295 | https://api.github.com/repos/openclaw/openclaw/issues/8295 | [Feature]: Add allowBots support for Telegram groups (parity with Discord/Slack) | Discord and Slack providers already support allowBots to enable bot-to-bot communication in channels/groups. Telegram groups don't have this option, which makes multi-bot setups (e.g. two MoltBot instances in the same Telegram group) unable to interact directly.
Use case: Two MoltBot agents in a shared Telegram group,... | open | null | false | 5 | [
"enhancement"
] | [] | 2026-02-03T21:20:53Z | 2026-05-05T00:08:19Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | arekqq | 19,227,389 | MDQ6VXNlcjE5MjI3Mzg5 | User | false |
openclaw/openclaw | 3,893,298,682 | I_kwDOQb6kR87oDwX6 | 8,299 | https://github.com/openclaw/openclaw/issues/8299 | https://api.github.com/repos/openclaw/openclaw/issues/8299 | Feature request: config option to suppress sub-agent announce | ## Problem
When using `sessions_spawn`, the sub-agent announce step runs after completion and posts a summary to the requester chat. The only way to suppress this is for the sub-agent model to reply exactly `ANNOUNCE_SKIP` during the announce step.
**The problem:** The sub-agent model frequently ignores the `ANNOUNCE... | open | null | false | 6 | [
"enhancement"
] | [] | 2026-02-03T21:21:44Z | 2026-05-05T08:50:21Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | jzOcb | 258,660,657 | U_kgDOD2rZMQ | User | false |
openclaw/openclaw | 3,893,437,100 | I_kwDOQb6kR87oESKs | 8,321 | https://github.com/openclaw/openclaw/issues/8321 | https://api.github.com/repos/openclaw/openclaw/issues/8321 | Feature: Per-job cache control for cron jobs | ## Problem
Currently, all isolated cron jobs inherit the global `agents.defaults.contextPruning` settings from the config. This means every isolated session caches context, even for simple script-running jobs that don't need LLM continuity.
**Impact:** High token costs from unnecessary caching
- Example: nightly migr... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-03T21:59:29Z | 2026-05-05T00:08:20Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | cortexuvula | 2,161,372 | MDQ6VXNlcjIxNjEzNzI= | User | false |
openclaw/openclaw | 3,894,301,252 | I_kwDOQb6kR87oHlJE | 8,495 | https://github.com/openclaw/openclaw/issues/8495 | https://api.github.com/repos/openclaw/openclaw/issues/8495 | Feature: Auto-leave WhatsApp groups not in allowlist + group-join webhook | ## Feature Request
### Problem
When using WhatsApp with `groupPolicy: allowlist` and `groupAllowFrom`, unauthorized groups are blocked from triggering agent sessions — but the bot remains a member of those groups. This creates a security/privacy concern: the bot is still technically present and visible to group member... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-04T03:06:03Z | 2026-05-05T11:54:17Z | null | NONE | null | 20260505T120033Z | 2026-05-05T12:00:33Z | motochan | 344,802 | MDQ6VXNlcjM0NDgwMg== | User | false |
openclaw/openclaw | 3,894,338,255 | I_kwDOQb6kR87oHuLP | 8,508 | https://github.com/openclaw/openclaw/issues/8508 | https://api.github.com/repos/openclaw/openclaw/issues/8508 | Feature: configurable/dynamic ack reaction emojis | Feature Request: Dynamic/Contextual Ack Reactions
Currently, OpenClaw uses a single fixed emoji for ackReaction (default 👀) to acknowledge messages before processing. This works but feels repetitive and robotic in group chats.
Proposal:
Allow the agent to choose contextually appropriate reactions based on message co... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-04T03:19:21Z | 2026-05-05T00:08:24Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | NubzTuna | 244,624,827 | U_kgDODpStuw | User | false |
openclaw/openclaw | 3,894,558,264 | I_kwDOQb6kR87oIj44 | 8,547 | https://github.com/openclaw/openclaw/issues/8547 | https://api.github.com/repos/openclaw/openclaw/issues/8547 | Feature Request: Font Size Configuration for Web UI | ## Feature Request
### Summary
Add configuration options for font size in the web chat interface to improve accessibility and user experience.
### Background
Currently, the Clawdbot/OpenClaw web interface uses fixed font sizes that cannot be adjusted by users. While browser zoom works as a workaround, native font siz... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-04T04:36:21Z | 2026-05-05T00:08:25Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | Hackerdomarketing | 213,138,820 | U_kgDODLQ9hA | User | false |
openclaw/openclaw | 3,894,887,034 | I_kwDOQb6kR87oJ0J6 | 8,606 | https://github.com/openclaw/openclaw/issues/8606 | https://api.github.com/repos/openclaw/openclaw/issues/8606 | Feature request: Hook event for context pruning | ## Summary
When context pruning runs (trimming old tool outputs), there's no hook event exposed. This makes it impossible to trigger follow-up actions like indexing the pruned content for searchability.
## Use Case
We want to maintain searchable access to content even after it's pruned from active context. The ideal... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-04T06:16:15Z | 2026-05-05T06:01:27Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | halthelobster | 257,232,409 | U_kgDOD1UOGQ | User | false |
openclaw/openclaw | 3,895,076,876 | I_kwDOQb6kR87oKigM | 8,635 | https://github.com/openclaw/openclaw/issues/8635 | https://api.github.com/repos/openclaw/openclaw/issues/8635 | Feature Request: Token Usage API for session_status Tool | ## Summary
Expose per-session token usage data through the `session_status` tool so agents can programmatically track and budget their token consumption.
## Current Behavior
- `session_status` shows percentage-based usage (`5h 28% left`, `Week 3% left`)
- Per-turn token data exists in session JSONL files (`.message.... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-04T07:20:43Z | 2026-05-05T00:08:27Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | saphid | 4,596,216 | MDQ6VXNlcjQ1OTYyMTY= | User | false |
openclaw/openclaw | 3,895,639,539 | I_kwDOQb6kR87oMr3z | 8,719 | https://github.com/openclaw/openclaw/issues/8719 | https://api.github.com/repos/openclaw/openclaw/issues/8719 | [Feature] OpenClaw Security Profile v1.1 (Data-centric, secure-by-default) | ## Summary
I’d like to propose a production-ready security model for OpenClaw that significantly reduces real-world risks (wallet theft, DB wipe, cookie exfiltration, plugin abuse) **without killing UX** and **without relying on LLM “judgment.”**
The core idea is **data-centric security with hard enforcement outside ... | open | null | false | 6 | [
"enhancement"
] | [] | 2026-02-04T09:48:57Z | 2026-05-05T03:09:39Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | Ar3ss12 | 150,044,748 | U_kgDOCPGATA | User | false |
openclaw/openclaw | 3,895,686,824 | I_kwDOQb6kR87oM3ao | 8,724 | https://github.com/openclaw/openclaw/issues/8724 | https://api.github.com/repos/openclaw/openclaw/issues/8724 | Feature: Add per-model generation timeout config | ## Problem
Google models (especially Gemini Flash) can get stuck in infinite thinking/planning loops without actually generating output or calling tools.
Example scenario:
- User asks: "Check my GitHub issues"
- Flash enters loop: "Analyzing..." → "Reviewing..." → "Evaluating..." (repeats 50+ times)
- Never calls `me... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-04T10:00:34Z | 2026-05-05T00:08:28Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | riskatcher | 257,757,463 | U_kgDOD10RFw | User | false |
openclaw/openclaw | 3,896,289,232 | I_kwDOQb6kR87oPKfQ | 8,812 | https://github.com/openclaw/openclaw/issues/8812 | https://api.github.com/repos/openclaw/openclaw/issues/8812 | [Feature Request] Auto-linkify URLs in tool output cards (Control UI & WebChat) | Problem
When tools like web_fetch, web_search, or browser return results containing URLs, those URLs appear as plain text in the chat interface and are not clickable.
Current behavior:
{
"url": "https://example.com/article",
"finalUrl": "https://example.com/article?ref=123",
"status": 200
}
The URLs above are d... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-04T12:24:07Z | 2026-05-05T09:52:02Z | null | NONE | null | 20260505T120033Z | 2026-05-05T12:00:33Z | applre | 10,095,151 | MDQ6VXNlcjEwMDk1MTUx | User | false |
openclaw/openclaw | 3,896,298,658 | I_kwDOQb6kR87oPMyi | 8,813 | https://github.com/openclaw/openclaw/issues/8813 | https://api.github.com/repos/openclaw/openclaw/issues/8813 | System prompt should prefer Skill tool over read when available | Core Problem:
OpenClaw treats skills as low-priority guidance rather than high-priority requirements in the way that Claude Code does.
Summary
When a Pi extension registers a tool called `Skill` (which replicates Claude Code's skill-loading behavior), the built-in skills section of the system prompt still tells the m... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-04T12:26:28Z | 2026-05-05T07:23:00Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | adamjdavidson | 47,649,576 | MDQ6VXNlcjQ3NjQ5NTc2 | User | false |
openclaw/openclaw | 3,897,106,465 | I_kwDOQb6kR87oSSAh | 8,892 | https://github.com/openclaw/openclaw/issues/8892 | https://api.github.com/repos/openclaw/openclaw/issues/8892 | Feature: --agent flag for TUI to select which agent handles the session | ## Summary
When running `openclaw tui`, there's currently no way to directly connect to a specific agent defined in `agents.list`. The `--session` flag only sets the session key/name, but the conversation is still handled by `agent main`.
## Current Behavior
```bash
openclaw tui --session agent:ollama
```
Results i... | open | null | false | 4 | [
"enhancement",
"tui"
] | [] | 2026-02-04T15:28:42Z | 2026-05-05T04:22:41Z | null | CONTRIBUTOR | null | 20260505T060034Z | 2026-05-05T06:00:34Z | nomadonwheels196 | 36,867,340 | MDQ6VXNlcjM2ODY3MzQw | User | false |
openclaw/openclaw | 3,897,225,279 | I_kwDOQb6kR87oSvA_ | 8,913 | https://github.com/openclaw/openclaw/issues/8913 | https://api.github.com/repos/openclaw/openclaw/issues/8913 | [Feature] Configurable thinking/reasoning block format per channel | ## Summary
Add a configuration option to customize how thinking/reasoning blocks are formatted per channel.
## Problem
On Telegram, reasoning blocks are rendered in italics (`_text_`), which is hard to read — especially for longer reasoning. The TUI console shows them in green, which is much more visible.
## Proposed... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-04T15:54:57Z | 2026-05-05T08:11:54Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | lucletoffe | 15,689,941 | MDQ6VXNlcjE1Njg5OTQx | User | false |
openclaw/openclaw | 3,897,571,561 | I_kwDOQb6kR87oUDjp | 8,959 | https://github.com/openclaw/openclaw/issues/8959 | https://api.github.com/repos/openclaw/openclaw/issues/8959 | Feature: Cross-platform multi-select component in message tool | ## Summary
Add a cross-platform **multi-select** capability to the `message` tool, allowing agents to present a list of options that users can batch-select and submit in **a single interaction** — instead of triggering a full agent turn per click.
## Problem
Currently, the `message` tool's `buttons` parameter only s... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-04T17:15:35Z | 2026-05-05T04:30:47Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | behindthebeat | 8,576,839 | MDQ6VXNlcjg1NzY4Mzk= | User | false |
openclaw/openclaw | 3,897,613,299 | I_kwDOQb6kR87oUNvz | 8,969 | https://github.com/openclaw/openclaw/issues/8969 | https://api.github.com/repos/openclaw/openclaw/issues/8969 | Enhancement: Developer experience quick wins from code analysis | ## Summary
Three low-hanging improvements identified during code analysis that would improve debugging and operator experience.
---
### 1. Add `skills.validate` CLI command
**What:** A command to report missing bins/env/config requirements for skills.
**Why:** The logic already exists in `shouldIncludeSkill` (`dist... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-04T17:26:50Z | 2026-05-05T00:08:29Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | miguelmartillo69-design | 259,191,127 | U_kgDOD3LxVw | User | false |
openclaw/openclaw | 3,897,622,769 | I_kwDOQb6kR87oUQDx | 8,972 | https://github.com/openclaw/openclaw/issues/8972 | https://api.github.com/repos/openclaw/openclaw/issues/8972 | [Feature]: Pattern-based channel allowlist for Discord (glob/wildcard support) | ## Summary
Add support for pattern-based channel names in the Discord allowlist, so agents can listen to dynamically-created channels without requiring a config patch + restart for each one.
## Use Case
When building work-stream workflows that spin off dedicated Discord channels (e.g., `#ws-task-1`, `#ws-task-2`), t... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-04T17:29:20Z | 2026-05-05T10:46:44Z | null | NONE | null | 20260505T120033Z | 2026-05-05T12:00:33Z | milldr | 14,060,048 | MDQ6VXNlcjE0MDYwMDQ4 | User | false |
openclaw/openclaw | 3,898,027,626 | I_kwDOQb6kR87oVy5q | 9,016 | https://github.com/openclaw/openclaw/issues/9016 | https://api.github.com/repos/openclaw/openclaw/issues/9016 | Feature: Expose OpenRouter usage cost to agent runtime | ## Feature Request: Expose OpenRouter Usage Cost to Agent Runtime
### Summary
Add per-message cost tracking from OpenRouter API responses and expose it to agents so they can append cost information to replies.
### Background
OpenRouter includes detailed usage information in every API response, including:
- Token coun... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-04T19:12:13Z | 2026-05-05T04:38:01Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | eric-wien | 63,148,689 | MDQ6VXNlcjYzMTQ4Njg5 | User | false |
openclaw/openclaw | 3,898,614,149 | I_kwDOQb6kR87oYCGF | 9,120 | https://github.com/openclaw/openclaw/issues/9120 | https://api.github.com/repos/openclaw/openclaw/issues/9120 | Feature: Message continuation marker for batched input (cost savings) | ## Problem
When a user sends multiple messages in quick succession (e.g., providing context across several messages), **each message triggers a separate agent turn** with full context loading, costing input tokens unnecessarily.
**Current behavior:**
- Message 1 "(cont)" → Agent processes, outputs NO_REPLY (but pays ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-04T22:02:19Z | 2026-05-05T00:08:31Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | FingerCramp | 14,649,980 | MDQ6VXNlcjE0NjQ5OTgw | User | false |
openclaw/openclaw | 3,898,688,486 | I_kwDOQb6kR87oYUPm | 9,132 | https://github.com/openclaw/openclaw/issues/9132 | https://api.github.com/repos/openclaw/openclaw/issues/9132 | Control UI: Add status indicator for sessions (active/completed/timed out) | **Problem**
The Control UI sessions list shows all sessions (main, hook, cron, subagent) but there's no visual indicator of whether a session is currently active, completed, or timed out. This makes it hard to distinguish the live main session from dozens of finished hook/cron/subagent sessions.
**Suggestion**
Add a s... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-04T22:24:07Z | 2026-05-05T04:44:41Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | Claudius-the-II | 259,450,935 | U_kgDOD3boNw | User | false |
openclaw/openclaw | 3,898,802,315 | I_kwDOQb6kR87oYwCL | 9,155 | https://github.com/openclaw/openclaw/issues/9155 | https://api.github.com/repos/openclaw/openclaw/issues/9155 | Feature Request: cron.defaults.delivery config for default delivery target | ## Summary
Add a `cron.defaults.delivery` config option so users can set a default delivery channel/target for all cron jobs, avoiding repetitive per-job configuration.
## Motivation
Currently, every isolated cron job with `delivery.mode: "announce"` requires explicit `delivery.channel` and `delivery.to` fields. Wit... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-04T23:05:09Z | 2026-05-05T06:15:09Z | null | CONTRIBUTOR | null | 20260505T090045Z | 2026-05-05T09:00:45Z | gildo | 133,645 | MDQ6VXNlcjEzMzY0NQ== | User | false |
openclaw/openclaw | 3,899,762,150 | I_kwDOQb6kR87ocaXm | 9,394 | https://github.com/openclaw/openclaw/issues/9394 | https://api.github.com/repos/openclaw/openclaw/issues/9394 | [imessage] Support rich text formatting and effects | ## Feature Request
iMessage channel should support sending rich text formatting and effects (bold, italics, underline, strikethrough, text size, and animations like Shake/Nod/Explode/Ripple/Bloom/Jitter).
**Current Behavior**
The imsg CLI tool only supports plain text. Messages sent through the OpenClaw iMessage chan... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T05:32:26Z | 2026-05-05T10:54:10Z | null | NONE | null | 20260505T120033Z | 2026-05-05T12:00:33Z | marvclaw | 259,501,594 | U_kgDOD3euGg | User | false |
openclaw/openclaw | 3,899,857,433 | I_kwDOQb6kR87ocxoZ | 9,409 | https://github.com/openclaw/openclaw/issues/9409 | https://api.github.com/repos/openclaw/openclaw/issues/9409 | Improve context overflow error message with specifics | ## Current Behavior
When a session exceeds the model's context limit, the user sees:
```
Context overflow: prompt too large for the model. Try again with less input or a larger-context model.
```
This message lacks critical information for diagnosis.
## Proposed Improvement
```
Context overflow in #channel-name (s... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T06:02:56Z | 2026-05-05T10:22:36Z | null | CONTRIBUTOR | null | 20260505T120033Z | 2026-05-05T12:00:33Z | liaosvcaf | 51,533,973 | MDQ6VXNlcjUxNTMzOTcz | User | false |
openclaw/openclaw | 3,899,989,050 | I_kwDOQb6kR87odRw6 | 9,443 | https://github.com/openclaw/openclaw/issues/9443 | https://api.github.com/repos/openclaw/openclaw/issues/9443 | Request: Prebuilt Android APK releases | *This issue is submitted on behalf of Lysen, by his AI assistant QING.*
As a user who wants to run the OpenClaw Android companion app, I would like to have prebuilt APK downloads available in GitHub releases.
Currently, the repository includes Android source code in `apps/android`, but no precompiled APK is distribut... | open | null | false | 23 | [
"enhancement"
] | [] | 2026-02-05T06:51:08Z | 2026-05-05T07:50:22Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | AstridQing-AI | 259,513,701 | U_kgDOD3fdZQ | User | false |
openclaw/openclaw | 3,900,092,187 | I_kwDOQb6kR87odq8b | 9,455 | https://github.com/openclaw/openclaw/issues/9455 | https://api.github.com/repos/openclaw/openclaw/issues/9455 | [Feature]: Cron Jobs UI: add filter bar above Jobs list (status/kind/search) | ## Summary
Please add a filter bar above the Cron Jobs list in the dashboard/UI so users can quickly filter and search jobs as the list grows.
## Problem
The Cron Jobs list becomes hard to manage once there are many jobs (recurring + one-shot + disabled). Users need a quick way to narrow the list to what matters (acti... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T07:22:40Z | 2026-05-05T07:42:45Z | null | CONTRIBUTOR | null | 20260505T090045Z | 2026-05-05T09:00:45Z | guxu11 | 53,551,744 | MDQ6VXNlcjUzNTUxNzQ0 | User | false |
openclaw/openclaw | 3,900,155,962 | I_kwDOQb6kR87od6g6 | 9,465 | https://github.com/openclaw/openclaw/issues/9465 | https://api.github.com/repos/openclaw/openclaw/issues/9465 | Feature Request: Cron Job Hooks System | ## Summary
Add a hooks system to cron jobs that allows executing custom logic before and after trigger execution. This enables advanced use cases like forced reminders, escalation policies, and conditional notifications.
## Motivation
Currently, cron jobs are fire-and-forget. Users need more control:
1. **Forced Re... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-05T07:41:41Z | 2026-05-05T06:37:02Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | guyhuang | 14,291,489 | MDQ6VXNlcjE0MjkxNDg5 | User | false |
openclaw/openclaw | 3,900,780,285 | I_kwDOQb6kR87ogS79 | 9,544 | https://github.com/openclaw/openclaw/issues/9544 | https://api.github.com/repos/openclaw/openclaw/issues/9544 | [Feature]: Move away from useMultiFileAuthState for Baileys | ## Summary
The useMultiFileAuthState in Baileys was never intended for production. It uses JSON files to store signal keys which depending on the way you deploy this exposes the project to corrupted sessions, ever inflating CPU/IO usage and even security risks.
## Proposed solution
Implement a custom auth state tha... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T10:15:44Z | 2026-05-05T10:15:40Z | null | NONE | null | 20260505T120033Z | 2026-05-05T12:00:33Z | purpshell | 52,720,753 | MDQ6VXNlcjUyNzIwNzUz | User | false |
openclaw/openclaw | 3,900,790,352 | I_kwDOQb6kR87ogVZQ | 9,546 | https://github.com/openclaw/openclaw/issues/9546 | https://api.github.com/repos/openclaw/openclaw/issues/9546 | [Feature Request] Add option to disable 'new messages' notification banner in WebChat | ## Description
The WebChat UI shows a 'new messages' notification banner at the top of the chat screen whenever new messages arrive. This banner obstructs the chat view and disrupts the user experience, especially when reading long responses.
## Current Behavior
When using WebChat, a 'new messages' notification bann... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T10:18:05Z | 2026-05-05T04:51:31Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | cheqianyi | 259,542,359 | U_kgDOD3hNVw | User | false |
openclaw/openclaw | 3,901,437,710 | I_kwDOQb6kR87oizcO | 9,607 | https://github.com/openclaw/openclaw/issues/9607 | https://api.github.com/repos/openclaw/openclaw/issues/9607 | Himalaya skill: missing email formatting philosophy and incorrect command syntax | ## Summary
The Himalaya email skill documentation has several gaps and inaccuracies that can lead to confusion:
### Issues Found
1. **Missing email formatting philosophy** - No clear guidance on when to use plain text vs styled HTML emails
2. **Missing HTML email reference** - The SKILL.md references `html-email.md`... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T12:55:21Z | 2026-05-05T04:59:53Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | frankdierolf | 207,307,187 | U_kgDODFtBsw | User | false |
openclaw/openclaw | 3,901,799,075 | I_kwDOQb6kR87okLqj | 9,637 | https://github.com/openclaw/openclaw/issues/9637 | https://api.github.com/repos/openclaw/openclaw/issues/9637 | Add accessibility config option to disable emojis and unicode symbols in TUI | ## Problem
The OpenClaw TUI uses emojis and unicode symbols (box-drawing characters, arrows) that create a poor experience for screenreader users. These symbols generate verbose, confusing output that makes navigating the interface difficult.
## Investigation
I've identified the sources of unicode/emoji symbols in t... | open | null | false | 5 | [
"enhancement",
"tui"
] | [] | 2026-02-05T14:13:07Z | 2026-05-05T06:44:31Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | robin24 | 8,311,608 | MDQ6VXNlcjgzMTE2MDg= | User | false |
openclaw/openclaw | 3,901,963,814 | I_kwDOQb6kR87okz4m | 9,656 | https://github.com/openclaw/openclaw/issues/9656 | https://api.github.com/repos/openclaw/openclaw/issues/9656 | Feature: Handle Signal message edits (editMessage envelope) | ## Problem
When a Signal message is edited, signal-cli provides an `editMessage` envelope containing:
- `targetSentTimestamp` — the original message timestamp being edited
- `dataMessage` — the new message content
Currently, OpenClaw treats edited messages as entirely new messages, which:
1. Creates duplicate context... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T14:48:57Z | 2026-05-05T09:43:35Z | null | NONE | null | 20260505T120033Z | 2026-05-05T12:00:33Z | NOVA-Openclaw | 258,138,572 | U_kgDOD2LhzA | User | false |
openclaw/openclaw | 3,901,980,444 | I_kwDOQb6kR87ok38c | 9,658 | https://github.com/openclaw/openclaw/issues/9658 | https://api.github.com/repos/openclaw/openclaw/issues/9658 | Feature Request: Support streaming or file-based tool arguments to bypass JSON serialization limits | # Feature Request: Support streaming or file-based tool arguments to bypass JSON serialization limits
## Summary
Tool calls with large arguments (>10KB) fail with "Bad escaped character in JSON" errors because the entire command string must be JSON-serialized. This affects `exec`, `write`, and other tools when workin... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T14:52:15Z | 2026-05-05T03:14:56Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | zlxdengtianhei | 131,322,965 | U_kgDOB9PUVQ | User | false |
openclaw/openclaw | 3,902,594,051 | I_kwDOQb6kR87onNwD | 9,764 | https://github.com/openclaw/openclaw/issues/9764 | https://api.github.com/repos/openclaw/openclaw/issues/9764 | Google Chat: Add user OAuth support for reactions and media uploads | ## Summary
The Google Chat channel currently only supports Service Account authentication (scope `chat.bot`). This limits functionality:
- ❌ **Reactions** — `spaces.messages.reactions.create` requires user-level OAuth scopes
- ❌ **Media uploads** — `attachments:upload` fails with service account auth
- ❌ **Proactiv... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-05T17:00:04Z | 2026-05-05T05:07:22Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | ndohuu | 1,586,139 | MDQ6VXNlcjE1ODYxMzk= | User | false |
openclaw/openclaw | 3,902,850,370 | I_kwDOQb6kR87ooMVC | 9,797 | https://github.com/openclaw/openclaw/issues/9797 | https://api.github.com/repos/openclaw/openclaw/issues/9797 | [Feature]: `queue_status` tool for intelligent task dispatch | ## Summary
Agents can spawn subagents via `sessions_spawn`, but have no visibility into queue/lane status. This makes intelligent task dispatch impossible — agents must blindly spawn without knowing if the target agent is already overloaded with queued tasks.
Without this visibility, users cannot build a "Dispatcher ... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-05T18:01:22Z | 2026-05-05T00:08:32Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | pahud | 278,432 | MDQ6VXNlcjI3ODQzMg== | User | false |
openclaw/openclaw | 3,902,945,975 | I_kwDOQb6kR87oojq3 | 9,820 | https://github.com/openclaw/openclaw/issues/9820 | https://api.github.com/repos/openclaw/openclaw/issues/9820 | feat: add jsonl support to memory indexer | ## problem
memory indexer only supports `.md` files:
```typescript
// internal.ts line 96-99
if (!absPath.endsWith(".md")) {
return;
}
```
this means jsonl files in `extraPaths` are ignored.
## use case
raw message logs stored as jsonl (via message:received/message:sent hooks) cannot be indexed for semantic sear... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-05T18:28:57Z | 2026-05-05T06:22:53Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | matskevich | 7,224,690 | MDQ6VXNlcjcyMjQ2OTA= | User | false |
openclaw/openclaw | 3,903,054,311 | I_kwDOQb6kR87oo-Hn | 9,835 | https://github.com/openclaw/openclaw/issues/9835 | https://api.github.com/repos/openclaw/openclaw/issues/9835 | Feature: Telegram Reactions - Two-Way Support with Prompt Guidance | Add bidirectional Telegram reaction support to OpenClaw with prompt-based guidance.
CONTEXT:
- OpenClaw already has `reactionNotifications` config (values: "off", "own", "all")
- When enabled, incoming reactions appear as system events: "Telegram reaction added: {emoji} by {user} on msg {id}"
- The `message` tool al... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T18:56:14Z | 2026-05-05T03:09:42Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | sepivip | 72,915,556 | MDQ6VXNlcjcyOTE1NTU2 | User | false |
openclaw/openclaw | 3,903,291,192 | I_kwDOQb6kR87op384 | 9,865 | https://github.com/openclaw/openclaw/issues/9865 | https://api.github.com/repos/openclaw/openclaw/issues/9865 | Feature Request: Batch API support for background tasks | ## Use Case
For non-real-time background work (heartbeats, cron jobs, scheduled maintenance), batch APIs from providers like OpenAI and Anthropic offer significant cost savings (typically 50% off) with acceptable latency tradeoffs.
## Proposal
Add support for batch API endpoints that could be:
- Opted into for speci... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-05T19:48:34Z | 2026-05-05T11:02:10Z | null | NONE | null | 20260505T120033Z | 2026-05-05T12:00:33Z | diffskybot | 259,501,990 | U_kgDOD3evpg | User | false |
openclaw/openclaw | 3,903,380,690 | I_kwDOQb6kR87oqNzS | 9,876 | https://github.com/openclaw/openclaw/issues/9876 | https://api.github.com/repos/openclaw/openclaw/issues/9876 | [Feature]: Show session and requester context in macOS exec approval popup | ## Summary
The macOS companion app's exec approval popup (`ExecApprovalsPromptPresenter`) shows the command, working directory, agent ID, executable path, host, and security/ask mode — but does not show which session triggered the request or any sender/channel context. When multiple sessions or channels are active, th... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T20:09:31Z | 2026-05-05T00:08:35Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | leiyangyou | 104,397 | MDQ6VXNlcjEwNDM5Nw== | User | false |
openclaw/openclaw | 3,903,712,278 | I_kwDOQb6kR87orewW | 9,912 | https://github.com/openclaw/openclaw/issues/9912 | https://api.github.com/repos/openclaw/openclaw/issues/9912 | Feature: Add maxTurns/maxToolCalls config option to limit agent iterations | ## Summary
Request a `maxTurns` or `maxToolCalls` configuration option to limit the number of tool call iterations an agent can make before being forced to respond.
## Use Case
When using certain LLM providers/models (e.g., KIMI K2 via NVIDIA API), the model may ignore system prompt instructions to stop calling tool... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T21:38:06Z | 2026-05-05T11:16:31Z | null | NONE | null | 20260505T120033Z | 2026-05-05T12:00:33Z | AlexGn | 5,334,352 | MDQ6VXNlcjUzMzQzNTI= | User | false |
openclaw/openclaw | 3,904,101,856 | I_kwDOQb6kR87os93g | 9,986 | https://github.com/openclaw/openclaw/issues/9986 | https://api.github.com/repos/openclaw/openclaw/issues/9986 | Feature: Trigger model fallback on context length exceeded | ## Problem
When the primary model's context limit is exceeded, OpenClaw currently freezes/errors instead of falling back to a configured fallback model.
The `agents.defaults.model.fallbacks` config exists, but fallback only triggers on API errors (529 overloaded, timeouts, etc.) — not on context overflow.
## Use Cas... | open | null | false | 5 | [
"enhancement"
] | [] | 2026-02-05T23:48:29Z | 2026-05-05T11:09:03Z | null | NONE | null | 20260505T120033Z | 2026-05-05T12:00:33Z | amirarenco-create | 258,618,597 | U_kgDOD2o05Q | User | false |
openclaw/openclaw | 3,904,118,866 | I_kwDOQb6kR87otCBS | 9,987 | https://github.com/openclaw/openclaw/issues/9987 | https://api.github.com/repos/openclaw/openclaw/issues/9987 | [Feature]: CLI command to list pending exec approval requests | ## Summary
Add a CLI command like `openclaw approvals pending` to list currently pending exec approval requests, including the command being requested.
## Use case
When running a headless agent (e.g., Molty on Discord), pending approvals show in the UI/Discord as:
```
exec - Approval required (id 8332daed)
```
But ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-05T23:54:31Z | 2026-05-05T05:22:27Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | schickling | 1,567,498 | MDQ6VXNlcjE1Njc0OTg= | User | false |
openclaw/openclaw | 3,904,177,821 | I_kwDOQb6kR87otQad | 9,993 | https://github.com/openclaw/openclaw/issues/9993 | https://api.github.com/repos/openclaw/openclaw/issues/9993 | Feature Request: config:pre-apply hook for validating config changes | ## Problem
When an AI agent uses the `gateway config.patch` tool to modify OpenClaw config, there's no validation layer. A bad config (wrong model name format, missing required fields, invalid values) can crash the gateway immediately.
**Real incident (2026-02-05):**
- Agent tried to upgrade from Claude Opus 4.5 to 4... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-06T00:20:20Z | 2026-05-05T05:29:03Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | jzOcb | 258,660,657 | U_kgDOD2rZMQ | User | false |
openclaw/openclaw | 3,904,220,426 | I_kwDOQb6kR87ota0K | 10,005 | https://github.com/openclaw/openclaw/issues/10005 | https://api.github.com/repos/openclaw/openclaw/issues/10005 | Add option to hide session status card | ## Feature Request: Add option to hide session status card
### Use Case
When running `/status` or calling `session_status` tool, OpenClaw returns a rendered status card (runtime, tokens, model, etc.). While useful for debugging, this adds visual clutter in regular conversations.
### Proposed Solution
Add a configurat... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-06T00:41:35Z | 2026-05-05T05:35:23Z | null | NONE | null | 20260505T060034Z | 2026-05-05T06:00:34Z | kiss2223320-cmyk | 259,466,562 | U_kgDOD3clQg | User | false |
openclaw/openclaw | 4,329,802,420 | I_kwDOQb6kR88AAAABAhOKtA | 71,889 | https://github.com/openclaw/openclaw/issues/71889 | https://api.github.com/repos/openclaw/openclaw/issues/71889 | [Bug]: agents add wizard pre-fills nested workspace path (workspace/<id>) instead of documented peer-level (workspace-<id>) | ### Bug type
Behavior bug (incorrect output/state without crash)
### Beta release blocker
No
### Summary
The interactive `agents add <id>` wizard pre-fills the "Workspace directory"
prompt with a nested path under the main agent's workspace
(`~/.openclaw/workspace/<id>`) instead of the peer-level form
(`~/.open... | open | null | false | 1 | [
"bug"
] | [] | 2026-04-26T02:05:38Z | 2026-05-03T02:15:13Z | null | NONE | null | 20260503T030030Z | 2026-05-03T03:00:30Z | pushpendrachauhan | 8,198,972 | MDQ6VXNlcjgxOTg5NzI= | User | false |
openclaw/openclaw | 3,904,350,203 | I_kwDOQb6kR87ot6f7 | 10,029 | https://github.com/openclaw/openclaw/issues/10029 | https://api.github.com/repos/openclaw/openclaw/issues/10029 | [Feature]: Add syntax highlighting for code blocks in WebChat | ## Summary
When sharing code snippets in OpenClaw's WebChat interface, code blocks appear as plain monospace text without any syntax highlighting. This makes it difficult to quickly read and understand code structure, especially for languages like C++ with complex syntax. Other platforms (Discord, GitHub, etc.) automa... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-06T01:32:26Z | 2026-05-04T23:11:04Z | null | NONE | null | 20260505T000034Z | 2026-05-05T00:00:34Z | cxxzsh | 62,886,679 | MDQ6VXNlcjYyODg2Njc5 | User | false |
openclaw/openclaw | 4,329,806,247 | I_kwDOQb6kR88AAAABAhOZpw | 71,890 | https://github.com/openclaw/openclaw/issues/71890 | https://api.github.com/repos/openclaw/openclaw/issues/71890 | [Bug] Web UI: Duplicate messages after WebSocket reconnect (2026.4.24) |
## Bug Description
After gateway restart or WebSocket reconnect, messages in the web UI (openclaw-control-ui) are displayed twice — once labeled as 'openclaw-control-ui' and once as 'You' — with identical content and timestamps.
## Steps to Reproduce
1. Open web UI (http://127.0.0.1:18789/)
2. Send a message via web ... | closed | completed | true | 2 | [] | [] | 2026-04-26T02:08:07Z | 2026-04-29T04:42:26Z | 2026-04-26T07:06:12Z | NONE | null | 20260429T060025Z | 2026-04-29T06:00:25Z | Mibslee | 63,028,878 | MDQ6VXNlcjYzMDI4ODc4 | User | false |
openclaw/openclaw | 3,904,821,498 | I_kwDOQb6kR87ovtj6 | 10,118 | https://github.com/openclaw/openclaw/issues/10118 | https://api.github.com/repos/openclaw/openclaw/issues/10118 | TUI: Support Shift+Enter for newline (Enter to send) | ## Summary
In `openclaw tui`, Enter currently sends the message immediately. Please add support for composing multi-line messages with `Shift+Enter` (or a configurable keybinding), while keeping Enter as send.
## Why
When using terminal chat heavily, users often need multi-line prompts/instructions. Right now this is ... | open | null | false | 4 | [
"enhancement",
"tui"
] | [] | 2026-02-06T04:36:30Z | 2026-05-05T06:53:05Z | null | NONE | null | 20260505T090045Z | 2026-05-05T09:00:45Z | david-wooo | 58,894,358 | MDQ6VXNlcjU4ODk0MzU4 | User | false |
openclaw/openclaw | 3,904,967,934 | I_kwDOQb6kR87owRT- | 10,137 | https://github.com/openclaw/openclaw/issues/10137 | https://api.github.com/repos/openclaw/openclaw/issues/10137 | api=openai-responses (Codex/Responses): gateway online but no replies; repeated 400 (no body) — request normalization / compat options | ## Summary\nWhen using the Codex/Responses integration path (), the gateway can appear healthy/connected but produce no assistant output. Logs repeatedly show . This is commonly encountered when using third-party proxy/relay (Responses-compatible) endpoints.\n\n## Repro\n1. Configure a provider/model with and set it a... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-06T05:28:30Z | 2026-05-05T01:18:05Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | willgeek | 15,921,581 | MDQ6VXNlcjE1OTIxNTgx | User | false |
openclaw/openclaw | 4,329,762,622 | I_kwDOQb6kR88AAAABAhLvPg | 71,879 | https://github.com/openclaw/openclaw/issues/71879 | https://api.github.com/repos/openclaw/openclaw/issues/71879 | [Bug]: Docker gateway can crash-loop when Bonjour/ciao runs on bridge networking | ## Bug type
Crash (process/app exits or hangs)
## Beta release blocker
No — this is not a plugin beta-release blocker per the issue template wording, but it is potentially wide-impact for Docker users because it affects the default Docker bridge networking path.
## Summary
After upgrading a Docker Compose gateway ... | closed | completed | true | 3 | [] | [] | 2026-04-26T01:41:03Z | 2026-04-30T04:41:40Z | 2026-04-26T04:53:56Z | NONE | null | 20260430T060018Z | 2026-04-30T06:00:18Z | gbballpack | 39,963,341 | MDQ6VXNlcjM5OTYzMzQx | User | false |
openclaw/openclaw | 3,904,992,655 | I_kwDOQb6kR87owXWP | 10,142 | https://github.com/openclaw/openclaw/issues/10142 | https://api.github.com/repos/openclaw/openclaw/issues/10142 | Feature: session:end internal hook event | ## Summary
Add a `session:end` internal hook event that fires when a session completes.
## Use Case
Integrating OpenClaw with workflow orchestration systems like Temporal. When an agent session completes, we want to signal back to a waiting workflow so it can proceed to the next step.
Current workaround requires ag... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-06T05:35:25Z | 2026-05-05T01:18:06Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | spk-alex | 205,000,749 | U_kgDODDgQLQ | User | false |
openclaw/openclaw | 3,905,297,963 | I_kwDOQb6kR87oxh4r | 10,213 | https://github.com/openclaw/openclaw/issues/10213 | https://api.github.com/repos/openclaw/openclaw/issues/10213 | feat: Add Anthropic server-side compaction API support | ## Summary
Anthropic's Opus 4.6 introduced a new server-side compaction API that handles context summarization automatically. OpenClaw currently has its own client-side compaction, but should support the new API option for better performance and summarization quality.
## Motivation
Server-side compaction offers seve... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-06T07:09:21Z | 2026-05-05T03:09:44Z | null | CONTRIBUTOR | null | 20260505T060034Z | 2026-05-05T06:00:34Z | GodsBoy | 5,792,287 | MDQ6VXNlcjU3OTIyODc= | User | false |
openclaw/openclaw | 3,905,514,574 | I_kwDOQb6kR87oyWxO | 10,253 | https://github.com/openclaw/openclaw/issues/10253 | https://api.github.com/repos/openclaw/openclaw/issues/10253 | [Feature]: Configurable auto-cleanup and archival policy for webhook/hook sessions | ## Summary
Hook-triggered sessions (for example Gmail webhook sessions) remain stored indefinitely after they complete. They accumulate in the session store and transcript files without automatic archival or cleanup. Over time this produces storage growth and UI clutter in `sessions_list`.
## Proposed solution
Add a... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-06T08:14:55Z | 2026-05-05T01:18:08Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | cazbrat | 59,489,879 | MDQ6VXNlcjU5NDg5ODc5 | User | false |
openclaw/openclaw | 3,906,227,516 | I_kwDOQb6kR87o1E08 | 10,354 | https://github.com/openclaw/openclaw/issues/10354 | https://api.github.com/repos/openclaw/openclaw/issues/10354 | [Feature]: Add description and enum to message tool's channel parameter | ## Summary
The `message` tool's `channel` parameter (`src/agents/tools/message-tool.ts`, line 40 in `buildRoutingSchema`) is defined as a bare `Type.Optional(Type.String())` with no description or enum constraint. This causes AI agents to pass Slack/Discord channel IDs (e.g. `C07ABC1234X`) instead of provider names (e... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-06T11:10:05Z | 2026-05-05T01:18:09Z | null | CONTRIBUTOR | null | 20260505T030055Z | 2026-05-05T03:00:55Z | carrotRakko | 24,588,751 | MDQ6VXNlcjI0NTg4NzUx | User | false |
openclaw/openclaw | 3,907,144,077 | I_kwDOQb6kR87o4kmN | 10,467 | https://github.com/openclaw/openclaw/issues/10467 | https://api.github.com/repos/openclaw/openclaw/issues/10467 | [Feature Request]: Multi-lane concurrency support for sub-agents via sessions_spawn | ## Summary
Currently, all sub-agents spawned via the `sessions_spawn` tool are funneled into a single global queue lane: `subagent`. The concurrency is governed by a single limit `agents.defaults.subagents.maxConcurrent`.
This creates a bottleneck and reliability risk for complex multi-agent workflows (e.g., 'Evoluti... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-06T14:54:17Z | 2026-05-05T01:18:10Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | lexobe | 148,517,205 | U_kgDOCNoxVQ | User | false |
openclaw/openclaw | 3,907,223,481 | I_kwDOQb6kR87o43-5 | 10,480 | https://github.com/openclaw/openclaw/issues/10480 | https://api.github.com/repos/openclaw/openclaw/issues/10480 | Support Workers AI model selection during onboard | ## Problem
When using `openclaw onboard --non-interactive` with `--auth-choice cloudflare-ai-gateway-api-key`, there is no way to specify which models to use. This means users who want to use Cloudflare Workers AI models (e.g. `@cf/meta/llama-3.3-70b-instruct-fp8-fast`) through AI Gateway are left with whatever defaul... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-06T15:14:58Z | 2026-05-05T01:18:12Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | sidharthachatterjee | 7,701,981 | MDQ6VXNlcjc3MDE5ODE= | User | false |
openclaw/openclaw | 3,907,465,609 | I_kwDOQb6kR87o5zGJ | 10,519 | https://github.com/openclaw/openclaw/issues/10519 | https://api.github.com/repos/openclaw/openclaw/issues/10519 | Telegram: Add language class to code blocks for native copy button support | ## Feature Request
**Summary:** Add language class to Telegram HTML code blocks to enable native copy buttons
### Current Behavior
When sending messages with code blocks to Telegram, OpenClaw converts markdown fenced code blocks to HTML:
```markdown
```bash
echo "hello"
```
```
Becomes:
```html
<pre><code>echo "hel... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-06T16:20:29Z | 2026-05-05T01:18:13Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | botko1botko | 257,185,516 | U_kgDOD1RW7A | User | false |
openclaw/openclaw | 3,908,317,659 | I_kwDOQb6kR87o9DHb | 10,659 | https://github.com/openclaw/openclaw/issues/10659 | https://api.github.com/repos/openclaw/openclaw/issues/10659 | Feature Request: Masked Secrets - Prevent Agent from Accessing Raw API Keys | ## Summary
Add a "masked secrets" system that allows agents to **use** API keys without being able to **see** them. This prevents accidental leaks and protects against prompt injection attacks designed to extract credentials.
## Problem
Currently, secrets stored in `~/.openclaw/.env` are fully accessible to agents:
... | open | null | false | 12 | [
"enhancement"
] | [] | 2026-02-06T20:26:06Z | 2026-05-05T01:18:15Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | jmkritt | 8,386,890 | MDQ6VXNlcjgzODY4OTA= | User | false |
openclaw/openclaw | 3,908,682,072 | I_kwDOQb6kR87o-cFY | 10,737 | https://github.com/openclaw/openclaw/issues/10737 | https://api.github.com/repos/openclaw/openclaw/issues/10737 | Feature Request: Auto-acknowledgment for iMessage channel | ## Summary
When a message arrives via iMessage, automatically send a brief acknowledgment (e.g., "…") before the agent starts processing. This simulates a typing indicator since Apple doesn't expose typing indicators programmatically.
## Use Case
- User sends iMessage to agent
- Agent immediately sends "…" (within mil... | open | null | false | 5 | [
"enhancement"
] | [] | 2026-02-06T22:33:38Z | 2026-05-05T01:18:16Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | steve-glaser-ops | 258,502,852 | U_kgDOD2hwxA | User | false |
openclaw/openclaw | 3,909,317,275 | I_kwDOQb6kR87pA3Kb | 10,872 | https://github.com/openclaw/openclaw/issues/10872 | https://api.github.com/repos/openclaw/openclaw/issues/10872 | iMessage: Include reply/thread context in incoming messages | iMessage: Include reply/thread context in incoming messages
## Labels
`enhancement`, `channel:imessage`
## Body
### Problem
When a user replies to a specific message in iMessage (using the inline reply feature), the OpenClaw agent receives **only the new message text** with no context about which message was being... | open | null | false | 4 | [
"enhancement"
] | [] | 2026-02-07T04:03:44Z | 2026-05-05T01:18:18Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | r80hk | 259,976,724 | U_kgDOD37uFA | User | false |
openclaw/openclaw | 3,909,552,179 | I_kwDOQb6kR87pBwgz | 10,944 | https://github.com/openclaw/openclaw/issues/10944 | https://api.github.com/repos/openclaw/openclaw/issues/10944 | Feature Request: Add parseMode config for Telegram channels | ## Problem
OpenClaw currently hardcodes `parse_mode: 'Markdown'` for Telegram messages. This causes issues:
- Some emoji render incorrectly as �� characters
- Markdown syntax can break formatted text
- No way to use plain text or HTML mode
## Proposed Solution
Add optional `parseMode` config for Telegram channels:
`... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-07T06:25:38Z | 2026-05-05T02:55:22Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | Ratzzz33 | 47,256,936 | MDQ6VXNlcjQ3MjU2OTM2 | User | false |
openclaw/openclaw | 3,909,642,906 | I_kwDOQb6kR87pCGqa | 10,957 | https://github.com/openclaw/openclaw/issues/10957 | https://api.github.com/repos/openclaw/openclaw/issues/10957 | 📱 Request Android Beta Access - Node for existing Gateway setup | ## GitHub Username
robbie1209
## Email
johnny.clawd@outlook.com
## Use Case
- Running a dedicated Motorola G15 as a 24/7 Clawdbot node (always-on, plugged in)
- Gateway already running on Mac mini (macOS, Sequoia)
- giffgaff SIM active in the phone
- Want camera, location, SMS, and canvas capabilities
- Currently run... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-07T07:05:45Z | 2026-05-05T01:18:19Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | robbie1209 | 255,968,837 | U_kgDOD0HGRQ | User | false |
openclaw/openclaw | 3,909,647,740 | I_kwDOQb6kR87pCH18 | 10,960 | https://github.com/openclaw/openclaw/issues/10960 | https://api.github.com/repos/openclaw/openclaw/issues/10960 | Feature request: Mid-stream message injection (soft steer) | ## Feature Request
**Problem:**
Currently, `steer` mode only injects messages at tool boundaries. If the agent is mid-generation or waiting on a long tool call, the user's message queues until the next boundary. This defeats the purpose of real-time steering.
**Desired behavior:**
True mid-stream injection — when a u... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-07T07:07:30Z | 2026-05-05T01:18:20Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | marcuswylde | 202,796,584 | U_kgDODBZuKA | User | false |
openclaw/openclaw | 3,909,904,855 | I_kwDOQb6kR87pDGnX | 11,040 | https://github.com/openclaw/openclaw/issues/11040 | https://api.github.com/repos/openclaw/openclaw/issues/11040 | Feature: First-class session/task chain tracking (parent/root/trace/depth) | ## Summary
Add first-class session/task chain tracking so OpenClaw can reconstruct subagent execution DAGs without external loggers.
## Problem
Today we can partially infer spawn relationships via `spawnedBy`, `sessions.list(spawnedBy=...)`, and subagent registry, but we cannot reliably reconstruct an end-to-end cha... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-07T09:46:40Z | 2026-05-05T01:18:22Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | alfadb | 3,182,249 | MDQ6VXNlcjMxODIyNDk= | User | false |
openclaw/openclaw | 3,909,961,650 | I_kwDOQb6kR87pDUey | 11,055 | https://github.com/openclaw/openclaw/issues/11055 | https://api.github.com/repos/openclaw/openclaw/issues/11055 | Feature: Visible warning before compaction (memoryFlush.visibleWarning) | ## Summary
Add an optional **visible warning** to the agent before auto-compaction triggers, giving it a chance to consciously save in-progress work.
## Current Behavior
The `memoryFlush` system runs a silent turn (`NO_REPLY`) before compaction. The agent receives a prompt to save memories, but:
- The agent doesn't ... | closed | not_planned | false | 3 | [
"enhancement"
] | [] | 2026-02-07T10:20:48Z | 2026-05-04T03:21:48Z | 2026-05-04T03:21:48Z | NONE | null | 20260504T060031Z | 2026-05-04T06:00:31Z | One-Up-Dev | 225,904,118 | U_kgDODXcF9g | User | false |
openclaw/openclaw | 3,909,965,667 | I_kwDOQb6kR87pDVdj | 11,057 | https://github.com/openclaw/openclaw/issues/11057 | https://api.github.com/repos/openclaw/openclaw/issues/11057 | Support for 1M context window with Claude Max subscription | ## Description
Claude Opus 4.6 was announced with a 1M token context window in beta for Claude Max subscribers. However, OpenClaw currently reports a 195k context limit when using Opus 4.6.
## Current Behavior
```
$ openclaw models list | grep opus-4-6
anthropic/claude-opus-4-6 text+image 195k n... | closed | not_planned | true | 6 | [
"enhancement"
] | [] | 2026-02-07T10:23:55Z | 2026-05-02T04:28:45Z | 2026-04-29T09:08:16Z | NONE | null | 20260502T060142Z | 2026-05-02T06:01:42Z | Valinainformatica | 213,354,025 | U_kgDODLeGKQ | User | false |
openclaw/openclaw | 3,910,255,945 | I_kwDOQb6kR87pEcVJ | 11,146 | https://github.com/openclaw/openclaw/issues/11146 | https://api.github.com/repos/openclaw/openclaw/issues/11146 | Dashboard UX Improvements |
1. Keyboard Shortcuts for Dashboard Navigation
Problem
Currently, navigating the OpenClaw dashboard requires constant mouse interaction. The "See New Messages" button must be clicked manually to jump to new messages, which interrupts workflow for keyboard-focused users.
Proposed Solution
Add keyboard shortcuts fo... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-07T13:11:41Z | 2026-05-05T01:18:23Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | Elholandes1980 | 188,880,826 | U_kgDOC0IXug | User | false |
openclaw/openclaw | 3,910,340,160 | I_kwDOQb6kR87pEw5A | 11,172 | https://github.com/openclaw/openclaw/issues/11172 | https://api.github.com/repos/openclaw/openclaw/issues/11172 | Feature Request: Set OPENCLAW_AGENT_ID environment variable in subagent sessions | ## Summary
When spawning a subagent via `sessions_spawn`, set an environment variable (e.g., `OPENCLAW_AGENT_ID` or `CLAWDBOT_AGENT_ID`) to the agentId so spawned sessions can identify themselves to external tools.
## Use Case
I have a git pre-push hook that blocks direct pushes from my main agent (NOVA) and require... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-07T13:59:48Z | 2026-05-05T01:18:24Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | NOVA-Openclaw | 258,138,572 | U_kgDOD2LhzA | User | false |
openclaw/openclaw | 3,910,801,356 | I_kwDOQb6kR87pGhfM | 11,301 | https://github.com/openclaw/openclaw/issues/11301 | https://api.github.com/repos/openclaw/openclaw/issues/11301 | Add /workspace symlink to container configuration | ## Problem
The `/workspace` symlink referenced throughout AGENTS.md and SOUL.md does not exist in agent containers. This causes friction every session as agents must use the full path `/home/node/.openclaw/workspace`.
## Solution
Add to Dockerfile or container entrypoint:
```bash
ln -sf /home/node/.openclaw/workspace ... | open | null | false | 2 | [
"enhancement"
] | [] | 2026-02-07T17:33:22Z | 2026-05-05T02:55:23Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | tukunangh | 134,852,420 | U_kgDOCAmvRA | User | false |
openclaw/openclaw | 3,910,866,468 | I_kwDOQb6kR87pGxYk | 11,326 | https://github.com/openclaw/openclaw/issues/11326 | https://api.github.com/repos/openclaw/openclaw/issues/11326 | [Feature]: Add GBK/GB2312 encoding support to web_fetch tool for Chinese websites | ### Problem Description
The `web_fetch` tool currently has issues with Chinese websites that use GBK/GB2312 encoding, resulting in garbled text when extracting content.
### Use Case
- Fetching financial data from Chinese sites (eastmoney.com, 10jqka.com.cn, etc.)
- Accessing Chinese news and information websites
- Gen... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-07T18:06:18Z | 2026-05-01T04:41:06Z | 2026-04-29T00:07:56Z | NONE | null | 20260501T113057Z | 2026-05-01T11:30:57Z | Prc2020xie | 260,108,815 | U_kgDOD4DyDw | User | false |
openclaw/openclaw | 3,911,063,940 | I_kwDOQb6kR87pHhmE | 11,414 | https://github.com/openclaw/openclaw/issues/11414 | https://api.github.com/repos/openclaw/openclaw/issues/11414 | Message plugin: support Telegram forum topic management (create/edit/close/reopen) | ## Feature Request
### Summary
The message plugin currently supports `send`, `edit`, `delete`, and `react` actions, but does not expose Telegram's forum topic management API.
### Requested Actions
Add support for the following Telegram Bot API methods via the message plugin:
- **`createForumTopic`** — create a new t... | open | null | false | 2 | [
"enhancement",
"dedupe:parent"
] | [] | 2026-02-07T19:45:52Z | 2026-05-05T02:00:18Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | hariseldonsradiant | 259,080,940 | U_kgDOD3FC7A | User | false |
openclaw/openclaw | 3,911,222,213 | I_kwDOQb6kR87pIIPF | 11,460 | https://github.com/openclaw/openclaw/issues/11460 | https://api.github.com/repos/openclaw/openclaw/issues/11460 | WhatsApp: Add support for querying message reactions | ## Summary
When using WhatsApp as a channel, there's currently no way to query reactions on messages.
## Current Behavior
- ✅ **Sending reactions works** via `message react` action
- ⚠️ **Receiving reaction notifications**: Config exists (`reactionNotifications: "own"`) but unclear if events surface to agent
- ❌ **Que... | open | null | false | 3 | [
"enhancement"
] | [] | 2026-02-07T20:57:55Z | 2026-05-05T01:18:26Z | null | NONE | null | 20260505T030055Z | 2026-05-05T03:00:55Z | ario | 952,509 | MDQ6VXNlcjk1MjUwOQ== | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.