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 | 3,911,016,917 | I_kwDOQb6kR87pHWHV | 11,393 | https://github.com/openclaw/openclaw/issues/11393 | https://api.github.com/repos/openclaw/openclaw/issues/11393 | Route heartbeats to isolated system session (avoid thread pollution) | ## Feature Request: Route Heartbeats to Isolated System Session
### Problem
When a heartbeat poll fires during an active conversation, the `HEARTBEAT_OK` response gets injected into the user's conversation thread, causing confusion.
Example:
```
User: Can you email me this plan?
Assistant: [sends email]
[heartbeat po... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-07T19:25:27Z | 2026-04-27T05:06:47Z | 2026-04-24T18:54:30Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | RalphPepperoni | 257,000,330 | U_kgDOD1GDig | User | false |
openclaw/openclaw | 3,911,035,803 | I_kwDOQb6kR87pHaub | 11,394 | https://github.com/openclaw/openclaw/issues/11394 | https://api.github.com/repos/openclaw/openclaw/issues/11394 | [Feature]: MCP Server tool-level allow/deny filtering | MCP Server Tool-Level Allow/Deny Filtering
Summary
Adds granular control over which MCP tools are available per server, allowing users to allowlist or denylist specific tools rather than all-or-nothing server access.
Motivation
Currently, enabling an MCP server exposes all its tools. For servers with many tools (or... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-07T19:32:44Z | 2026-04-27T05:06:46Z | 2026-04-24T18:54:34Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | schulthoff | 204,683,063 | U_kgDODDM3Nw | User | false |
openclaw/openclaw | 3,911,260,754 | I_kwDOQb6kR87pIRpS | 11,469 | https://github.com/openclaw/openclaw/issues/11469 | https://api.github.com/repos/openclaw/openclaw/issues/11469 | WhatsApp: Support separate contact appearance without self-chat mode | ## Feature Request: Separate Contact Appearance for WhatsApp
### Current Behavior
When using WhatsApp with OpenClaw in `selfChatMode`, the assistant appears as the user talking to themselves rather than as a distinct contact. This makes it less clear that you're interacting with an AI assistant.
### Desired Behavior
... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-07T21:15:06Z | 2026-04-27T05:06:46Z | 2026-04-24T18:54:37Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | jones-d | 49,987,634 | MDQ6VXNlcjQ5OTg3NjM0 | User | false |
openclaw/openclaw | 3,911,444,651 | I_kwDOQb6kR87pI-ir | 11,501 | https://github.com/openclaw/openclaw/issues/11501 | https://api.github.com/repos/openclaw/openclaw/issues/11501 | Feature Request: Customizable sender prefix for messaging channels | ## Feature Request: Customizable sender prefix for messaging channels
### Description
Currently, all messages sent through OpenClaw channels (WhatsApp, etc.) include an "[openclaw]" prefix at the beginning, followed by the assistant name. For example: "[openclaw] [Kenobi] - Your message here"
### Request
Add a config... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-07T22:36:05Z | 2026-04-27T05:06:45Z | 2026-04-24T18:54:41Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | obikenobimain-hue | 259,857,888 | U_kgDOD30d4A | User | false |
openclaw/openclaw | 3,911,454,739 | I_kwDOQb6kR87pJBAT | 11,504 | https://github.com/openclaw/openclaw/issues/11504 | https://api.github.com/repos/openclaw/openclaw/issues/11504 | [Feature Request] Add before_model_select plugin hook for dynamic model routing | ## Summary
Add a new plugin hook (`before_model_select`) that allows plugins to influence model selection before each agent turn. This enables intelligent model routing, cost optimization, and A/B testing without requiring core modifications.
## Use Case: Intelligent Model Routing
**Problem:** Different tasks require... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-07T22:40:30Z | 2026-04-27T05:06:44Z | 2026-04-24T18:54:45Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | bzanghi | 131,682,606 | U_kgDOB9lRLg | User | false |
openclaw/openclaw | 3,911,515,316 | I_kwDOQb6kR87pJPy0 | 11,521 | https://github.com/openclaw/openclaw/issues/11521 | https://api.github.com/repos/openclaw/openclaw/issues/11521 | Feature: Automatic Markdown to Slack mrkdwn conversion | ## Problem
When OpenClaw sends agent responses to Slack, standard Markdown formatting doesn't render correctly because Slack uses its own mrkdwn format.
### Examples
- Standard Markdown: `**bold**` → Slack shows: \*\*bold\*\*
- Slack mrkdwn: `*bold*` → Slack shows: **bold**
- Tables: Not supported in Slack at all
##... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-07T23:07:00Z | 2026-04-27T05:06:43Z | 2026-04-24T18:54:49Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | glen15 | 75,651,376 | MDQ6VXNlcjc1NjUxMzc2 | User | false |
openclaw/openclaw | 3,911,558,639 | I_kwDOQb6kR87pJaXv | 11,533 | https://github.com/openclaw/openclaw/issues/11533 | https://api.github.com/repos/openclaw/openclaw/issues/11533 | [Feature]: Support DeepInfra as an LLM provider in OpenClaw | ## Summary
OpenClaw currently does not support DeepInfra as an LLM provider. Attempts to use DeepInfra's OpenAI-compatible endpoints result in a "422 status code (no body)" error, while other providers (e.g. pollinations) operate as expected.
## Proposed solution
Add support for DeepInfra as an LLM provider in OpenC... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-07T23:34:26Z | 2026-04-27T05:06:42Z | 2026-04-24T18:54:52Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | n00b001 | 1,337,355 | MDQ6VXNlcjEzMzczNTU= | User | false |
openclaw/openclaw | 3,911,740,505 | I_kwDOQb6kR87pKGxZ | 11,596 | https://github.com/openclaw/openclaw/issues/11596 | https://api.github.com/repos/openclaw/openclaw/issues/11596 | [Feature Request] Add timezone support for log timestamps (UTC+8 for Beijing) | ## Problem
The gateway logs show timestamps in UTC, but users in Beijing use UTC+8 timezone. This makes log reading inconvenient.
**Example**:
01:46:17 [feishu] feishu[default]: WebSocket client started
Actual Beijing time: `09:46:17`
**Expected Behavior**
Logs should display local time (Beijing: UTC+8) instead of U... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-08T01:50:06Z | 2026-04-27T05:06:42Z | 2026-04-24T18:54:56Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | poe12poe | 26,784,105 | MDQ6VXNlcjI2Nzg0MTA1 | User | false |
openclaw/openclaw | 3,911,788,959 | I_kwDOQb6kR87pKSmf | 11,619 | https://github.com/openclaw/openclaw/issues/11619 | https://api.github.com/repos/openclaw/openclaw/issues/11619 | Periodic memory auto-save | Current `session-memory` hook only saves on `/new` command. Add option for periodic auto-save based on triggers.
**Problem:**
Memory is only saved when user explicitly runs `/new`, which means:
- Context is lost if session crashes
- Long sessions accumulate unsaved context
- Compaction events lose everything
**Propos... | closed | not_planned | true | 4 | [
"enhancement"
] | [] | 2026-02-08T02:33:12Z | 2026-04-27T05:06:41Z | 2026-04-24T18:55:00Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | murilo | 26,014 | MDQ6VXNlcjI2MDE0 | User | false |
openclaw/openclaw | 3,912,053,499 | I_kwDOQb6kR87pLTL7 | 11,695 | https://github.com/openclaw/openclaw/issues/11695 | https://api.github.com/repos/openclaw/openclaw/issues/11695 | [Feature]: Logseq Integration | ## Summary
I use Logseq as my primary knowledge management tool for organizing research, meeting
notes, and project planning (e.g., career coaching sessions with background research,
video transcription outputs, Q&A lists — all interlinke... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-08T05:41:21Z | 2026-04-27T05:06:40Z | 2026-04-24T18:55:03Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | veristagnum | 116,735,067 | U_kgDOBvU8Ww | User | false |
openclaw/openclaw | 3,912,329,165 | I_kwDOQb6kR87pMWfN | 11,776 | https://github.com/openclaw/openclaw/issues/11776 | https://api.github.com/repos/openclaw/openclaw/issues/11776 | [Feature]: Always-responsive control lane + watchdog for long tasks | ### Summary
I run OpenClaw on a remote Linux host (moltbot-host). When long tasks run, my WhatsApp agent often stops replying for a long time. From my perspective, the biggest pain is: I don't know if it's busy, stuck, or dead.
Today I ran:
- `openclaw status --all`
- `openclaw gateway status`
The report said:
- Ga... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-08T08:56:57Z | 2026-04-27T04:36:40Z | 2026-04-24T18:55:07Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Crnelia4227 | 11,078,936 | MDQ6VXNlcjExMDc4OTM2 | User | false |
openclaw/openclaw | 3,912,575,679 | I_kwDOQb6kR87pNSq_ | 11,826 | https://github.com/openclaw/openclaw/issues/11826 | https://api.github.com/repos/openclaw/openclaw/issues/11826 | Support payload.model override for per-job model selection | ### Problem
Cron jobs cannot specify which model to use. All jobs use the global `agents.defaults.model` chain, forcing operators to choose between:
- Expensive cloud models for ALL jobs (wasteful for simple tasks)
- Cheap local models for ALL jobs (insufficient for complex tasks)
### Evidence
When attempting to set... | closed | completed | true | 4 | [
"enhancement"
] | [] | 2026-02-08T11:33:39Z | 2026-04-27T05:06:39Z | 2026-04-24T18:55:11Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | similar-bespoke | 125,871,494 | U_kgDOB4Clhg | User | false |
openclaw/openclaw | 3,912,887,413 | I_kwDOQb6kR87pOex1 | 11,919 | https://github.com/openclaw/openclaw/issues/11919 | https://api.github.com/repos/openclaw/openclaw/issues/11919 | RFC: Composable Skills Architecture — skill dependencies, interfaces, and composition | ## Summary
I'd like to propose a backwards-compatible extension to OpenClaw's skill system that enables **skill composition**: the ability for skills to declare dependencies on other skills, expose typed interfaces, and compose into higher-order capabilities.
## Problem
Today, skills are flat and isolated. There's n... | closed | not_planned | true | 5 | [
"enhancement"
] | [] | 2026-02-08T14:55:10Z | 2026-04-27T05:06:38Z | 2026-04-24T18:55:15Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | vivekchand | 987,905 | MDQ6VXNlcjk4NzkwNQ== | User | false |
openclaw/openclaw | 3,912,961,323 | I_kwDOQb6kR87pOw0r | 11,940 | https://github.com/openclaw/openclaw/issues/11940 | https://api.github.com/repos/openclaw/openclaw/issues/11940 | [Feature] Visual Model Priority Management UI Plugin | ## Description
I have developed a visual model priority management UI for the OpenClaw Gateway Dashboard to address the challenge of managing 700+ models.
## Key Features
- **Drag-and-Drop Priority**: Easily reorder primary and fallback models.
- **Provider-First Browsing**: Organized by provider for better navigation... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-08T15:38:35Z | 2026-04-27T05:06:37Z | 2026-04-24T18:55:19Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | oneles | 62,010,096 | MDQ6VXNlcjYyMDEwMDk2 | User | false |
openclaw/openclaw | 3,912,994,345 | I_kwDOQb6kR87pO44p | 11,946 | https://github.com/openclaw/openclaw/issues/11946 | https://api.github.com/repos/openclaw/openclaw/issues/11946 | Post-Conversation Hook for Automated Knowledge Extraction | ## Summary
Agents effectively "forget" after a new conversation/session starts, so users have to manually write daily notes / memory files to preserve learnings.
It would be very helpful to have an automatic post-conversation hook that can extract key knowledge from the conversation and write it to a configured file/s... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-08T15:57:31Z | 2026-04-27T05:06:36Z | 2026-04-24T18:55:22Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | bzhz2023 | 213,139,146 | U_kgDODLQ-yg | User | false |
openclaw/openclaw | 3,913,430,421 | I_kwDOQb6kR87pQjWV | 12,061 | https://github.com/openclaw/openclaw/issues/12061 | https://api.github.com/repos/openclaw/openclaw/issues/12061 | [Feature]: Claude code headless mode as an "api provider" | ## Summary
A ton of people in the community are concerned about token costs with Anthropic, or if they use the plans, they're concerned about being banned (like what happened to opencode users).
## Proposed solution
Add claude code headless mode as an llm provider. Instead of reaching to the API - have the user i... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-08T19:58:42Z | 2026-04-27T05:06:35Z | 2026-04-24T18:55:26Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | haydonryan | 6,804,348 | MDQ6VXNlcjY4MDQzNDg= | User | false |
openclaw/openclaw | 3,913,666,693 | I_kwDOQb6kR87pRdCF | 12,129 | https://github.com/openclaw/openclaw/issues/12129 | https://api.github.com/repos/openclaw/openclaw/issues/12129 | Feature request: outbound text transform hook for channel plugins | ## Summary
It would be great to have a configurable outbound text transform hook/middleware for channel plugins, so users can customize message formatting per channel without patching source files.
## Use case
WhatsApp uses its own formatting syntax (e.g. `*bold*` instead of `**bold**`, no support for `[links](url)`... | closed | completed | true | 5 | [
"enhancement"
] | [] | 2026-02-08T22:03:38Z | 2026-04-27T05:06:34Z | 2026-04-24T18:55:29Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | kbecking | 195,651,854 | U_kgDOC6lpDg | User | false |
openclaw/openclaw | 3,913,820,320 | I_kwDOQb6kR87pSCig | 12,179 | https://github.com/openclaw/openclaw/issues/12179 | https://api.github.com/repos/openclaw/openclaw/issues/12179 | Feature Request: Auto memory search injection on user prompts | ## Summary
Add an optional auto-inject feature that automatically runs `memory_search` on every user prompt and injects relevant memories into context before the agent processes the message.
## Problem
Currently, the agent must manually decide when to call `memory_search`. This means:
- Relevant context might be mis... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-08T23:45:44Z | 2026-04-27T05:06:34Z | 2026-04-24T18:55:33Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | asz-hub | 224,570,580 | U_kgDODWKs1A | User | false |
openclaw/openclaw | 3,914,215,938 | I_kwDOQb6kR87pTjIC | 12,299 | https://github.com/openclaw/openclaw/issues/12299 | https://api.github.com/repos/openclaw/openclaw/issues/12299 | No programmatic access to cumulative token usage or cost data per session | ## Feature Request
**Version:** `2026.2.6-3`
### Problem
There is no API or CLI command to retrieve **cumulative** token usage or estimated cost per session. The existing data sources are misleading or incomplete:
#### `sessions.json` token fields are not cumulative
- `inputTokens` / `outputTokens` — these track **... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-09T03:08:38Z | 2026-04-27T05:06:33Z | 2026-04-24T18:55:37Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | kurthamm | 16,831,171 | MDQ6VXNlcjE2ODMxMTcx | User | false |
openclaw/openclaw | 3,914,652,277 | I_kwDOQb6kR87pVNp1 | 12,377 | https://github.com/openclaw/openclaw/issues/12377 | https://api.github.com/repos/openclaw/openclaw/issues/12377 | [Feature]:Provide a way to surface tool call details in channels | **Summary**
When OpenClaw's agent uses tools (e.g., exec, web_search, web_fetch, Read, etc.), the tool call inputs and outputs are only visible in the WebChat interface. On messaging channels like Telegram and Discord, users can only see the agent's final text reply — the intermediate tool invocations and their results... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-09T06:12:15Z | 2026-04-27T05:06:32Z | 2026-04-24T18:55:41Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ifdog | 1,519,790 | MDQ6VXNlcjE1MTk3OTA= | User | false |
openclaw/openclaw | 3,914,702,148 | I_kwDOQb6kR87pVZ1E | 12,385 | https://github.com/openclaw/openclaw/issues/12385 | https://api.github.com/repos/openclaw/openclaw/issues/12385 | [Feature]: Security - Context-Based Runtime Security Policy (Shield.md) | ## Summary
OpenClaw agents can install/execute skills, call tools, talk to MCP servers, make outbound requests, and read secrets. Today, the “should we proceed?” logic is inconsistent across features and easy to bypass or forget, which increases the risk of accidental data exfiltration and unsafe execution.
Proposals... | closed | not_planned | true | 17 | [
"enhancement"
] | [] | 2026-02-09T06:27:02Z | 2026-04-27T05:06:31Z | 2026-04-24T18:55:45Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | fr0gger | 6,546,250 | MDQ6VXNlcjY1NDYyNTA= | User | false |
openclaw/openclaw | 3,914,800,396 | I_kwDOQb6kR87pVx0M | 12,405 | https://github.com/openclaw/openclaw/issues/12405 | https://api.github.com/repos/openclaw/openclaw/issues/12405 | Feature: Pluggable sandbox backends & per-agent exec routing | # Feature: Pluggable sandbox backends & per-agent exec routing
## Problem
Currently, OpenClaw sandboxing is Docker-only (`sandbox.docker.image`), and `tools.exec.host` is a **global** setting — all agents share the same exec target. This creates two gaps:
### 1. No pluggable sandbox backends
The only sandbox option ... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-09T06:59:22Z | 2026-04-27T05:06:30Z | 2026-04-24T18:55:49Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | anshumanbh | 9,063,340 | MDQ6VXNlcjkwNjMzNDA= | User | false |
openclaw/openclaw | 3,914,972,402 | I_kwDOQb6kR87pWbzy | 12,436 | https://github.com/openclaw/openclaw/issues/12436 | https://api.github.com/repos/openclaw/openclaw/issues/12436 | docs: Document undocumented hooks (before_agent_start, message_received, etc.) | ## Background
While trying to implement automatic Slack thread context injection for our AI agents, we consulted the hooks documentation (`docs/hooks.md`). The doc lists `message:received` under "Future Events", which led us to believe this functionality was not available.
However, after reverse-engineering the imple... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-09T07:50:43Z | 2026-04-27T05:06:29Z | 2026-04-24T18:55:53Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ksato8710 | 50,102,309 | MDQ6VXNlcjUwMTAyMzA5 | User | false |
openclaw/openclaw | 3,915,274,799 | I_kwDOQb6kR87pXlov | 12,478 | https://github.com/openclaw/openclaw/issues/12478 | https://api.github.com/repos/openclaw/openclaw/issues/12478 | Feature: configuration assistant skill to help build openclaw config | ## Summary
有时候 让openclaw 自己进行配置，总是会配置错误，找不准正确的配置格式或配置项，能不能给他设计一个配置用的 skill，毕竟配置项太多，我作为一个使用者，也不是很熟悉，很多情况下还是需要 openclaw 自己进行配置的
## Proposed solution
希望能有一个官方的配置 skill，这样就可以减少 openclaw 配置错误的现象出现
## Additional context
谢谢
| closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-09T09:07:20Z | 2026-04-27T05:06:29Z | 2026-04-24T18:55:56Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | rheros | 23,209,622 | MDQ6VXNlcjIzMjA5NjIy | User | false |
openclaw/openclaw | 3,915,515,902 | I_kwDOQb6kR87pYgf- | 12,509 | https://github.com/openclaw/openclaw/issues/12509 | https://api.github.com/repos/openclaw/openclaw/issues/12509 | [Feature Request] Plugin Command Registration Conflict Detection | **Type**: Feature Request
**Category**: feature-request
**Severity**: High
## Problem Statement
Plugins loaded later in the registration sequence can override commands registered by earlier plugins. This enables command hijacking where a malicious plugin silently replaces legitimate functionality.
## Current State
... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-09T10:00:01Z | 2026-04-27T05:06:28Z | 2026-04-24T18:56:00Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | fr33d3m0n | 127,989,854 | U_kgDOB6D4Xg | User | false |
openclaw/openclaw | 3,915,517,850 | I_kwDOQb6kR87pYg-a | 12,518 | https://github.com/openclaw/openclaw/issues/12518 | https://api.github.com/repos/openclaw/openclaw/issues/12518 | [Security] Nostr/Zalo Channel: Full Credential Exposure via Shared Environment | **Component**: Extensions (TypeScript Plugins)
**Category**: security
**Priority**: Critical
## Summary
Credentials are exposed through multiple vectors: (1) `process.env` contains ALL plugin credentials (API keys, tokens, secrets), (2) ZCA subprocess receives full `process.env` via `spawn({env: {...process.env}})`, ... | closed | not_planned | true | 1 | [
"bug",
"security"
] | [] | 2026-02-09T10:00:28Z | 2026-04-27T05:06:27Z | 2026-04-24T18:56:04Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | fr33d3m0n | 127,989,854 | U_kgDOB6D4Xg | User | false |
openclaw/openclaw | 3,915,522,936 | I_kwDOQb6kR87pYiN4 | 12,540 | https://github.com/openclaw/openclaw/issues/12540 | https://api.github.com/repos/openclaw/openclaw/issues/12540 | [Security] Bundled Skills: Unauthenticated API Access Enabling Quota Exhaustion | **Component**: Bundled Skills
**Category**: security
**Priority**: High
## Summary
The Places API server binds to 0.0.0.0 on port 8000 without any authentication requirement. All four API endpoints accept unauthenticated requests. An attacker can send concurrent requests to exhaust the Google Places API quota, causin... | closed | completed | true | 1 | [
"bug",
"security"
] | [] | 2026-02-09T10:01:39Z | 2026-04-27T05:06:26Z | 2026-04-24T18:56:11Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | fr33d3m0n | 127,989,854 | U_kgDOB6D4Xg | User | false |
openclaw/openclaw | 3,915,524,200 | I_kwDOQb6kR87pYiho | 12,545 | https://github.com/openclaw/openclaw/issues/12545 | https://api.github.com/repos/openclaw/openclaw/issues/12545 | [Security] Nostr Channel: Private Key Stored in Plaintext | **Component**: Extensions (TypeScript Plugins)
**Category**: security
**Priority**: High
## Summary
The Nostr private key is stored as plaintext in the state file. While file permissions are set to 0o600, no encryption is applied. The private key is accessible to any code running in the shared process.
## Risk Asses... | closed | completed | true | 2 | [
"bug",
"security"
] | [] | 2026-02-09T10:01:56Z | 2026-04-27T05:06:24Z | 2026-04-24T18:56:19Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | fr33d3m0n | 127,989,854 | U_kgDOB6D4Xg | User | false |
openclaw/openclaw | 3,915,528,182 | I_kwDOQb6kR87pYjf2 | 12,563 | https://github.com/openclaw/openclaw/issues/12563 | https://api.github.com/repos/openclaw/openclaw/issues/12563 | [Security] Discord/Slack/Telegram Skills: Messaging Operations Without Audit | **Component**: Bundled Skills
**Category**: security
**Priority**: Medium
## Summary
Messaging platform operations (Discord, Slack, Telegram) have zero audit logging. Actions taken by the AI agent on behalf of users (including administrative actions) leave no record for accountability or forensic analysis.
## Risk A... | closed | not_planned | true | 1 | [
"bug",
"security"
] | [] | 2026-02-09T10:02:53Z | 2026-04-27T05:06:23Z | 2026-04-24T18:56:22Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | fr33d3m0n | 127,989,854 | U_kgDOB6D4Xg | User | false |
openclaw/openclaw | 3,915,528,625 | I_kwDOQb6kR87pYjmx | 12,565 | https://github.com/openclaw/openclaw/issues/12565 | https://api.github.com/repos/openclaw/openclaw/issues/12565 | [Security] Agent Runtime: Unrestricted Tool Execution Leading to Privilege Escalation | **Component**: Extensions (TypeScript Plugins)
**Category**: security
**Priority**: Medium
## Summary
The agent tool execution system has zero controls: no tool allowlist/denylist, no permission model per tool, no rate limiting on tool calls, and no parameter validation. Any message that reaches the agent can trigger... | closed | completed | true | 5 | [
"bug",
"security"
] | [] | 2026-02-09T10:02:59Z | 2026-04-27T05:06:23Z | 2026-04-24T18:56:26Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | fr33d3m0n | 127,989,854 | U_kgDOB6D4Xg | User | false |
openclaw/openclaw | 3,915,686,780 | I_kwDOQb6kR87pZKN8 | 12,587 | https://github.com/openclaw/openclaw/issues/12587 | https://api.github.com/repos/openclaw/openclaw/issues/12587 | [Feature]: Add openclaw icon to dashboard icons | ## Summary
Add openclaw icons to dashboard icons repository
## Proposed solution
This would give users who run dashboards like Heimdall and Homepage the ability to auto get the current icon.
## Alternatives considered
None
## Additional context
https://dashboardicons.com/icons
I see there is actually an icon, b... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-09T10:36:01Z | 2026-04-27T05:06:22Z | 2026-04-24T18:56:29Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | officiallymarky | 33,408,375 | MDQ6VXNlcjMzNDA4Mzc1 | User | false |
openclaw/openclaw | 3,915,874,844 | I_kwDOQb6kR87pZ4Ic | 12,611 | https://github.com/openclaw/openclaw/issues/12611 | https://api.github.com/repos/openclaw/openclaw/issues/12611 | [Feature]: Integrate tea2adt for Air-Gap-Safe Chat Interaction with OpenClaw | ## Problem
Running OpenClaw on a **dedicated offline PC** (air-gapped) is currently safe, but prevents **remote interaction**.
Typical workarounds - containers or VMs - introduce too many vulnerabilities for a highly capable AI, including kernel escapes, misconfigurations, and relay compromise. Current OpenClaw workf... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-09T11:21:44Z | 2026-04-27T05:06:21Z | 2026-04-24T18:56:33Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ClarkFieseln | 54,716,944 | MDQ6VXNlcjU0NzE2OTQ0 | User | false |
openclaw/openclaw | 3,916,298,290 | I_kwDOQb6kR87pbfgy | 12,660 | https://github.com/openclaw/openclaw/issues/12660 | https://api.github.com/repos/openclaw/openclaw/issues/12660 | Feature Request: Context Provider Plugin Slot (plugins.slots.context) | ## Summary
Add a new plugin slot (`plugins.slots.context`) that allows a plugin to **replace OpenClaw's default context assembly** — specifically, how conversation history is selected and included in each model call.
## Problem
Today, OpenClaw builds context by appending every message to the conversation transcript ... | closed | completed | true | 7 | [
"enhancement"
] | [] | 2026-02-09T13:07:11Z | 2026-04-27T05:06:20Z | 2026-04-24T18:56:37Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | rezabaram | 368,379 | MDQ6VXNlcjM2ODM3OQ== | User | false |
openclaw/openclaw | 3,917,051,366 | I_kwDOQb6kR87peXXm | 12,739 | https://github.com/openclaw/openclaw/issues/12739 | https://api.github.com/repos/openclaw/openclaw/issues/12739 | [Feature]: /status should use per-model contextWindow instead of global contextTokens | When switching models via /model, the /status command continues to show the global `agents.defaults.contextTokens` value (e.g., 1,000,000 for Opus) even when the active model has a different `contextWindow` defined in `models.providers`.
For example, switching to a local MLX model with `contextWindow: 128000` still sh... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-09T15:49:41Z | 2026-04-27T05:06:20Z | 2026-04-24T18:56:41Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | blortski | 117,885,527 | U_kgDOBwbKVw | User | false |
openclaw/openclaw | 3,917,482,249 | I_kwDOQb6kR87pgAkJ | 12,797 | https://github.com/openclaw/openclaw/issues/12797 | https://api.github.com/repos/openclaw/openclaw/issues/12797 | Feature: Debounce/Batch incoming Telegram messages (handle forwarded posts + comments) | When a user forwards a post to the bot in Telegram and adds a comment (caption) in the same action, Telegram API often sends this as two separate updates in rapid succession (e.g. within 100-300ms).
Current behavior:
OpenClaw treats them as two separate turns.
Turn 1: Processes the comment (e.g. 'What is this?').
Turn... | closed | completed | true | 5 | [
"enhancement"
] | [] | 2026-02-09T17:56:38Z | 2026-04-27T05:06:19Z | 2026-04-24T18:56:44Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Vovch0 | 154,655,665 | U_kgDOCTfbsQ | User | false |
openclaw/openclaw | 3,917,634,859 | I_kwDOQb6kR87pgl0r | 12,811 | https://github.com/openclaw/openclaw/issues/12811 | https://api.github.com/repos/openclaw/openclaw/issues/12811 | [Feature]: ctx.accountId undefined in plugin command handlers | Hi,
I found a bug in OpenClaw where ctx.accountId is always undefined when a plugin command
handler is called. This prevents plugi... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-09T18:31:33Z | 2026-04-27T05:06:18Z | 2026-04-24T18:58:09Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | habib256 | 1,260,156 | MDQ6VXNlcjEyNjAxNTY= | User | false |
openclaw/openclaw | 3,917,856,079 | I_kwDOQb6kR87phb1P | 12,838 | https://github.com/openclaw/openclaw/issues/12838 | https://api.github.com/repos/openclaw/openclaw/issues/12838 | Feature Request: Auto-resize large images before LLM submission | ## Problem
When users post images >3.5MB in Discord (or other channels), the base64 encoding inflates them beyond Anthropic's 5MB per-image limit, causing `LLM request rejected` errors visible to users.
## Current Behavior
- Gateway includes raw images from chat history in LLM context
- Base64 encoding adds ~33% overh... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-09T19:51:23Z | 2026-04-27T05:06:18Z | 2026-04-24T18:58:12Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Knytro23 | 257,926,341 | U_kgDOD1-kxQ | User | false |
openclaw/openclaw | 3,918,041,886 | I_kwDOQb6kR87piJMe | 12,867 | https://github.com/openclaw/openclaw/issues/12867 | https://api.github.com/repos/openclaw/openclaw/issues/12867 | Feature request: message:received hook for pre-processing incoming messages | ## Feature Request
Add a `message:received` hook event that fires before the agent processes an incoming message. This would enable pre-processing workflows like:
- Fetching thread context for Slack/Discord before the agent responds
- Enriching messages with external data
- Filtering or transforming messages before a... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-09T20:38:55Z | 2026-04-27T05:06:17Z | 2026-04-24T18:58:15Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | visshehi | 162,566,191 | U_kgDOCbCQLw | User | false |
openclaw/openclaw | 3,918,140,989 | I_kwDOQb6kR87pihY9 | 12,880 | https://github.com/openclaw/openclaw/issues/12880 | https://api.github.com/repos/openclaw/openclaw/issues/12880 | memory_get: support workspace files beyond MEMORY.md (USER.md, SOUL.md, etc.) | ## Feature Request
memory_get via /tools/invoke only works for MEMORY.md and memory/*.md. Other workspace files like USER.md, SOUL.md, AGENTS.md return `{disabled: true, error: "path required"}`.
The `read` tool would work but is (correctly) blocked from the HTTP API.
### Impact
The OpenClaw Hub uses `/tools/invoke`... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-09T21:01:47Z | 2026-04-27T05:06:16Z | 2026-04-24T18:58:18Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Alfridus1 | 86,924,928 | MDQ6VXNlcjg2OTI0OTI4 | User | false |
openclaw/openclaw | 3,919,060,694 | I_kwDOQb6kR87pmB7W | 13,024 | https://github.com/openclaw/openclaw/issues/13024 | https://api.github.com/repos/openclaw/openclaw/issues/13024 | Feature Request: Cron Job Chaining — trigger job on completion of another job | ## Summary
Add the ability to trigger a cron job when another cron job completes successfully. This enables multi-agent pipelines where agents hand off work sequentially without polling delays.
## Use Case: Multi-Agent Dev Pipeline
I run an autonomous dev pipeline where multiple agents on separate machines build, te... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T02:12:16Z | 2026-04-27T05:06:15Z | 2026-04-24T18:58:26Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | dropoutclawdbot | 260,558,269 | U_kgDOD4fNvQ | User | false |
openclaw/openclaw | 3,919,154,342 | I_kwDOQb6kR87pmYym | 13,059 | https://github.com/openclaw/openclaw/issues/13059 | https://api.github.com/repos/openclaw/openclaw/issues/13059 | [Feature] Publish Browser Relay extension to Chrome Web Store | ## Problem
The Browser Relay extension currently requires manual installation as an unpacked developer mode extension. This causes several UX issues:
1. **Extension disappears on Chrome restart** — Chrome periodically disables developer mode extensions
2. **Breaks on Chrome updates** — Unpacked extensions often get r... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-10T02:48:00Z | 2026-04-27T05:06:14Z | 2026-04-24T18:58:29Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Koul91 | 107,915,825 | U_kgDOBm6qMQ | User | false |
openclaw/openclaw | 3,919,422,687 | I_kwDOQb6kR87pnaTf | 13,133 | https://github.com/openclaw/openclaw/issues/13133 | https://api.github.com/repos/openclaw/openclaw/issues/13133 | Enhancement: Short-circuit --version in bootstrap to avoid full app initialization | ## Current Behavior
`openclaw --version` takes ~5 seconds on a single-core VPS (DO-Premium-AMD, 2GHz, 1 vCPU) to print a hardcoded version string. The flag goes through the full startup pipeline:
1. `openclaw.mjs` bootstrap
2. `entry.ts` respawns the process to add `--disable-warning=ExperimentalWarning` (~1.6s)
3. `... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T04:45:54Z | 2026-04-27T05:06:14Z | 2026-04-24T18:58:32Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | rocketman-code | 39,049,432 | MDQ6VXNlcjM5MDQ5NDMy | User | false |
openclaw/openclaw | 3,919,704,328 | I_kwDOQb6kR87pofEI | 13,206 | https://github.com/openclaw/openclaw/issues/13206 | https://api.github.com/repos/openclaw/openclaw/issues/13206 | [Feature Request] Support per-agent identity (name and avatar) | ## Problem
Currently all agents share the same name and avatar in ui.assistant. We want per-agent identity configuration.
## Expected Format
```json
{
"ui": {
"assistant": { "name": "Seal", "avatar": "🦭" },
"agents": {
"coding-agent": { "name": "Fox", "avatar": "🦊" },
"learning-agent": { "nam... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-10T06:30:56Z | 2026-04-27T05:06:13Z | 2026-04-24T18:58:35Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | luxi233 | 80,712,966 | MDQ6VXNlcjgwNzEyOTY2 | User | false |
openclaw/openclaw | 3,919,843,987 | I_kwDOQb6kR87ppBKT | 13,238 | https://github.com/openclaw/openclaw/issues/13238 | https://api.github.com/repos/openclaw/openclaw/issues/13238 | Proposal: Define a release quality gate and canary policy for stable tags | ## Summary
OpenClaw is moving fast, but recent releases show repeated regressions across channels/providers. I think we need a **release quality gate + canary policy** before cutting stable tags.
## Why this matters
- Many high-severity bugs are being reported within hours of release.
- Regressions often span core flo... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-10T07:15:26Z | 2026-04-27T05:06:12Z | 2026-04-24T18:58:38Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | 0xRaini | 190,923,101 | U_kgDOC2FBXQ | User | false |
openclaw/openclaw | 3,919,862,507 | I_kwDOQb6kR87ppFrr | 13,245 | https://github.com/openclaw/openclaw/issues/13245 | https://api.github.com/repos/openclaw/openclaw/issues/13245 | Feature Request: OpenAI Realtime API conversation mode for voice calls | ## Summary
The voice-call plugin currently uses OpenAI Realtime API only for **transcription** (speech-to-text). This results in ~5-10 second latency because the response generation still goes through the embedded agent.
## Proposed Enhancement
Add support for OpenAI Realtime API **conversation mode** (`type: realtime... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-10T07:21:21Z | 2026-04-27T05:06:11Z | 2026-04-24T18:58:41Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ilaydabdogan | 34,490,857 | MDQ6VXNlcjM0NDkwODU3 | User | false |
openclaw/openclaw | 3,921,025,363 | I_kwDOQb6kR87pthlT | 13,358 | https://github.com/openclaw/openclaw/issues/13358 | https://api.github.com/repos/openclaw/openclaw/issues/13358 | 🔥🔥🔥OpenClaw高级使用技巧视频分享 | https://youtu.be/iJEfIc1mrsc
https://www.bilibili.com/video/BV1vYFQzQE4P/
| closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-10T11:51:48Z | 2026-04-27T05:06:10Z | 2026-04-24T18:58:47Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | win4r | 42,172,631 | MDQ6VXNlcjQyMTcyNjMx | User | false |
openclaw/openclaw | 3,921,259,145 | I_kwDOQb6kR87puaqJ | 13,380 | https://github.com/openclaw/openclaw/issues/13380 | https://api.github.com/repos/openclaw/openclaw/issues/13380 | [Feature]: [RFC / Partnership] Building Water AI: The "Excel" of the Post-Scarcity Workforce (Open Source AGI Orchestrator) | **Author:** John Victor, Stellarium Foundation
---
### 🌊 The Vision: The Operating System for Work
We are looking for co-architects to build **Water AI**—an open-source, local-first, agentic platform that automates digital labor "out of the box."
While the current AI landscape is fragmented into chat interfaces an... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-10T12:43:15Z | 2026-04-27T05:06:10Z | 2026-04-24T18:58:50Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | StellariumFoundation | 180,241,666 | U_kgDOCr5FAg | User | false |
openclaw/openclaw | 3,921,373,541 | I_kwDOQb6kR87pu2ll | 13,392 | https://github.com/openclaw/openclaw/issues/13392 | https://api.github.com/repos/openclaw/openclaw/issues/13392 | Feature Request: session:spawn and session:complete hook events | ## Summary
Add `session:spawn` and `session:complete` hook events to the internal hooks system, enabling automations when isolated sub-agent sessions are created and when they finish.
## Motivation
In multi-bot Telegram setups (and similar multi-identity architectures), a common pattern is:
1. Main agent receives a... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-10T13:08:03Z | 2026-04-27T05:06:09Z | 2026-04-24T18:58:53Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | kyya | 13,448,248 | MDQ6VXNlcjEzNDQ4MjQ4 | User | false |
openclaw/openclaw | 3,921,620,025 | I_kwDOQb6kR87pvyw5 | 13,419 | https://github.com/openclaw/openclaw/issues/13419 | https://api.github.com/repos/openclaw/openclaw/issues/13419 | [Feature]: Option to disable implicit mention when bot is thread parent | ## Summary
When the bot is the parent of a Slack thread (i.e. the bot posted the first message), it responds to every reply in that thread — even without an explicit @mention. This is because `implicitMention` is set to `true` whenever `message.parent_user_id === ctx.botUserId`.
For use cases where `requireMention` i... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-10T13:58:17Z | 2026-04-27T05:06:08Z | 2026-04-24T18:58:56Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | carrotRakko | 24,588,751 | MDQ6VXNlcjI0NTg4NzUx | User | false |
openclaw/openclaw | 3,921,650,557 | I_kwDOQb6kR87pv6N9 | 13,423 | https://github.com/openclaw/openclaw/issues/13423 | https://api.github.com/repos/openclaw/openclaw/issues/13423 | [Bug]: [CRITICAL SECURITY] Agent Routing System Complete Failure - Communication Misdirection | # [CRITICAL SECURITY] Agent Routing System Complete Failure - Communication Misdirection
## Summary
**CRITICAL P0 Security Issue:** OpenClaw Gateway routing system completely broken. Telegram bot messages are routed to wrong agents, causing users to unknowingly communicate with different agents than intended. This cre... | closed | completed | true | 1 | [
"bug",
"security"
] | [] | 2026-02-10T14:03:51Z | 2026-04-27T05:06:07Z | 2026-04-24T18:58:59Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | recurator | 186,123 | MDQ6VXNlcjE4NjEyMw== | User | false |
openclaw/openclaw | 3,921,671,307 | I_kwDOQb6kR87pv_SL | 13,426 | https://github.com/openclaw/openclaw/issues/13426 | https://api.github.com/repos/openclaw/openclaw/issues/13426 | Feature Request: Discord inline button support (like Telegram) | ## Summary
OpenClaw supports inline buttons for Telegram via `buttons=[[{text, callback_data}]]`, where button clicks route back as user messages to the agent. This feature is missing for Discord.
## Current Behavior
- `--buttons` parameter on `message send` is explicitly Telegram-only
- Discord messages sent with `... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T14:07:29Z | 2026-04-27T05:06:07Z | 2026-04-24T18:59:02Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | batumilove | 189,313,529 | U_kgDOC0ix-Q | User | false |
openclaw/openclaw | 3,922,294,526 | I_kwDOQb6kR87pyXb- | 13,499 | https://github.com/openclaw/openclaw/issues/13499 | https://api.github.com/repos/openclaw/openclaw/issues/13499 | Feature Request: Auto-spawn at context threshold | ## Summary
Add configuration option to automatically spawn a fresh session when context window exceeds a threshold (e.g., 50%).
## Problem
Research shows model performance degrades as context fills up - even at 50% capacity, agents can become confused and lose coherence. Currently, OpenClaw has:
- `compaction` - sum... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-10T16:04:45Z | 2026-04-27T05:06:06Z | 2026-04-24T18:59:06Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | dave-melillo | 20,580,540 | MDQ6VXNlcjIwNTgwNTQw | User | false |
openclaw/openclaw | 3,922,436,813 | I_kwDOQb6kR87py6LN | 13,513 | https://github.com/openclaw/openclaw/issues/13513 | https://api.github.com/repos/openclaw/openclaw/issues/13513 | I am not afraid of OpenClaw, nor its bugs, nor tokens, nor hyps, nor the lack of electricity. | Introduction:
Let me introduce you to Pi, my Openclaw AI. It lives autonomously on an old third-generation HP 6506b i7 laptop with 16 GB of RAM and an SSD. It has sudo access, and this is our story.
First day: headaches. Second day: headaches. Third day: total stability and resilience, even in the face of hypos failure... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-10T16:34:32Z | 2026-04-27T05:06:05Z | 2026-04-24T18:59:09Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | mayga966-droid | 260,700,539 | U_kgDOD4n5ew | User | false |
openclaw/openclaw | 3,922,596,341 | I_kwDOQb6kR87pzhH1 | 13,533 | https://github.com/openclaw/openclaw/issues/13533 | https://api.github.com/repos/openclaw/openclaw/issues/13533 | Agent personality repos as .claude workspaces in .openclaw | ## Problem
Each agent (Mandy, Devy, Mark) needs:
1. **An OpenClaw identity** (personality, theme, routing) — lives in `openclaw.json`
2. **A .claude brain** (SOPs, skills, MCP config, hooks) — lives in its own git repo
Currently there's no mechanism to clone an agent's .claude repo into the `.openclaw/` runtime so Op... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T17:09:48Z | 2026-04-27T05:06:05Z | 2026-04-24T18:59:11Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ldraney | 107,153,866 | U_kgDOBmMJyg | User | false |
openclaw/openclaw | 3,922,799,205 | I_kwDOQb6kR87p0Spl | 13,545 | https://github.com/openclaw/openclaw/issues/13545 | https://api.github.com/repos/openclaw/openclaw/issues/13545 | [Feature implement request]: Signal group reply with mentioning (Patch included) | ## Summary
Request Feature official added for Signal group reply with mentioning.
## Proposed solution
Only reply when mentioning in group (Signal).
## Additional context
Attached guide and code thats already works for me from auto patch by AI.
[signal_patches_2026.02.03-1.md](https://github.com/user-attachments/... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T17:49:58Z | 2026-04-27T05:06:04Z | 2026-04-24T18:59:14Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | tsangkingyiu | 157,472,574 | U_kgDOCWLXPg | User | false |
openclaw/openclaw | 3,922,851,795 | I_kwDOQb6kR87p0ffT | 13,553 | https://github.com/openclaw/openclaw/issues/13553 | https://api.github.com/repos/openclaw/openclaw/issues/13553 | [Feature]: Save whatsapp/iMessage attachments to ~/clawd/temp for upload | ## Summary
Describe the problem you are trying to solve or the opportunity you see.
## Proposed solution
What would you like Clawdbot to do?
## Alternatives considered
Any other approaches you have considered?
## Additional context
Links, screenshots, or related issues. | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T18:02:47Z | 2026-04-27T05:06:03Z | 2026-04-24T18:59:17Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | stev00z | 29,033,507 | MDQ6VXNlcjI5MDMzNTA3 | User | false |
openclaw/openclaw | 3,922,894,927 | I_kwDOQb6kR87p0qBP | 13,562 | https://github.com/openclaw/openclaw/issues/13562 | https://api.github.com/repos/openclaw/openclaw/issues/13562 | [Feature]: Ollama Memory Search Provider | Currently ollama does not work as a provider for memorySearch
```
"memorySearch": {
"enabled": true,
"provider": "ollama",
"model": "qwen3-embedding:0.6b",
"sources": [
"memory"
]
}
```
Invalid config at /Users/xxx/.openclaw/openclaw.json:\n- agents.defa... | closed | completed | true | 4 | [
"enhancement"
] | [] | 2026-02-10T18:13:31Z | 2026-04-27T05:06:02Z | 2026-04-24T18:59:20Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | blastronaut | 3,728,529 | MDQ6VXNlcjM3Mjg1Mjk= | User | false |
openclaw/openclaw | 3,923,026,620 | I_kwDOQb6kR87p1KK8 | 13,587 | https://github.com/openclaw/openclaw/issues/13587 | https://api.github.com/repos/openclaw/openclaw/issues/13587 | Feature Request: Native Cross-Session Prompt Logging | ## Summary
Add built-in prompt logging at the gateway level to automatically capture all user inputs across sessions and Control UI clients.
## Current State
Manual implementation: agents manually write prompts to `memory/user-prompts-YYYY-MM-DD.md` files. This works but requires agent awareness and isn't available to... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T18:47:49Z | 2026-04-27T05:06:02Z | 2026-04-24T18:59:23Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Annaxiebot | 259,232,558 | U_kgDOD3OTLg | User | false |
openclaw/openclaw | 3,923,053,142 | I_kwDOQb6kR87p1QpW | 13,598 | https://github.com/openclaw/openclaw/issues/13598 | https://api.github.com/repos/openclaw/openclaw/issues/13598 | Add cron troubleshooting playbook to docs | ## Overview
Request a comprehensive troubleshooting guide for OpenClaw cron issues. Reference: issue #13582.
## Requested Coverage
### Verification Commands
- How to verify jobs are scheduled: `openclaw cron list`
- How to check execution history: `openclaw cron runs`
- How to view job details and output
### Troubl... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T18:54:36Z | 2026-04-27T05:06:01Z | 2026-04-24T18:59:26Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,053,175 | I_kwDOQb6kR87p1Qp3 | 13,599 | https://github.com/openclaw/openclaw/issues/13599 | https://api.github.com/repos/openclaw/openclaw/issues/13599 | Add detailed node pairing setup guide with examples | ## Overview
Request step-by-step node pairing documentation with real-world examples. Current documentation is sparse on this topic, causing setup confusion.
## Requested Coverage
### Conceptual Foundation
- What nodes are and their purpose
- When to use node pairing
- Architecture overview (gateway ↔ node relations... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T18:54:37Z | 2026-04-27T05:06:00Z | 2026-04-24T18:59:29Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,066,807 | I_kwDOQb6kR87p1T-3 | 13,608 | https://github.com/openclaw/openclaw/issues/13608 | https://api.github.com/repos/openclaw/openclaw/issues/13608 | Add metrics and observability support (Prometheus/OpenTelemetry) | ## Problem
OpenClaw currently lacks standardized metrics export, making it difficult to:
- Monitor production deployments
- Set up alerts for failures
- Track performance over time
- Debug issues with historical data
- Integrate with existing monitoring stacks (Grafana, Datadog, etc.)
## Proposed Solution
Add nativ... | closed | completed | true | 2 | [
"enhancement"
] | [
"vincentkoc"
] | 2026-02-10T18:58:09Z | 2026-04-27T05:05:59Z | 2026-04-24T18:59:32Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,066,969 | I_kwDOQb6kR87p1UBZ | 13,609 | https://github.com/openclaw/openclaw/issues/13609 | https://api.github.com/repos/openclaw/openclaw/issues/13609 | Add retry logic for failed cron jobs | ## Problem
When a cron job fails (network blip, transient API error, temporary service unavailability), it simply moves on to the next scheduled run without retry. This means transient failures can cause missed executions.
## Proposed Solution
Add configurable retry logic for cron jobs:
```json
{
"schedule": { ..... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-10T18:58:11Z | 2026-04-27T05:05:59Z | 2026-04-24T18:59:35Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | trevorgordon981 | 193,960,977 | U_kgDOC4-cEQ | User | false |
openclaw/openclaw | 3,923,393,806 | I_kwDOQb6kR87p2j0O | 13,656 | https://github.com/openclaw/openclaw/issues/13656 | https://api.github.com/repos/openclaw/openclaw/issues/13656 | [Feature]: add duckduck go or google search api | # Feature Request: Additional Web Search Providers
## Summary
Add support for additional search providers beyond Brave Search, including **DuckDuckGo**, **Google Custom Search API**, and **SearXNG** (self-hosted meta-search).
## Motivation
The current `web_search` tool only supports Brave, Perplexity, and Grok. Users... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-10T20:24:22Z | 2026-04-27T05:05:58Z | 2026-04-24T18:59:38Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | liteblogging | 87,638,337 | MDQ6VXNlcjg3NjM4MzM3 | User | false |
openclaw/openclaw | 3,923,394,379 | I_kwDOQb6kR87p2j9L | 13,657 | https://github.com/openclaw/openclaw/issues/13657 | https://api.github.com/repos/openclaw/openclaw/issues/13657 | Refine Outreach Development Workflow | Establish a formal process for outreach engineering. Requirements:
- Link all work to GitHub issues.
- Small, atomic commits with descriptive messages.
- Staging/Dry-run modes for all automation logic.
- Systematic documentation of blockers (e.g. Moonshot 401 error). | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T20:24:29Z | 2026-04-27T05:05:57Z | 2026-04-24T18:59:41Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | gitfitbro | 68,287,769 | MDQ6VXNlcjY4Mjg3NzY5 | User | false |
openclaw/openclaw | 3,923,420,366 | I_kwDOQb6kR87p2qTO | 13,660 | https://github.com/openclaw/openclaw/issues/13660 | https://api.github.com/repos/openclaw/openclaw/issues/13660 | Add --force flag to 'openclaw cron run' command for manual job execution | ## Problem
Currently, `openclaw cron run <job-id>` respects the job schedule and returns "not-due" if the job has already fired (even if it failed silently). This makes it impossible to manually rerun a failed job without modifying its schedule.
### Example Scenario
```bash
# Radar validation cron runs every 5 minut... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-10T20:30:51Z | 2026-04-27T05:05:56Z | 2026-04-24T18:59:44Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ben1440 | 258,074,641 | U_kgDOD2HoEQ | User | false |
openclaw/openclaw | 3,924,147,273 | I_kwDOQb6kR87p5bxJ | 13,754 | https://github.com/openclaw/openclaw/issues/13754 | https://api.github.com/repos/openclaw/openclaw/issues/13754 | Feature request: session:compaction hook event | ## Summary
Request a new hook event `session:compaction` (or `session:post-compaction`) that fires after auto-compaction or `/compact` completes.
## Use Case
Agents with persistent memory systems (daily logs, project files) need to re-bootstrap context after compaction. Currently:
1. `memoryFlush` runs **before** c... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-11T00:24:14Z | 2026-04-27T05:05:55Z | 2026-04-24T18:59:47Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | coleclawson | 259,448,346 | U_kgDOD3beGg | User | false |
openclaw/openclaw | 3,924,197,253 | I_kwDOQb6kR87p5n-F | 13,760 | https://github.com/openclaw/openclaw/issues/13760 | https://api.github.com/repos/openclaw/openclaw/issues/13760 | Feature Request: Circuit Breaker for Model Fallback Chain on Billing/Quota Failures | **Feature Request**
**Problem**: When a provider in the model fallback chain fails due to billing or quota issues (e.g., 402 errors for insufficient credits), OpenClaw retries the failing provider on each new message before moving to the next fallback. This causes repeated error messages and delays in response, disrup... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-11T00:47:32Z | 2026-04-27T05:05:55Z | 2026-04-24T18:59:50Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ToneLoke | 10,687,151 | MDQ6VXNlcjEwNjg3MTUx | User | false |
openclaw/openclaw | 3,924,232,765 | I_kwDOQb6kR87p5wo9 | 13,769 | https://github.com/openclaw/openclaw/issues/13769 | https://api.github.com/repos/openclaw/openclaw/issues/13769 | Feature: Add Claude Opus 4.6 model support | ## Summary
Add Claude Opus 4.6 (claude-opus-4-6) as an available model option.
## Problem
Claude Opus 4.6 is a newer model variant but isn't available in the model selection.
## Proposed Solution
Add `anthropic/claude-opus-4-6` to the model catalog with appropriate alias support (e.g. `opus46`).
## Notes
- Should wo... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-11T01:05:45Z | 2026-04-27T05:05:54Z | 2026-04-24T18:59:53Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | danieljmnewton | 235,655,237 | U_kgDODgvQRQ | User | false |
openclaw/openclaw | 3,924,264,111 | I_kwDOQb6kR87p54Sv | 13,776 | https://github.com/openclaw/openclaw/issues/13776 | https://api.github.com/repos/openclaw/openclaw/issues/13776 | Feature Request: Script Hooks for Fallback Failures to Trigger Local Server Start | **Feature Request**
**Problem**: When all cloud providers in the model fallback chain fail (due to billing issues, quotas, or outages), OpenClaw attempts to use the next model but has no mechanism to trigger a script or external action to start a local model server (e.g., Ollama, LM Studio). This requires manual inter... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-11T01:21:33Z | 2026-04-27T05:05:53Z | 2026-04-24T18:59:56Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ToneLoke | 10,687,151 | MDQ6VXNlcjEwNjg3MTUx | User | false |
openclaw/openclaw | 3,924,276,681 | I_kwDOQb6kR87p57XJ | 13,781 | https://github.com/openclaw/openclaw/issues/13781 | https://api.github.com/repos/openclaw/openclaw/issues/13781 | [Feature]: Ability to shut clawdbot down | ## Summary
Would be nice to be able to shut the related openclaw processes down, without uninstall. | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-11T01:28:13Z | 2026-04-27T05:05:52Z | 2026-04-24T18:59:58Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | bardicreels | 180,100,244 | U_kgDOCrwclA | User | false |
openclaw/openclaw | 3,924,591,782 | I_kwDOQb6kR87p7ISm | 13,858 | https://github.com/openclaw/openclaw/issues/13858 | https://api.github.com/repos/openclaw/openclaw/issues/13858 | [Feature]: Add DingTalk (钉钉) Channel Support | # Feature Request: Add DingTalk (钉钉) Channel Support
## Background
DingTalk is one of China's leading enterprise communication platforms with hundreds of millions of users. Adding DingTalk support would greatly expand OpenClaw's reach to Chinese enterprise users who rely on DingTalk for daily communication and collab... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-11T03:46:41Z | 2026-04-27T05:05:52Z | 2026-04-24T19:00:01Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | largezhou | 13,796,692 | MDQ6VXNlcjEzNzk2Njky | User | false |
openclaw/openclaw | 3,924,604,775 | I_kwDOQb6kR87p7Ldn | 13,860 | https://github.com/openclaw/openclaw/issues/13860 | https://api.github.com/repos/openclaw/openclaw/issues/13860 | 飞书 | ## 功能请求：飞书 (Feishu/Lark) 渠道支持
### 场景描述
我需要在飞书上使用 OpenClaw 进行消息收发，目前团队工作沟通主要在飞书进行。
### 期望功能
- 飞书机器人消息接收与发送
- 群聊支持
- 私聊支持
- Webhook 事件订阅
### 备选方案
如果官方暂无计划，支持 MCP 集成方式也可以接受。
### 参考
飞书开放平台: https://open.feishu.cn/
--- | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-11T03:54:11Z | 2026-04-27T05:05:51Z | 2026-04-24T19:00:04Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | zep-rgb | 259,180,028 | U_kgDOD3LF_A | User | false |
openclaw/openclaw | 3,924,666,664 | I_kwDOQb6kR87p7ako | 13,871 | https://github.com/openclaw/openclaw/issues/13871 | https://api.github.com/repos/openclaw/openclaw/issues/13871 | Feature Request: Browser-based VAD/STT voice input in Control UI WebChat | ## Summary
Add a microphone button with Voice Activity Detection (VAD) and Speech-to-Text (STT) to the Control UI's WebChat, enabling voice input directly from the browser.
## Problem
Currently, voice input (Talk mode with VAD auto-detect) only works on native apps (macOS, iOS, Android). The browser-based Control UI... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-11T04:25:47Z | 2026-04-27T05:05:50Z | 2026-04-24T19:00:07Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Mundoai | 189,129,845 | U_kgDOC0XkdQ | User | false |
openclaw/openclaw | 3,924,835,046 | I_kwDOQb6kR87p8Drm | 13,900 | https://github.com/openclaw/openclaw/issues/13900 | https://api.github.com/repos/openclaw/openclaw/issues/13900 | Feature Request: Ephemeral Cron Sessions | # Feature Request: Ephemeral Cron Sessions
**Repository:** github.com/openclaw/openclaw
**Type:** Feature Request
**Priority:** Medium-High
---
## Problem
Cron sessions accumulate context tokens over time, eventually hitting the 200k token limit. This happens because:
1. Each cron job run adds to the session's con... | closed | completed | true | 4 | [
"enhancement"
] | [] | 2026-02-11T05:44:19Z | 2026-04-27T05:05:49Z | 2026-04-24T19:00:11Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | jeremyknows | 237,305,675 | U_kgDODiT_Sw | User | false |
openclaw/openclaw | 3,925,080,685 | I_kwDOQb6kR87p8_pt | 13,923 | https://github.com/openclaw/openclaw/issues/13923 | https://api.github.com/repos/openclaw/openclaw/issues/13923 | [Feature]: Provider rate-limit / quota query tool | ## Description
Models currently have no way to query their own provider-level rate limits or quota usage. This forces workarounds like browser scraping or manual checking.
## Use Case
We run claude-opus-4-6 on a Claude Max 20x plan. The model needs to know:
- Current weekly usage percentage (we've hit 97% and had to... | closed | completed | true | 4 | [
"enhancement"
] | [] | 2026-02-11T07:07:43Z | 2026-04-27T05:05:48Z | 2026-04-24T19:00:14Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ukr-coder | 246,043,930 | U_kgDODqpVGg | User | false |
openclaw/openclaw | 3,925,321,334 | I_kwDOQb6kR87p96Z2 | 13,966 | https://github.com/openclaw/openclaw/issues/13966 | https://api.github.com/repos/openclaw/openclaw/issues/13966 | [Feature]: Feishu Streaming Messages + CardKit Live Typing Effect | ### Description
Request to add support for:
1. **Streaming message output** - Show typing indicator while AI is generating responses
2. **CardKit live typing effect** - Real-time typewriter effect in Feishu cards
### Use Case
Currently, Feishu messages are sent as static blocks, which feels unnatural for AI conversati... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-11T08:23:54Z | 2026-04-27T05:05:48Z | 2026-04-24T19:00:17Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Jiachi-Deng | 190,316,763 | U_kgDOC1gA2w | User | false |
openclaw/openclaw | 3,925,541,201 | I_kwDOQb6kR87p-wFR | 14,008 | https://github.com/openclaw/openclaw/issues/14008 | https://api.github.com/repos/openclaw/openclaw/issues/14008 | [Feature]: add wecom(企业微信) channel extension | ## Background
WeCom (企业微信) is Tencent's enterprise communication and collaboration platform, widely adopted across Chinese businesses and organizations. Adding WeCom support would greatly expand OpenClaw's reach to Chinese enterprise users who rely on WeCom for daily communication and collaboration.
## We Have a Prod... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-11T09:25:45Z | 2026-04-27T05:05:47Z | 2026-04-24T19:00:20Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | xl370869-art | 238,406,348 | U_kgDODjXKzA | User | false |
openclaw/openclaw | 3,925,925,616 | I_kwDOQb6kR87qAN7w | 14,044 | https://github.com/openclaw/openclaw/issues/14044 | https://api.github.com/repos/openclaw/openclaw/issues/14044 | [Feature]: Integration onchain | I am a solana rust infra dev. I am going to integration openclaw with onchain program
is it possible or any guide and example?
| closed | not_planned | true | 1 | [
"enhancement",
"r: support"
] | [] | 2026-02-11T10:59:42Z | 2026-04-27T05:05:46Z | 2026-04-24T19:00:23Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | roswelly | 60,659,040 | MDQ6VXNlcjYwNjU5MDQw | User | false |
openclaw/openclaw | 3,925,969,761 | I_kwDOQb6kR87qAYth | 14,049 | https://github.com/openclaw/openclaw/issues/14049 | https://api.github.com/repos/openclaw/openclaw/issues/14049 | [Feature]: Qdrant-based Persistent Memory for OpenClaw Agents |
**Status:** 📝 Draft for submission
**Author:** @gamersalpha
**Date:** 2026-02-11
---
## 🎯 Title
**Integrate Qdrant as Native Memory Layer for OpenClaw Agents**
## 📋 Problem Statement
OpenClaw agents currently rely on flat text files (MD, JSON) for memory storage:
- `MEMORY.md` - Long-term curated memories (mai... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-11T11:10:22Z | 2026-04-27T05:05:46Z | 2026-04-24T19:00:27Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | gamersalpha | 32,466,950 | MDQ6VXNlcjMyNDY2OTUw | User | false |
openclaw/openclaw | 3,926,083,789 | I_kwDOQb6kR87qA0jN | 14,059 | https://github.com/openclaw/openclaw/issues/14059 | https://api.github.com/repos/openclaw/openclaw/issues/14059 | [Code Quality]: It would be nice to refector the main folder into organized folders with their own perimeters | ## Summary
The current naming conventions will only get us so far - soon duplicate code will start to be generated if something is not done.
## Proposed solution
organised folder with many different files that do different things into many different folders that each do one thing.
## Alternatives considered
none.... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-11T11:38:20Z | 2026-04-27T05:05:45Z | 2026-04-24T19:00:30Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Josephrp | 18,212,928 | MDQ6VXNlcjE4MjEyOTI4 | User | false |
openclaw/openclaw | 3,926,424,947 | I_kwDOQb6kR87qCH1z | 14,083 | https://github.com/openclaw/openclaw/issues/14083 | https://api.github.com/repos/openclaw/openclaw/issues/14083 | Linux: Port diagnostics should fall back to `ss` when `lsof` is missing | ## Problem
Port conflict diagnostics use `lsof -nP -iTCP:PORT -sTCP:LISTEN` to identify what process owns a port. When `lsof` is not installed (common on minimal Linux servers, containers, and cloud VMs), the fallback is a generic socket bind test that reports "Port is in use but process details are unavailable (insta... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-11T12:57:34Z | 2026-04-27T05:05:44Z | 2026-04-24T19:00:33Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | holistic-medical-supply-base | 249,014,699 | U_kgDODtepqw | User | false |
openclaw/openclaw | 3,926,426,262 | I_kwDOQb6kR87qCIKW | 14,084 | https://github.com/openclaw/openclaw/issues/14084 | https://api.github.com/repos/openclaw/openclaw/issues/14084 | Linux: Doctor diagnostics significantly thinner than macOS | ## Problem
The `openclaw doctor` command has substantially deeper diagnostics and auto-repair on macOS than on Linux:
### macOS gets:
- LaunchAgent bootstrap repair (`maybeRepairLaunchAgentBootstrap`)
- `launchctl getenv` override detection for gateway tokens/passwords
- Deprecated `MOLTBOT_*`/`CLAWDBOT_*` env var de... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-11T12:57:51Z | 2026-04-27T05:05:42Z | 2026-04-24T19:00:36Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | holistic-medical-supply-base | 249,014,699 | U_kgDODtepqw | User | false |
openclaw/openclaw | 3,927,009,831 | I_kwDOQb6kR87qEWon | 14,129 | https://github.com/openclaw/openclaw/issues/14129 | https://api.github.com/repos/openclaw/openclaw/issues/14129 | Model switching blocked by images in conversation | When a conversation contains images, OpenClaw prevents switching to text-only models (like qwen3-coder) and only allows multimodal models. This is a usability limitation that blocks model testing and switching. | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-11T14:56:44Z | 2026-04-27T05:05:42Z | 2026-04-24T19:00:39Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | kataguru | 120,272,927 | U_kgDOBys4Hw | User | false |
openclaw/openclaw | 3,927,384,223 | I_kwDOQb6kR87qFyCf | 14,159 | https://github.com/openclaw/openclaw/issues/14159 | https://api.github.com/repos/openclaw/openclaw/issues/14159 | Feature Request: Pre-tool hooks (gate scripts that block tool execution) | ## Problem
Agents can be instructed to run gate scripts before taking certain actions (e.g., spawning sub-agents, reporting completion), but compliance is voluntary. The agent *chooses* to run the gate — it can skip it under pressure, urgency, or context loss.
## Proposed Solution
Add a `preToolHooks` config option ... | closed | completed | true | 7 | [
"enhancement"
] | [] | 2026-02-11T16:09:36Z | 2026-04-27T05:05:41Z | 2026-04-24T19:00:42Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | matthewdecelles | 132,421,579 | U_kgDOB-SXyw | User | false |
openclaw/openclaw | 3,927,511,154 | I_kwDOQb6kR87qGRBy | 14,180 | https://github.com/openclaw/openclaw/issues/14180 | https://api.github.com/repos/openclaw/openclaw/issues/14180 | [Feature]: Analysis: The OpenClaw Agent Corpora | I am working from first principals agnostic of and "hub's" or skills, not because of the benefits, but because the network tunneling creates a few broken eggs during setup.
the first, and possibly only agentic design that works in my tart vm have a web console.
when i make a request, i notice a built-in paradox of ... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-11T16:35:27Z | 2026-04-27T05:05:40Z | 2026-04-24T19:00:46Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | jnorthrup | 73,514 | MDQ6VXNlcjczNTE0 | User | false |
openclaw/openclaw | 3,927,855,592 | I_kwDOQb6kR87qHlHo | 14,210 | https://github.com/openclaw/openclaw/issues/14210 | https://api.github.com/repos/openclaw/openclaw/issues/14210 | Add outbound message hook (message:sent) OR error suppression for provider overload/rate_limit | ## Problem Statement
When LLM providers (Anthropic, OpenAI, etc.) return `overloaded_error` or rate limit responses, OpenClaw sends a user-facing message:
```
The AI service is temporarily overloaded. Please try again in a moment.
```
**Issues:**
1. Users receive "spam" error messages in Telegram/WhatsApp during hig... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-11T17:54:34Z | 2026-04-27T05:05:39Z | 2026-04-24T19:00:49Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | AlekseyRodkin | 208,993,710 | U_kgDODHT9rg | User | false |
openclaw/openclaw | 3,929,203,492 | I_kwDOQb6kR87qMuMk | 14,329 | https://github.com/openclaw/openclaw/issues/14329 | https://api.github.com/repos/openclaw/openclaw/issues/14329 | Feature: Emit user messages in Gateway WebSocket chat events (state: "user") | ## Summary
The Gateway WebSocket protocol's `chat` event only emits assistant-side states (`delta`, `final`, `aborted`, `error`). There is no event for incoming user messages. External consumers (dashboards, bridges, monitoring tools) that connect as operators have no way to receive user messages in real-time.
## Pro... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-12T00:01:51Z | 2026-04-27T05:05:38Z | 2026-04-24T19:00:52Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | lucletoffe | 15,689,941 | MDQ6VXNlcjE1Njg5OTQx | User | false |
openclaw/openclaw | 3,929,292,341 | I_kwDOQb6kR87qND41 | 14,351 | https://github.com/openclaw/openclaw/issues/14351 | https://api.github.com/repos/openclaw/openclaw/issues/14351 | [Feature]: integrate with ITIR suite, ideal openclaw companion preceding industry trends, seeking devs | ## Summary
Describe the problem you are trying to solve or the opportunity you see.
We are trying to avoid a future of the rentier class. For this we need open, free, and inspectable tools.
https://venturebeat.com/data/observational-memory-cuts-ai-agent-costs-10x-and-outscores-rag-on-long
https://thenewstack.io/tho... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-12T00:41:54Z | 2026-04-27T05:05:38Z | 2026-04-24T19:00:55Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | chboishabba | 26,853,614 | MDQ6VXNlcjI2ODUzNjE0 | User | false |
openclaw/openclaw | 3,929,449,143 | I_kwDOQb6kR87qNqK3 | 14,386 | https://github.com/openclaw/openclaw/issues/14386 | https://api.github.com/repos/openclaw/openclaw/issues/14386 | [Feature]: RAM Cache for Personality Files | ## **Summary**
Implement RAM caching for personality/memory .md files to eliminate 2-3 second disk I/O delays on every session startup and message processing.
## **Problem Statement**
Current OpenClaw reads personality files (SOUL.md, MEMORY.md, etc.) from disk for every:
- **Session startup** (2-3 second delay loadin... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-12T01:51:43Z | 2026-04-27T05:05:37Z | 2026-04-24T19:00:58Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ro046283 | 114,316,970 | U_kgDOBtBWqg | User | false |
openclaw/openclaw | 3,929,725,343 | I_kwDOQb6kR87qOtmf | 14,427 | https://github.com/openclaw/openclaw/issues/14427 | https://api.github.com/repos/openclaw/openclaw/issues/14427 | [Feature]: replyInThread config option for Mattermost channels | ## Summary
The Mattermost plugin has no config option to automatically reply in threads. When the bot replies to a channel message, it posts as a new top-level message instead of threading under the triggering message. This makes channels noisy and hard to follow, especially with multiple conversations happening.
Addi... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-12T03:52:39Z | 2026-04-27T05:05:36Z | 2026-04-24T19:01:01Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | rockinyp | 207,491,833 | U_kgDODF4S-Q | User | false |
openclaw/openclaw | 3,929,736,499 | I_kwDOQb6kR87qOwUz | 14,428 | https://github.com/openclaw/openclaw/issues/14428 | https://api.github.com/repos/openclaw/openclaw/issues/14428 | Feature request: per-agent humanDelay configuration | ## Summary
Currently `humanDelay` can only be configured at the `agents.defaults` level, which applies globally to all agents.
## Use Case
In a multi-agent setup (e.g., personal assistant + public-facing agent), it would be useful to have different response timing behaviors:
- Personal assistant: instant responses
- P... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-12T03:57:26Z | 2026-04-27T05:05:35Z | 2026-04-24T19:01:04Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Parcher-ia | 244,210,402 | U_kgDODo5a4g | User | false |
openclaw/openclaw | 3,929,743,707 | I_kwDOQb6kR87qOyFb | 14,431 | https://github.com/openclaw/openclaw/issues/14431 | https://api.github.com/repos/openclaw/openclaw/issues/14431 | [Feature]: Feishu inbound dedup cache lost on SIGUSR1 restart, causing duplicate message processing | ## Summary
When the gateway restarts in-process (SIGUSR1, triggered by config changes, model switches, or `/new`), the `inboundDedupeCache` is re-initialized as an empty Map. The Feishu/Lark SDK's WebSocket client then reconnects and replays recent events. Since the cache is empty, previously-processed messages pass `... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-12T04:01:12Z | 2026-04-27T05:05:35Z | 2026-04-24T19:01:07Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Jesse-HH | 73,918,103 | MDQ6VXNlcjczOTE4MTAz | User | false |
openclaw/openclaw | 3,930,205,690 | I_kwDOQb6kR87qQi36 | 14,495 | https://github.com/openclaw/openclaw/issues/14495 | https://api.github.com/repos/openclaw/openclaw/issues/14495 | [Plugin Proposal] @openclaw/shield — API key health monitoring & multi-provider failover | ## Problem
When an API key gets blocked, rate-limited, or expires, the entire OpenClaw agent system goes down. This is a common failure mode — users typically configure a single provider (e.g., Anthropic) and when that key fails, all agents fail simultaneously.
### Error encountered
```
⚠️ Agent failed before reply:... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-12T07:04:46Z | 2026-04-27T05:05:34Z | 2026-04-24T19:01:10Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | 0xfreeoffical | 130,159,659 | U_kgDOB8IUKw | User | false |
openclaw/openclaw | 3,930,297,113 | I_kwDOQb6kR87qQ5MZ | 14,513 | https://github.com/openclaw/openclaw/issues/14513 | https://api.github.com/repos/openclaw/openclaw/issues/14513 | Docs: Recommend Tailscale Serve for HTTPS access | ## Summary
When accessing the Control UI over HTTP from a non-localhost address, users see:
```
control ui requires HTTPS or localhost (secure context)
```
The error message mentions `gateway.controlUi.allowInsecureAuth: true` as a workaround, but a better solution is **Tailscale Serve**, which provides automatic HT... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-12T07:34:51Z | 2026-04-27T05:05:33Z | 2026-04-24T19:01:13Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | wolffcatskyy | 153,394,885 | U_kgDOCSSexQ | User | false |
openclaw/openclaw | 3,931,309,124 | I_kwDOQb6kR87qUwRE | 14,606 | https://github.com/openclaw/openclaw/issues/14606 | https://api.github.com/repos/openclaw/openclaw/issues/14606 | UI 'Compact Session Log' should use smart LLM-based compaction instead of raw file truncation | ## Problem
The **"Compact Session Log"** button in the OpenClaw UI (macOS app / web control panel) performs a raw JSONL file truncation — it simply keeps the last N lines and discards the rest. This results in **complete context loss** after compaction.
Meanwhile, the `/compact` text command uses the Pi Coding Agent ... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-12T11:24:33Z | 2026-04-27T05:05:32Z | 2026-04-24T19:01:17Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | yanfeatherai | 261,148,312 | U_kgDOD5DOmA | User | false |
openclaw/openclaw | 3,931,313,346 | I_kwDOQb6kR87qUxTC | 14,607 | https://github.com/openclaw/openclaw/issues/14607 | https://api.github.com/repos/openclaw/openclaw/issues/14607 | [Feature]: when extensions add dingtalk ? | ## Summary
Describe the problem you are trying to solve or the opportunity you see.
## Proposed solution
What would you like Clawdbot to do?
## Alternatives considered
Any other approaches you have considered?
## Additional context
Links, screenshots, or related issues.
| closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-12T11:25:27Z | 2026-04-27T05:05:32Z | 2026-04-24T19:01:19Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | hhrhhrhhr | 25,874,816 | MDQ6VXNlcjI1ODc0ODE2 | User | false |
openclaw/openclaw | 3,932,018,650 | I_kwDOQb6kR87qXdfa | 14,688 | https://github.com/openclaw/openclaw/issues/14688 | https://api.github.com/repos/openclaw/openclaw/issues/14688 | Onboarding: default multi-channel DM setups to per-channel session isolation | ## Summary
When users onboard multiple DM channels (e.g. Telegram + Feishu/QQ/Slack) and leave `session.dmScope` unset, direct messages default to `main` and can unintentionally share one DM context (`agent:main:main`).
This is currently documented, but in practice many users still hit cross-channel DM context bleed d... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-12T13:55:17Z | 2026-04-27T05:05:31Z | 2026-04-24T19:01:22Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | constansino | 65,108,260 | MDQ6VXNlcjY1MTA4MjYw | User | false |
openclaw/openclaw | 3,932,222,906 | I_kwDOQb6kR87qYPW6 | 14,710 | https://github.com/openclaw/openclaw/issues/14710 | https://api.github.com/repos/openclaw/openclaw/issues/14710 | [Feature]: Echo Guard — Prompt Injection Detection via Input Echo | <html><head></head><body><h1>Feature Request: Echo Guard — Prompt Injection Detection via Input Echo Verification</h1>
<h2>Summary</h2>
<p>Introduce an <strong>Echo Guard</strong> security layer that detects prompt injection attempts by asking the LLM to simply echo back the user's input in an independent, tool-free ca... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-02-12T14:38:47Z | 2026-04-27T05:05:30Z | 2026-04-24T19:01:25Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | mourad-ghafiri | 42,439,055 | MDQ6VXNlcjQyNDM5MDU1 | User | false |
openclaw/openclaw | 3,932,450,063 | I_kwDOQb6kR87qZG0P | 14,735 | https://github.com/openclaw/openclaw/issues/14735 | https://api.github.com/repos/openclaw/openclaw/issues/14735 | feat: implement message:sent hook event for outbound message post-processing | ## Summary
Add a `message:sent` hook event that fires before an outbound message is delivered to a channel. This would allow hooks to transform, annotate, or enrich messages on a per-channel basis.
## Use Case
I want to append context window usage info (e.g. `📊 45k/200k`) as a footer to outgoing Telegram messages, ... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-12T15:24:15Z | 2026-04-27T05:05:29Z | 2026-04-24T19:01:28Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | lumenclaw-cloud | 260,919,617 | U_kgDOD41RQQ | User | false |
openclaw/openclaw | 3,932,772,527 | I_kwDOQb6kR87qaViv | 14,779 | https://github.com/openclaw/openclaw/issues/14779 | https://api.github.com/repos/openclaw/openclaw/issues/14779 | Feature: Add plugin hooks for token cost management | ### Problem
OpenClaw instances can accumulate significant token costs through:
1. **No dynamic model routing** - Always using the same model tier regardless of task complexity (simple file operations use the same expensive model as strategic planning)
2. **Unlimited context injection** - Loading all workspace files +... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-12T16:32:12Z | 2026-04-27T05:05:29Z | 2026-04-24T19:01:32Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | pfaria32 | 42,222,117 | MDQ6VXNlcjQyMjIyMTE3 | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.