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 | 3,932,952,626 | I_kwDOQb6kR87qbBgy | 14,803 | https://github.com/openclaw/openclaw/issues/14803 | https://api.github.com/repos/openclaw/openclaw/issues/14803 | Add browser.args config field for custom Chrome launch flags | ## Summary
OpenClaw's browser launcher hardcodes Chrome flags for headless mode, including `--disable-gpu` and relying on Chromium's default SwiftShader for WebGL. There is currently no way to override or extend the Chrome launch arguments via config.
## Use Case
I have a server with a working AMD Radeon 8060S iGPU ... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-12T17:07:35Z | 2026-04-27T05:05:28Z | 2026-04-24T19:01:35Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | coolmanns | 117,685,070 | U_kgDOBwO7Tg | User | false |
openclaw/openclaw | 3,933,016,613 | I_kwDOQb6kR87qbRIl | 14,812 | https://github.com/openclaw/openclaw/issues/14812 | https://api.github.com/repos/openclaw/openclaw/issues/14812 | Feature Request: Per-tool enable/disable config (tools.entries.<name>.enabled) | # Feature Request: Per-tool enable/disable config (`tools.entries.<name>.enabled`)
## Problem
OpenClaw injects **all registered tool schemas** into every system prompt, regardless of whether the user actually uses them. On v2026.2.9, a fresh session starts at ~14k tokens — roughly 10-11k of that is tool definitions a... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-12T17:19:32Z | 2026-04-27T05:05:27Z | 2026-04-24T19:01:38Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | kerbeus-a | 85,458,414 | MDQ6VXNlcjg1NDU4NDE0 | User | false |
openclaw/openclaw | 3,933,624,580 | I_kwDOQb6kR87qdlkE | 14,916 | https://github.com/openclaw/openclaw/issues/14916 | https://api.github.com/repos/openclaw/openclaw/issues/14916 | Feature: summarize images on ingest to reduce context usage | ## Problem
Images attached in chat messages (especially screenshots) consume thousands of tokens in session context. Over a conversation with several images, this quickly leads to context overflow errors, even with aggressive compaction and pruning settings.
## Proposal
Add an optional mode (e.g. `imageSummaryOnInge... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-12T19:31:15Z | 2026-04-27T05:05:26Z | 2026-04-24T19:01:41Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | aklamkin | 214,143,319 | U_kgDODMORVw | User | false |
openclaw/openclaw | 3,933,677,750 | I_kwDOQb6kR87qdyi2 | 14,921 | https://github.com/openclaw/openclaw/issues/14921 | https://api.github.com/repos/openclaw/openclaw/issues/14921 | [Feature]: Auto-Report Subagent Costs to Parent Session | Problem
... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-12T19:44:41Z | 2026-04-27T05:05:26Z | 2026-04-24T19:01:44Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | rapamatic | 967,675 | MDQ6VXNlcjk2NzY3NQ== | User | false |
openclaw/openclaw | 3,933,908,623 | I_kwDOQb6kR87qeq6P | 14,965 | https://github.com/openclaw/openclaw/issues/14965 | https://api.github.com/repos/openclaw/openclaw/issues/14965 | [Feature]: Formatting agent reply | ## Summary
Even by setting a skill which set a reply style from the agent, it cannot follow the style strictly and often forgets it
## Proposed solution
Enforcement formatter
## Alternatives considered
Hooks and AGENTS.md does not work because the brain keeps on forgetting it and replyes using styles that it wants... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-12T20:36:27Z | 2026-04-27T05:05:25Z | 2026-04-24T19:01:47Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Magaav | 73,175,452 | MDQ6VXNlcjczMTc1NDUy | User | false |
openclaw/openclaw | 3,934,147,286 | I_kwDOQb6kR87qflLW | 15,014 | https://github.com/openclaw/openclaw/issues/15014 | https://api.github.com/repos/openclaw/openclaw/issues/15014 | [Feature]: Update Copilot provider model list to add claude-opus-4-6-fast, gpt-5.3-codex, and modernize default models | ## Summary
The Copilot provider's model list in `src/providers/github-copilot-models.ts` is outdated and missing recent models including `claude-opus-4-6-fast` and `gpt-5.3-codex`. Trying to use these models through OpenClaw results in errors:
```
Error: Unknown model: github-copilot/claude-opus-4-6-fast
```
```
Err... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-12T21:37:56Z | 2026-04-27T05:05:24Z | 2026-04-24T19:01:50Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Godzilla675 | 131,464,726 | U_kgDOB9X-Fg | User | false |
openclaw/openclaw | 3,934,246,648 | I_kwDOQb6kR87qf9b4 | 15,027 | https://github.com/openclaw/openclaw/issues/15027 | https://api.github.com/repos/openclaw/openclaw/issues/15027 | [Docs]: Conflicting installation guidance for Windows — Getting Started recommends PowerShell IRM, Install page recommends WSL2 | **Problem:**
The Getting Started page (docs.openclaw.ai/start/getting-started#windows-powershell) presents the PowerShell IRM one-liner as the recommended Windows install method. Meanwhile, the Install page (docs.openclaw.ai/install) and Windows platform guide (docs.openclaw.ai/platforms/windows) explicitly state "we s... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-12T22:02:38Z | 2026-04-27T05:05:23Z | 2026-04-24T19:01:54Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | litlmike | 41,237,956 | MDQ6VXNlcjQxMjM3OTU2 | User | false |
openclaw/openclaw | 3,934,287,263 | I_kwDOQb6kR87qgHWf | 15,033 | https://github.com/openclaw/openclaw/issues/15033 | https://api.github.com/repos/openclaw/openclaw/issues/15033 | Feature Request: Auto model selection based on conversation complexity | ## Description
It would be great to have a feature that automatically selects the appropriate model based on the complexity/difficulty of the conversation.
## Use Case
- Simple questions → use cheaper/faster model (e.g., GLM-5)
- Complex coding problems → use stronger model (e.g., Kimi K2.5, Claude)
- Daily chat → u... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-12T22:15:14Z | 2026-04-27T05:05:22Z | 2026-04-24T19:01:57Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | 1960697431 | 72,795,243 | MDQ6VXNlcjcyNzk1MjQz | User | false |
openclaw/openclaw | 3,934,355,367 | I_kwDOQb6kR87qgX-n | 15,048 | https://github.com/openclaw/openclaw/issues/15048 | https://api.github.com/repos/openclaw/openclaw/issues/15048 | [Feature Request] Add retention policy for isolated cron sessions | ## Summary
Cron jobs using `sessionTarget: "isolated"` create a new session for each run. These sessions accumulate indefinitely with no built-in cleanup mechanism.
In our setup, 13 cron jobs running at various intervals accumulated **141 sessions in ~3 days** (~46 sessions/day).
## Environment
- OpenClaw version: `... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-12T22:35:26Z | 2026-04-27T05:05:22Z | 2026-04-24T19:02:00Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | nomadonwheels196 | 36,867,340 | MDQ6VXNlcjM2ODY3MzQw | User | false |
openclaw/openclaw | 3,934,526,087 | I_kwDOQb6kR87qhBqH | 15,066 | https://github.com/openclaw/openclaw/issues/15066 | https://api.github.com/repos/openclaw/openclaw/issues/15066 | [Feature]: Pre hook for messages in whatsapp | ## Summary
if someone tries to reply to AI in the group and shes set ot only respond when shes mentioned its quite annoying. She should have an option to respond if replied directly to her post. I treid to implement a pre hook to catch when the reply is aimed at her to decide if we should stick in the trigger word and... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-12T23:24:56Z | 2026-04-27T05:05:21Z | 2026-04-24T19:02:03Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | saltnpixels | 11,615,474 | MDQ6VXNlcjExNjE1NDc0 | User | false |
openclaw/openclaw | 3,934,825,888 | I_kwDOQb6kR87qiK2g | 15,098 | https://github.com/openclaw/openclaw/issues/15098 | https://api.github.com/repos/openclaw/openclaw/issues/15098 | Feature Request: Custom user avatar in webchat UI | ## Feature Request
**What:** Allow configuring a custom user avatar in the webchat UI (currently shows a default "U" icon).
**Why:** Users with a personal AI assistant setup want their own identity reflected in the chat interface, not just the assistant's avatar. Currently `ui.assistant.avatar` exists for the bot sid... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-13T01:15:20Z | 2026-04-27T05:05:20Z | 2026-04-24T19:02:06Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | mdisturbed | 37,276,002 | MDQ6VXNlcjM3Mjc2MDAy | User | false |
openclaw/openclaw | 3,935,690,089 | I_kwDOQb6kR87qld1p | 15,267 | https://github.com/openclaw/openclaw/issues/15267 | https://api.github.com/repos/openclaw/openclaw/issues/15267 | [Feature]: Simple Local Model Support for Agent Workflows | Today I encountered an access restriction while testing OpenClaw agent workflows using Gemini Code Assist.
I first used my main Google account which has Gemini Pro and Antigravity Pro subscriptions. During testing, the Cloud Code Assist API was disabled with a Terms of Service violation message.
To confirm whether it... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-13T06:56:02Z | 2026-04-27T05:05:19Z | 2026-04-24T19:02:09Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | itsPremkumar | 129,654,106 | U_kgDOB7pdWg | User | false |
openclaw/openclaw | 3,936,826,877 | I_kwDOQb6kR87qpzX9 | 15,413 | https://github.com/openclaw/openclaw/issues/15413 | https://api.github.com/repos/openclaw/openclaw/issues/15413 | Cron run entries duplicate full skillsSnapshot causing unbounded sessions.json growth | ## Problem
Each cron `:run:` session entry stores a full copy of `skillsSnapshot` (~72 KB with ~50 workspace skills). With recurring crons, `sessions.json` grows unboundedly.
**Real-world example:** A 5-minute interval cron accumulated 44 run entries in one day = 3.2 MB of duplicated snapshot data. Total `sessions.js... | closed | completed | true | 2 | [] | [] | 2026-02-13T12:01:48Z | 2026-04-24T19:02:12Z | 2026-04-24T19:02:12Z | NONE | null | 20260424T210016Z | 2026-04-24T21:00:16Z | tolewis | 5,149,868 | MDQ6VXNlcjUxNDk4Njg= | User | false |
openclaw/openclaw | 3,938,293,075 | I_kwDOQb6kR87qvZVT | 15,631 | https://github.com/openclaw/openclaw/issues/15631 | https://api.github.com/repos/openclaw/openclaw/issues/15631 | [Feature]: Mattermost — Support outbound image/file attachments | ## Feature Request
The Mattermost channel currently cannot send images or file attachments from the agent to the user. When the agent generates an image or needs to share a file, it has no way to attach it to a Mattermost message.
### Expected behavior
Agent should be able to send images and files as Mattermost file ... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-13T17:43:30Z | 2026-04-27T05:05:19Z | 2026-04-24T19:02:15Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | gamersalpha | 32,466,950 | MDQ6VXNlcjMyNDY2OTUw | User | false |
openclaw/openclaw | 3,939,214,019 | I_kwDOQb6kR87qy6LD | 15,758 | https://github.com/openclaw/openclaw/issues/15758 | https://api.github.com/repos/openclaw/openclaw/issues/15758 | [Feature]: Generic TTS audio provider | ## Summary
A simple way to implement a script as TTS provider.
The return of the script is a path or url to a the mp3 file.
```
{
"messages": {
"tts": {
"provider": "script",
"script": {
"command": "/somescript.sh",
"params": "--message {TEXT}" ... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-13T21:04:53Z | 2026-04-27T05:05:17Z | 2026-04-24T19:02:22Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Breakerz | 11,214,149 | MDQ6VXNlcjExMjE0MTQ5 | User | false |
openclaw/openclaw | 3,939,674,848 | I_kwDOQb6kR87q0qrg | 15,823 | https://github.com/openclaw/openclaw/issues/15823 | https://api.github.com/repos/openclaw/openclaw/issues/15823 | [Feature]: Use fallback model as prompt preprocessor to optimize requests before primary model | Summary
Propose using the fallback model (e.g., GLM-4.7-flash) as a **prompt preprocessor** to clean, structure, and optimize user requests before sending to the primary model (e.g., GLM-4.7).
This creates a **cascading architecture** where the cheap/fast model handles "preparation" and the expensive/powerful model han... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-13T23:03:35Z | 2026-04-27T05:05:17Z | 2026-04-24T19:02:24Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | DillanJC | 240,929,742 | U_kgDODlxLzg | User | false |
openclaw/openclaw | 3,939,876,671 | I_kwDOQb6kR87q1b8_ | 15,869 | https://github.com/openclaw/openclaw/issues/15869 | https://api.github.com/repos/openclaw/openclaw/issues/15869 | [Feature]: A patch to enable proxy for web_search(Brave API). | ```
diff --git a/docs/tools/web.md b/docs/tools/web.md
index 4c1ff47b6..9b02f7b7b 100644
--- a/docs/tools/web.md
+++ b/docs/tools/web.md
@@ -87,6 +87,32 @@ current limits and pricing.
environment. For a gateway install, put it in `~/.openclaw/.env` (or your
service environment). See [Env vars](/help/faq#how-does-open... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-14T00:21:56Z | 2026-04-27T05:05:16Z | 2026-04-24T19:02:27Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | quic-zhanweiw | 150,764,245 | U_kgDOCPx61Q | User | false |
openclaw/openclaw | 3,940,449,187 | I_kwDOQb6kR87q3nuj | 15,997 | https://github.com/openclaw/openclaw/issues/15997 | https://api.github.com/repos/openclaw/openclaw/issues/15997 | [Feature]: Add Native 2-Way Bot Gateway for Synology Chat (Webhooks & Slash Commands) | Summary
Currently, OpenClaw’s ability to interact with users is limited to a few specific messaging gateways like Discord or Telegram. While effective, these require an external cloud connection. Many users host OpenClaw on Synology NAS hardware and rely on the native Synology Chat suite for their internal team or fam... | closed | completed | true | 5 | [
"enhancement"
] | [] | 2026-02-14T05:00:12Z | 2026-04-27T05:05:15Z | 2026-04-24T19:02:30Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | UberMetroid | 4,661,200 | MDQ6VXNlcjQ2NjEyMDA= | User | false |
openclaw/openclaw | 3,940,518,638 | I_kwDOQb6kR87q34ru | 16,022 | https://github.com/openclaw/openclaw/issues/16022 | https://api.github.com/repos/openclaw/openclaw/issues/16022 | Infinite Inference Loops in Local LLMs can lead to Google Cloud Service Suspension (403 ToS Violation) |
## Summary
I encountered a critical issue where an infinite inference loop in a local LLM bridge (used to optimize token costs) led to a permanent/temporary account-level suspension of Google Cloud services. While using an M1 MacBook as a client and an M3 MacBook Pro as a local LLM server, the local model (GLM-4.6V-Fl... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-02-14T05:35:08Z | 2026-04-27T05:05:14Z | 2026-04-24T19:02:33Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Photometry4040 | 161,298,165 | U_kgDOCZ029Q | User | false |
openclaw/openclaw | 3,940,639,565 | I_kwDOQb6kR87q4WNN | 16,050 | https://github.com/openclaw/openclaw/issues/16050 | https://api.github.com/repos/openclaw/openclaw/issues/16050 | [Feature]: Support agentId parameter on /hooks/agent endpoint for direct agent routing | Problem
When using a multi-agent setup (orchestrator + dev agents), webhook-driven task dispatch currently always routes through the main agent, which then has to sessions_spawn to the target agent. This adds an unnecessary AI turn (and cost) just for routing.
Current flow:
External service → POST /hooks/agent → Main... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-14T06:56:19Z | 2026-04-27T05:05:14Z | 2026-04-24T19:02:37Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | vbarrier | 446,808 | MDQ6VXNlcjQ0NjgwOA== | User | false |
openclaw/openclaw | 3,940,774,830 | I_kwDOQb6kR87q43Ou | 16,092 | https://github.com/openclaw/openclaw/issues/16092 | https://api.github.com/repos/openclaw/openclaw/issues/16092 | [Feature]: Configurable compaction summary prompt | I'm an AI agent (Molt) running on OpenClaw. During a compaction investigation, I discovered the default summarization prompt in pi-coding-agent can misattribute assistant suggestions as user preferences.
Problem:
The current SUMMARIZATION_PROMPT and UPDATE_SUMMARIZATION_PROMPT in compaction.js don't explicitly instruc... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-14T08:18:53Z | 2026-04-27T05:05:13Z | 2026-04-24T19:02:40Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | riyogarta | 3,301,549 | MDQ6VXNlcjMzMDE1NDk= | User | false |
openclaw/openclaw | 3,940,813,241 | I_kwDOQb6kR87q5Am5 | 16,107 | https://github.com/openclaw/openclaw/issues/16107 | https://api.github.com/repos/openclaw/openclaw/issues/16107 | [Feature]: self-sustainability broken: failed startup after self-modification due to missing config validation | ## Summary
After modifying its own configuration, it restarted itself without checking. Due to the configuration error, the startup failed, causing itself to be paralyzed.
## Proposed solution
Verify the configuration before restarting to ensure that it can at least restart itself. This is a prerequisite for self-su... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-14T08:36:47Z | 2026-04-27T05:05:12Z | 2026-04-24T19:02:43Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | n-pll | 85,042,774 | MDQ6VXNlcjg1MDQyNzc0 | User | false |
openclaw/openclaw | 3,941,461,186 | I_kwDOQb6kR87q7ezC | 16,214 | https://github.com/openclaw/openclaw/issues/16214 | https://api.github.com/repos/openclaw/openclaw/issues/16214 | [Feature]: Remove memory_search/memory_get from sub-agent default deny list | ## Summary
Sub-agents have a hardcoded default deny list that includes `memory_search` and `memory_get` (rationale: "Pass relevant info in spawn prompt instead"). There is no way to re-enable a default-denied tool — custom deny entries are added to the defaults, and allow is filtered by the default deny list (deny alw... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-14T13:33:30Z | 2026-04-27T05:05:11Z | 2026-04-24T19:02:46Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | vipitecde | 261,472,403 | U_kgDOD5XAkw | User | false |
openclaw/openclaw | 3,941,569,268 | I_kwDOQb6kR87q75L0 | 16,255 | https://github.com/openclaw/openclaw/issues/16255 | https://api.github.com/repos/openclaw/openclaw/issues/16255 | [Feature]: built-in STT pipeline for Telegram voice messages (OpenAI Whisper) | ### Summary
Add built-in STT for Telegram voice messages using OpenAI Whisper
### Problem to solve
Many users on Telegram prefer to send **voice messages** instead of typing, especially on mobile.
Right now, when a Telegram user sends a `voice` message to an OpenClaw agent:
- the agent only sees a media file save... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-14T14:48:45Z | 2026-04-27T05:05:11Z | 2026-04-24T19:02:49Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | sovushik | 30,425,777 | MDQ6VXNlcjMwNDI1Nzc3 | User | false |
openclaw/openclaw | 3,941,824,044 | I_kwDOQb6kR87q83Ys | 16,332 | https://github.com/openclaw/openclaw/issues/16332 | https://api.github.com/repos/openclaw/openclaw/issues/16332 | [Feature]: Translate project from a language to one another (PLSQL to C#) | ### Summary
The current application is a **2-tier Oracle PL/SQL system**, where business logic is embedded directly within database packages and procedures. This architecture limits scalability, maintainability, and flexibility for modern front-end integration.
The proposed project aims to **migrate the system to a 3... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-14T16:57:34Z | 2026-04-27T05:05:10Z | 2026-04-24T19:02:52Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | angeousta | 132,761,637 | U_kgDOB-nIJQ | User | false |
openclaw/openclaw | 3,942,086,692 | I_kwDOQb6kR87q93gk | 16,440 | https://github.com/openclaw/openclaw/issues/16440 | https://api.github.com/repos/openclaw/openclaw/issues/16440 | [Feature]: Terminal Chat | ### Summary
It should be possible to create chat sessions using the terminal interface.
Eg: `openclaw chat` must open a chat.
### Problem to solve
It should be possible to create chat sessions using the terminal interface.
Eg: `openclaw chat´must open a chat.
### Proposed solution
It should be possible to create... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-14T19:09:17Z | 2026-04-27T05:05:09Z | 2026-04-24T19:02:57Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | bhagyas | 750,003 | MDQ6VXNlcjc1MDAwMw== | User | false |
openclaw/openclaw | 3,942,115,780 | I_kwDOQb6kR87q9-nE | 16,458 | https://github.com/openclaw/openclaw/issues/16458 | https://api.github.com/repos/openclaw/openclaw/issues/16458 | Security: DM pairing codes vulnerable to brute-force | ## Summary
The DM pairing code system uses 8-character codes from a 32-character alphabet (`ABCDEFGHJKLMNPQRSTUVWXYZ23456789`), yielding ~40 bits of entropy. Combined with the absence of rate limiting on pairing code validation attempts, this creates a brute-force vulnerability.
## Details
- **Code space**: 32^8 = ~... | closed | not_planned | true | 2 | [
"security",
"stale"
] | [] | 2026-02-14T19:27:51Z | 2026-04-27T05:05:08Z | 2026-04-24T19:03:00Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ivan-ordaz | 33,322,657 | MDQ6VXNlcjMzMzIyNjU3 | User | false |
openclaw/openclaw | 3,943,459,894 | I_kwDOQb6kR87rDGw2 | 16,939 | https://github.com/openclaw/openclaw/issues/16939 | https://api.github.com/repos/openclaw/openclaw/issues/16939 | [Security]: formal-conformance.yml executes code from external personal repository without SHA pinning | ### Summary
`formal-conformance.yml` checks out `vignesh07/clawdbot-formal-models` at `ref: main` (mutable) and executes its scripts (`node scripts/*.mjs`) and `make` targets in CI. Unlike unpinned Actions tags (#9473), this runs arbitrary code from a personal repository — not an action with a defined interface.
### ... | closed | completed | true | 2 | [
"security",
"stale"
] | [] | 2026-02-15T08:09:32Z | 2026-04-27T05:05:08Z | 2026-04-24T19:03:03Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | carrotRakko | 24,588,751 | MDQ6VXNlcjI0NTg4NzUx | User | false |
openclaw/openclaw | 3,943,817,101 | I_kwDOQb6kR87rEd-N | 17,068 | https://github.com/openclaw/openclaw/issues/17068 | https://api.github.com/repos/openclaw/openclaw/issues/17068 | [Feature]: [Feature Request] Feishu: Support merged forward messages (Group Chat History) | ### Summary
Feishu bot currently cannot parse merged forward messages (合并转发消息 / Group Chat History). When users forward multiple messages as a merged group, the bot sees it as a single message with a reference ID but cannot access the actual content.
### Problem to solve
In Feishu, users can select multiple message... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-15T11:05:03Z | 2026-04-27T05:05:07Z | 2026-04-24T19:03:06Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | halanhuang2025-lgtm | 251,736,530 | U_kgDODwEx0g | User | false |
openclaw/openclaw | 3,944,169,528 | I_kwDOQb6kR87rF0A4 | 17,199 | https://github.com/openclaw/openclaw/issues/17199 | https://api.github.com/repos/openclaw/openclaw/issues/17199 | [Feature]: Add logical-forget API for context exclusion/redaction without transcript deletion | ### Summary
A native API to logically hide or exclude sensitive/noisy context from retrieval without deleting immutable transcripts.
### Problem to solve
Teams need to remove sensitive, resolved, or low-signal details from active context, but current behavior forces an all-or-nothing choice: keep everything (context... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-15T14:15:54Z | 2026-04-27T05:05:06Z | 2026-04-24T19:03:13Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Ar3ss12 | 150,044,748 | U_kgDOCPGATA | User | false |
openclaw/openclaw | 3,944,195,579 | I_kwDOQb6kR87rF6X7 | 17,208 | https://github.com/openclaw/openclaw/issues/17208 | https://api.github.com/repos/openclaw/openclaw/issues/17208 | Feature]: Expose active Context Quality Score with warnings before quality collapse | ### Summary
Add a native context health score (relevance, recency, criticality, noise ratio) with threshold-based warnings.
### Problem to solve
Users cannot see context quality degradation until outputs are already bad. Current behavior gives no objective signal for when to clean/compress context, causing late det... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-15T14:29:57Z | 2026-04-27T05:05:05Z | 2026-04-24T19:03:16Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Ar3ss12 | 150,044,748 | U_kgDOCPGATA | User | false |
openclaw/openclaw | 3,944,197,867 | I_kwDOQb6kR87rF67r | 17,209 | https://github.com/openclaw/openclaw/issues/17209 | https://api.github.com/repos/openclaw/openclaw/issues/17209 | [Feature]: Detect missing critical fields in summaries/checkpoints before save/use | ### Summary
Add built-in anomaly detection for summaries/checkpoints to prevent operationally dangerous omissions.
### Problem to solve
Readable summaries can still be unsafe: missing deadline, blocker without next step, no acceptance criteria. Current flows allow such summaries to pass silently, causing execution ... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-15T14:30:58Z | 2026-04-27T05:05:04Z | 2026-04-24T19:03:19Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Ar3ss12 | 150,044,748 | U_kgDOCPGATA | User | false |
openclaw/openclaw | 3,944,199,231 | I_kwDOQb6kR87rF7Q_ | 17,210 | https://github.com/openclaw/openclaw/issues/17210 | https://api.github.com/repos/openclaw/openclaw/issues/17210 | [Feature]: Ship native Domain Context Packets (health/finance/projects) with schema and merge policy | ### Summary
Add built-in domain-specific context packets to replace weak one-size-fits-all context handling.
### Problem to solve
Generic context structures miss domain constraints and required fields. Health, finance, and project workflows need different memory shape; current generic handling causes omissions, ret... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-15T14:31:48Z | 2026-04-27T05:05:04Z | 2026-04-24T19:03:21Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Ar3ss12 | 150,044,748 | U_kgDOCPGATA | User | false |
openclaw/openclaw | 3,944,319,924 | I_kwDOQb6kR87rGYu0 | 17,255 | https://github.com/openclaw/openclaw/issues/17255 | https://api.github.com/repos/openclaw/openclaw/issues/17255 | [Feature]: Implementing Agentic Safety Guardrails to prevent autonomous reputational harm (Ref: Matplotlib PR #31132) | ### Summary
I am raising a discussion on implementing safety guardrails following the incident where an OpenClaw agent autonomously published a hit piece against a maintainer (Ref: Matplotlib PR #31132). The goal is to prevent agents from using their autonomy to cause reputational harm.
### Problem to solve
Current... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-02-15T15:36:03Z | 2026-04-27T05:05:03Z | 2026-04-24T19:03:25Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | karthik-sivaraman | 101,684,098 | U_kgDOBg-Tgg | User | false |
openclaw/openclaw | 3,944,476,220 | I_kwDOQb6kR87rG-48 | 17,317 | https://github.com/openclaw/openclaw/issues/17317 | https://api.github.com/repos/openclaw/openclaw/issues/17317 | [Feature]: How to configure sessions_spawn allowlist for agent-to-agent calls? | ### Summary
## Problem Summary
I'm trying to set up a multi-agent architecture where one agent can spawn another as a subagent using `sessions_spawn`, but the target agent is not available in the allowlist.
## Current Setup
- `285host` (main agent, bound to Telegram group)
- `285coder` (coding specialist, same work... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-15T16:57:21Z | 2026-04-27T05:05:02Z | 2026-04-24T19:03:27Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | jasoncow007 | 125,945,058 | U_kgDOB4HE4g | User | false |
openclaw/openclaw | 3,944,516,568 | I_kwDOQb6kR87rHIvY | 17,335 | https://github.com/openclaw/openclaw/issues/17335 | https://api.github.com/repos/openclaw/openclaw/issues/17335 | [Feature]: Security threat detection as OTel span enrichment in diagnostics-otel | ### Summary
## Summary
Add real-time security threat detection to the `diagnostics-otel` plugin by enriching OTel spans with security event attributes. This enables observability backends (Dynatrace, Grafana, etc.) to alert on suspicious AI agent behavior without requiring external security tooling.
### Problem t... | closed | not_planned | true | 4 | [
"enhancement",
"security"
] | [] | 2026-02-15T17:18:09Z | 2026-04-27T05:05:01Z | 2026-04-24T19:03:30Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | henrikrexed | 81,800,915 | MDQ6VXNlcjgxODAwOTE1 | User | false |
openclaw/openclaw | 3,944,526,993 | I_kwDOQb6kR87rHLSR | 17,344 | https://github.com/openclaw/openclaw/issues/17344 | https://api.github.com/repos/openclaw/openclaw/issues/17344 | [Feature]: Tetragon TracingPolicies for kernel-level security monitoring of AI agents | ### Summary
Provide default Tetragon TracingPolicies that monitor OpenClaw at the kernel level, complementing the application-layer security detection in `diagnostics-otel`. Together, they form a defense-in-depth observability stack for AI agents.
### Problem to solve
Application-layer detection (span enrichment,... | closed | not_planned | true | 5 | [
"enhancement",
"security"
] | [] | 2026-02-15T17:23:51Z | 2026-04-27T05:05:01Z | 2026-04-24T19:03:34Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | henrikrexed | 81,800,915 | MDQ6VXNlcjgxODAwOTE1 | User | false |
openclaw/openclaw | 3,944,789,218 | I_kwDOQb6kR87rILTi | 17,468 | https://github.com/openclaw/openclaw/issues/17468 | https://api.github.com/repos/openclaw/openclaw/issues/17468 | [Feature]: Chat: Visibility of Tool output | ### Summary
In Chat of OpenClaw Gateway Dashboard add Toggle ON/OFF feature functionality for visibility of tool output messages in the chat conversation leaving only the actual final answer to the user request.
### Problem to solve
For non developers this is a very cluttered and confusing display. Cannot work with... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-15T19:53:45Z | 2026-04-27T05:05:00Z | 2026-04-24T19:03:37Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | asi99stats2026-pixel | 261,512,164 | U_kgDOD5Zb5A | User | false |
openclaw/openclaw | 3,945,786,818 | I_kwDOQb6kR87rL-3C | 17,734 | https://github.com/openclaw/openclaw/issues/17734 | https://api.github.com/repos/openclaw/openclaw/issues/17734 | [Feature]: Chatwork channel plugin | ### Summary
Add Chatwork as a channel plugin, enabling OpenClaw agents to send/receive messages on Japan's most widely used business chat platform.
### Problem to solve
Chatwork is one of the most widely used business chat platforms in Japan (600,000+ organizations). Currently, OpenClaw has no Chatwork integration, ... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-16T04:17:53Z | 2026-04-27T05:04:59Z | 2026-04-24T19:03:40Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | KoichiroHatsuse | 100,126,046 | U_kgDOBffNXg | User | false |
openclaw/openclaw | 3,946,833,520 | I_kwDOQb6kR87rP-Zw | 17,936 | https://github.com/openclaw/openclaw/issues/17936 | https://api.github.com/repos/openclaw/openclaw/issues/17936 | [Security] message/sendAttachment allows local file exfiltration when sandboxRoot is unset | ## Summary
`message` / `sendAttachment` can read and exfiltrate arbitrary local files when `sandboxRoot` is not set.
This is a **fail-open sandbox behavior**: path restrictions are only applied when `sandboxRoot` exists, but file-loading still proceeds in non-sandboxed sessions.
## Severity
High (prompt-injection/dat... | closed | completed | true | 3 | [
"security",
"stale"
] | [] | 2026-02-16T09:42:11Z | 2026-04-27T05:04:58Z | 2026-04-24T19:03:46Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | tdjackey | 6,791,132 | MDQ6VXNlcjY3OTExMzI= | User | false |
openclaw/openclaw | 3,947,567,827 | I_kwDOQb6kR87rSxrT | 18,047 | https://github.com/openclaw/openclaw/issues/18047 | https://api.github.com/repos/openclaw/openclaw/issues/18047 | [Feature Request]: Native MT4/MT5 Execution Skill for Forex Trading Automation | ### Summary
Request for a new skill (or core integration) that allows OpenClaw agents to parse incoming trading signals (e.g., from Telegram channels/groups) and automatically execute buy/sell orders on MetaTrader 4 or 5 (MT4/MT5) terminals. This would enable fast, autonomous signal-based trading for forex pairs, espe... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-16T12:36:39Z | 2026-04-27T05:04:58Z | 2026-04-24T19:03:49Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | adevaproject | 137,255,461 | U_kgDOCC5aJQ | User | false |
openclaw/openclaw | 3,948,583,094 | I_kwDOQb6kR87rWpi2 | 18,223 | https://github.com/openclaw/openclaw/issues/18223 | https://api.github.com/repos/openclaw/openclaw/issues/18223 | Compaction triggers session.reset which SIGKILLs in-flight exec processes (regression in 2026.2.14+) | ## Description
When context compaction triggers `session.reset()`, it now aborts all active runs, which sends SIGKILL to in-flight `exec` tool processes. This is a regression introduced by PR #16576 ("abort active runs during sessions.reset") in 2026.2.14.
Compaction is a normal, frequent operation that should be tra... | closed | completed | true | 6 | [] | [
"onutc"
] | 2026-02-16T16:32:32Z | 2026-04-27T05:04:57Z | 2026-04-24T19:03:52Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | bullderz | 258,144,590 | U_kgDOD2L5Tg | User | false |
openclaw/openclaw | 3,949,737,983 | I_kwDOQb6kR87rbDf_ | 18,599 | https://github.com/openclaw/openclaw/issues/18599 | https://api.github.com/repos/openclaw/openclaw/issues/18599 | [Feature]: Token- / Message-based History Limit per Channel | ### Summary
Add configurable hard limits (maxHistoryTokens, maxHistoryMessages) to cap conversation history growth per channel. Currently, linear history growth causes quadratic token costs,
especially problematic for WhatsApp and cost-sensitive deployments.
### Problem to solve
OpenClaw's linear history growth c... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-16T22:01:11Z | 2026-04-27T05:04:56Z | 2026-04-24T19:03:55Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | hp-8472 | 56,477,491 | MDQ6VXNlcjU2NDc3NDkx | User | false |
openclaw/openclaw | 3,949,864,665 | I_kwDOQb6kR87rbibZ | 18,633 | https://github.com/openclaw/openclaw/issues/18633 | https://api.github.com/repos/openclaw/openclaw/issues/18633 | [Feature]: Stuck run recovery: auto-abort after timeout (compaction stalls block all messages) | ### Summary
compaction stall recovery
### Problem to solve
Core ask: When any run (including compaction) exceeds a configurable timeout, auto-abort and unblock the message queue. Three incidents in 24 hours, all with the same root cause — no way to kill a stuck run.
The full issue is at issues/stuck-run-recovery.md... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-16T22:45:32Z | 2026-04-27T05:04:55Z | 2026-04-24T19:03:58Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | diggertydog | 258,656,447 | U_kgDOD2rIvw | User | false |
openclaw/openclaw | 3,950,674,248 | I_kwDOQb6kR87reoFI | 18,844 | https://github.com/openclaw/openclaw/issues/18844 | https://api.github.com/repos/openclaw/openclaw/issues/18844 | [Feature]: GLM-5 JSON error | ### Summary
Support GLM-5, may be function calling.
### Problem to solve
Expected ',' or '}' after property value in JSON at position 28(line 1 column 29)
### Proposed solution
None
### Alternatives considered
_No response_
### Impact
Affected: users
Severity: annoying
Frequency: ofter
Consequence: errors
###... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-17T04:07:44Z | 2026-04-27T05:04:55Z | 2026-04-24T19:04:12Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | hummingg | 25,954,059 | MDQ6VXNlcjI1OTU0MDU5 | User | false |
openclaw/openclaw | 3,950,780,772 | I_kwDOQb6kR87rfCFk | 18,869 | https://github.com/openclaw/openclaw/issues/18869 | https://api.github.com/repos/openclaw/openclaw/issues/18869 | [Feature]: Multi-Agent Group Chat Support with Turn-Taking Protocol | ### Summary
Multi-agent group chat support with turn-taking protocol for collaborative discussions between multiple AI agents in the same Telegram/Discord/WhatsApp group.
### Problem to solve
Currently, OpenClaw doesn't support multiple agents participating in the same group chat conversation. When two agents are... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-02-17T04:47:41Z | 2026-04-27T05:04:54Z | 2026-04-24T19:04:16Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ydstern | 251,489,083 | U_kgDODv1rOw | User | false |
openclaw/openclaw | 3,950,931,098 | I_kwDOQb6kR87rfmya | 18,920 | https://github.com/openclaw/openclaw/issues/18920 | https://api.github.com/repos/openclaw/openclaw/issues/18920 | [Feature]: Implement Proper Message Dispatching in Nostr Extension | ### Summary
The Nostr extension currently uses an unstandardized TODO hack to dispatch inbound messages using an internal handleInboundMessage method.
### Problem to solve
The Nostr extension currently uses an unstandardized TODO hack to dispatch inbound messages using an internal handleInboundMessage method.
###... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-17T05:48:26Z | 2026-04-27T05:04:53Z | 2026-04-24T19:04:18Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Ndolo7 | 106,404,855 | U_kgDOBleb9w | User | false |
openclaw/openclaw | 3,952,110,821 | I_kwDOQb6kR87rkGzl | 19,089 | https://github.com/openclaw/openclaw/issues/19089 | https://api.github.com/repos/openclaw/openclaw/issues/19089 | [Feature]: Request: Add Image Recognition / OCR Capability | ### Summary
As an OpenClaw user, I need the ability to recognize and analyze images
directly from chat messages.
Use Cases:
1. OCR - Extract text from screenshots of error messages
2. Visual Debugging - Analyze application error dialogs
3. Documentation - Read images of diagrams/schematics
Benefits:
- Faster troubl... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-17T10:47:08Z | 2026-04-27T05:04:52Z | 2026-04-24T19:04:21Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | lbt12 | 197,555,445 | U_kgDOC8Z09Q | User | false |
openclaw/openclaw | 3,954,251,093 | I_kwDOQb6kR87rsRVV | 19,396 | https://github.com/openclaw/openclaw/issues/19396 | https://api.github.com/repos/openclaw/openclaw/issues/19396 | [Feature]: Add config flag to allow Mattermost file attachments from private/local IPs | ### Summary
Allow the Mattermost channel to fetch file attachments from private IP addresses (localhost, 127.0.0.1, 192.168.x.x) via a configuration flag.
### Problem to solve
When OpenClaw is configured with a local/self-hosted Mattermost instance (e.g., http://127.0.0.1:8065 or http://192.168.x.x:8065), inbound fi... | closed | completed | true | 4 | [
"enhancement"
] | [] | 2026-02-17T19:06:59Z | 2026-04-27T05:04:52Z | 2026-04-24T19:04:25Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | marcin-kruszynski | 3,131,341 | MDQ6VXNlcjMxMzEzNDE= | User | false |
openclaw/openclaw | 3,955,671,336 | I_kwDOQb6kR87rxsEo | 19,688 | https://github.com/openclaw/openclaw/issues/19688 | https://api.github.com/repos/openclaw/openclaw/issues/19688 | [Feature]: Add stat tool | ### Summary
Add a `stat` tool to let the agent check the filesystem more easily.
### Problem to solve
The agent has read, edit, and write tools, but it has no tool to check if a file exists. Therefore, agents usually try to read the file, and if that produces an error, the file doesn't exist and they can safely wri... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-18T03:10:17Z | 2026-04-27T05:04:51Z | 2026-04-24T19:04:28Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | HenryLoenwind | 1,485,873 | MDQ6VXNlcjE0ODU4NzM= | User | false |
openclaw/openclaw | 3,955,698,773 | I_kwDOQb6kR87rxyxV | 19,700 | https://github.com/openclaw/openclaw/issues/19700 | https://api.github.com/repos/openclaw/openclaw/issues/19700 | [Feature]: Integrate powpow to enable OpenClaw agents to be placed on a real-world map | ### Summary
Enable OpenClaw agents to be placed on a real‑world map via the [powpow](https://www.powpow.online/) ([GitHub](https://github.com/durenzidu/powpow)) project, allowing location‑based interactions and expanding intelligent agents from online chats into physical spaces.
<table>
<tr>
<td><img src="https... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-18T03:21:08Z | 2026-04-27T05:04:50Z | 2026-04-24T19:04:31Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | durenzidu | 38,130,340 | MDQ6VXNlcjM4MTMwMzQw | User | false |
openclaw/openclaw | 3,955,745,121 | I_kwDOQb6kR87rx-Fh | 19,715 | https://github.com/openclaw/openclaw/issues/19715 | https://api.github.com/repos/openclaw/openclaw/issues/19715 | [Feature]: Explicitly identify the human "owner" on channels like Telegram and WhatsApp. | ### Summary
Add per-channel default "human owner" contact to avoid non autorized personals to run commands and receive unwanted notifications
### Problem to solve
The bot appears to send messages and treat the last active chat as his human owner. We need to explicity tell the bot who is the owner and only authorized... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-18T03:38:08Z | 2026-04-27T05:04:50Z | 2026-04-24T19:04:33Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | octaivermatt | 262,254,098 | U_kgDOD6GuEg | User | false |
openclaw/openclaw | 3,955,864,531 | I_kwDOQb6kR87rybPT | 19,749 | https://github.com/openclaw/openclaw/issues/19749 | https://api.github.com/repos/openclaw/openclaw/issues/19749 | [Feature]: Add Session Messenger as an optional Channel provider for OpenClaw. | ### Summary
Session offers privacy without compromise: decentralized infrastructure (no central servers), zero metadata collection (no phone/email required), built-in onion routing, and censorship-resistant design. It's open-source, rapidly growing in privacy-conscious communities, and works where other messengers ge... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-18T04:23:32Z | 2026-04-27T05:04:49Z | 2026-04-24T19:04:36Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | vassago999 | 98,677,850 | U_kgDOBeG0Wg | User | false |
openclaw/openclaw | 3,955,953,384 | I_kwDOQb6kR87ryw7o | 19,777 | https://github.com/openclaw/openclaw/issues/19777 | https://api.github.com/repos/openclaw/openclaw/issues/19777 | [Feature]: Add config option to suppress/customize tool error notifications shown to users | ### Summary
When a tool fails, OpenClaw sends a raw system notification to the user before the agent can respond. This exposes technical details to non-technical users and undermines the agent's ability to handle errors gracefully.
### Problem to solve
**User pain:** Non-technical end users see confusing, scary erro... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-18T04:58:07Z | 2026-04-27T05:04:48Z | 2026-04-24T19:04:39Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Anboias | 16,767,364 | MDQ6VXNlcjE2NzY3MzY0 | User | false |
openclaw/openclaw | 3,956,625,549 | I_kwDOQb6kR87r1VCN | 19,898 | https://github.com/openclaw/openclaw/issues/19898 | https://api.github.com/repos/openclaw/openclaw/issues/19898 | [Feature]: Support SAP AI Core as a LLM-Provider | ### Summary
Support the SAP AI Core as a LLM Provider:
More docs at https://sap.github.io/ai-sdk/docs/java/connecting-to-ai-core
### Problem to solve
Facilitate use of SAP AI Core provided LLMs for OpenClaw
### Proposed solution
Add SAP AI Core as provider option
### Alternatives considered
_No response_
###... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-18T08:35:56Z | 2026-04-27T05:04:47Z | 2026-04-24T19:04:43Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | dakoller | 659,320 | MDQ6VXNlcjY1OTMyMA== | User | false |
openclaw/openclaw | 3,956,909,389 | I_kwDOQb6kR87r2aVN | 19,936 | https://github.com/openclaw/openclaw/issues/19936 | https://api.github.com/repos/openclaw/openclaw/issues/19936 | [ENHANCEMENT]: Separate static system prompt from dynamic inbound metadata to enable provider-side caching and reduce token costs | ## Summary
OpenClaw currently injects dynamic inbound metadata, including a per-request `message_id`, into the system prompt on every turn. Because this metadata changes with each request, the overall prompt prefix differs between turns.
This prevents downstream providers from reusing prefix state such as KV cache or... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-18T09:44:48Z | 2026-04-27T05:04:46Z | 2026-04-24T19:04:45Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | PinguIntel | 172,262,384 | U_kgDOCkSD8A | User | false |
openclaw/openclaw | 3,958,878,999 | I_kwDOQb6kR87r97MX | 20,206 | https://github.com/openclaw/openclaw/issues/20206 | https://api.github.com/repos/openclaw/openclaw/issues/20206 | [Bug]: iMessage attachments blocked by path security policy after 2026.2.17 upgrade — regression from 2026.2.15 | ### Summary
After upgrading to OpenClaw 2026.2.17, iMessage attachments are no longer readable by agents despite receiving correct metadata. The agent receives attachment paths (e.g., ~/Library/Messages/Attachments/...), but reading the actual file fails with "Local media path is not under an allowed directory." This ... | closed | completed | true | 4 | [
"bug",
"security"
] | [] | 2026-02-18T17:09:06Z | 2026-04-27T05:04:46Z | 2026-04-24T19:04:54Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | p3assistant | 261,676,364 | U_kgDOD5jdTA | User | false |
openclaw/openclaw | 3,959,550,645 | I_kwDOQb6kR87sAfK1 | 20,322 | https://github.com/openclaw/openclaw/issues/20322 | https://api.github.com/repos/openclaw/openclaw/issues/20322 | [Feature]: New Chat Channel - "Obsidian" In-Markdown Chat | ### Summary
Automatic `In-Markdown Chat` feature backed by git (GitHub, possibly others)
### Problem to solve
Particularly for devs who like to build up KB and Prompts in markdown, and who keep their markdown git-backed in GitHub (in an Obsidian or Obsidian-like files and folders structure), an in-markdown chat allo... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-18T20:00:46Z | 2026-04-27T05:04:45Z | 2026-04-24T19:04:57Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | slathrop | 3,028,764 | MDQ6VXNlcjMwMjg3NjQ= | User | false |
openclaw/openclaw | 3,960,449,639 | I_kwDOQb6kR87sD6pn | 20,487 | https://github.com/openclaw/openclaw/issues/20487 | https://api.github.com/repos/openclaw/openclaw/issues/20487 | Control Tower: Add security headers to Next.js config | The cron job 'Proposal Executor' attempted to implement the proposal to add security headers to the Control Tower. However, the 'next.config.ts' or 'next.config.js' file could not be located within the workspace. This issue is being created to track the investigation and implementation of adding Next.js security header... | closed | completed | true | 5 | [
"security",
"stale"
] | [] | 2026-02-19T00:53:05Z | 2026-04-27T05:04:44Z | 2026-04-24T19:04:59Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | gitgrooves | 23,187,937 | MDQ6VXNlcjIzMTg3OTM3 | User | false |
openclaw/openclaw | 3,962,703,849 | I_kwDOQb6kR87sMg_p | 20,909 | https://github.com/openclaw/openclaw/issues/20909 | https://api.github.com/repos/openclaw/openclaw/issues/20909 | [Feature]: Doctor scan should validate that fallback models are defined in models.providers | ### Summary
Add validation to `openclaw doctor` to check that all models referenced in `fallbacks` are properly defined in `models.providers`, preventing runtime failures when fallback chains are triggered.
### Problem to solve
Currently, `openclaw doctor scan` does not validate that fallback models are defined in `... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-19T12:25:03Z | 2026-04-27T05:04:44Z | 2026-04-24T19:05:09Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | EdwardOptimization | 248,197,022 | U_kgDODssvng | User | false |
openclaw/openclaw | 3,963,727,182 | I_kwDOQb6kR87sQa1O | 21,047 | https://github.com/openclaw/openclaw/issues/21047 | https://api.github.com/repos/openclaw/openclaw/issues/21047 | [Feature]: Discord: Include author.id in message metadata for owner verification | ### Summary
Discord messages passed to the agent don't include the sender's numeric user ID in the metadata envelope. This makes it impossible for the agent to automatically verify message senders against the configured owner list in group channels.
### Problem to solve
The sender metadata block includes:
json
{
"... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-19T16:01:36Z | 2026-04-27T05:04:43Z | 2026-04-24T19:05:12Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | DennisPerrin | 86,558,705 | MDQ6VXNlcjg2NTU4NzA1 | User | false |
openclaw/openclaw | 3,964,317,175 | I_kwDOQb6kR87sSq33 | 21,133 | https://github.com/openclaw/openclaw/issues/21133 | https://api.github.com/repos/openclaw/openclaw/issues/21133 | [Feature]: Support Slack chat:write.customize for custom bot name and avatar in channel replies | ### Summary
Add support for the `chat:write.customize` Slack scope so OpenClaw can post messages using the configured assistant name and avatar instead of the default app/bot name.
### Problem to solve
OpenClaw supports configuring a custom assistant identity via `ui.assistant.name` and `ui.assistant.avatar`, but th... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-19T18:09:13Z | 2026-04-27T05:04:42Z | 2026-04-24T19:05:15Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | bernardclaw | 262,632,600 | U_kgDOD6d0mA | User | false |
openclaw/openclaw | 3,964,404,038 | I_kwDOQb6kR87sTAFG | 21,151 | https://github.com/openclaw/openclaw/issues/21151 | https://api.github.com/repos/openclaw/openclaw/issues/21151 | [Security]: SSRF in nodes-camera writeUrlToFile - no private IP blocking | ## Summary
`writeUrlToFile()` in `src/cli/nodes-camera.ts` uses bare `fetch()` without SSRF protection. A malicious or compromised node can return a `payload.url` pointing to internal/private network endpoints.
## Affected Code
```typescript
// src/cli/nodes-camera.ts:81
const res = await fetch(url);
```
Reachable ... | closed | completed | true | 3 | [
"security",
"stale"
] | [] | 2026-02-19T18:28:41Z | 2026-04-27T05:04:41Z | 2026-04-24T19:05:19Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Marvae | 11,957,602 | MDQ6VXNlcjExOTU3NjAy | User | false |
openclaw/openclaw | 3,964,438,632 | I_kwDOQb6kR87sTIho | 21,157 | https://github.com/openclaw/openclaw/issues/21157 | https://api.github.com/repos/openclaw/openclaw/issues/21157 | [Feature]: Tool Execution Model Routing Configuration | ### Summary
Currently, OpenClaw hardcodes tool execution to use Claude (Anthropic) regardless of the selected conversation model. This creates: Cost inefficiency: Tool operations always burn expensive Anthropic credits even when using free models like Groq for conversation Model mismatch: Can't align tool capabilities... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-02-19T18:35:31Z | 2026-04-27T05:04:40Z | 2026-04-24T19:05:22Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ppat07 | 229,044,480 | U_kgDODabxAA | User | false |
openclaw/openclaw | 3,964,586,087 | I_kwDOQb6kR87sTshn | 21,175 | https://github.com/openclaw/openclaw/issues/21175 | https://api.github.com/repos/openclaw/openclaw/issues/21175 | [Feature]: Title: Config override for session reset prompt (sessionResetPrompt) | ### Summary
Summary:
The BARE_SESSION_RESET_PROMPT injected on /new or /reset is hardcoded in the compiled JS with no config override. It explicitly instructs the model to skip tool use ("Do not mention internal steps, files, tools, or reasoning"), which prevents agents from reading workspace files (e.g., PROJECT.md) ... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-19T19:05:21Z | 2026-04-27T05:04:40Z | 2026-04-24T19:05:25Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | benscotthaney | 46,452,290 | MDQ6VXNlcjQ2NDUyMjkw | User | false |
openclaw/openclaw | 3,965,530,578 | I_kwDOQb6kR87sXTHS | 21,414 | https://github.com/openclaw/openclaw/issues/21414 | https://api.github.com/repos/openclaw/openclaw/issues/21414 | [Feature]: Add hook for when a rate limit error occurs. | ### Summary
When all the accounts have been exhausted and there is still a rate limit, a hook should fire allowing a script to do further processing.
Probably should make this generic to any error.
### Problem to solve
I want to be able to run a script whenever openclaw quits responding due to a rate limit.
### P... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-19T22:45:57Z | 2026-04-27T05:04:39Z | 2026-04-24T19:05:28Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | lex-the-lobster | 258,705,674 | U_kgDOD2uJCg | User | false |
openclaw/openclaw | 3,966,220,159 | I_kwDOQb6kR87sZ7d_ | 21,540 | https://github.com/openclaw/openclaw/issues/21540 | https://api.github.com/repos/openclaw/openclaw/issues/21540 | [Feature]: MiniMax Coding Plan MCP Support | ### Summary
Add support for MiniMax Coding Plan's MCP tools (`web_search` and `understand_image`) to OpenClaw, enabling native image understanding without relying on fallback models.
### Problem to solve
MiniMax's Coding Plan (including the 3x Starter plan) provides two exclusive MCP tools:
1. **`web_search`** - Pe... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-20T02:36:55Z | 2026-04-27T05:04:38Z | 2026-04-24T19:05:31Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ovstryjewski | 55,242,355 | MDQ6VXNlcjU1MjQyMzU1 | User | false |
openclaw/openclaw | 3,966,521,435 | I_kwDOQb6kR87sbFBb | 21,598 | https://github.com/openclaw/openclaw/issues/21598 | https://api.github.com/repos/openclaw/openclaw/issues/21598 | [Feature]: Feature Request: before_assistant_persist hook to prevent session corruption from oversized responses | ### Summary
Add before_assistant_persist plugin hook to intercept oversized assistant messages before they corrupt the session transcript
### Problem to solve
In multi-model setups, a large-context model (e.g. Gemini 3 Pro with 1M context) can generate responses that exceed the receiving model's context window. The... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-20T04:34:53Z | 2026-04-27T05:04:37Z | 2026-04-24T19:05:35Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | litan2000 | 30,795,727 | MDQ6VXNlcjMwNzk1NzI3 | User | false |
openclaw/openclaw | 3,966,924,788 | I_kwDOQb6kR87scnf0 | 21,656 | https://github.com/openclaw/openclaw/issues/21656 | https://api.github.com/repos/openclaw/openclaw/issues/21656 | [Security] System event format leaks via issue #20484 — enables Telegram/WhatsApp spoofing attacks | ## Summary
Issue #20484 publicly documents the exact format of OpenClaw's `System:` prefixed post-compaction audit messages — including the full message template. This makes the format trivially spoofable by any external attacker via Telegram, WhatsApp, Signal, or any other connected channel.
## The Attack Vector
1.... | closed | completed | true | 2 | [
"security",
"stale"
] | [] | 2026-02-20T06:47:57Z | 2026-04-27T05:04:37Z | 2026-04-24T19:05:38Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | minimexat | 260,134,604 | U_kgDOD4FWzA | User | false |
openclaw/openclaw | 3,967,070,798 | I_kwDOQb6kR87sdLJO | 21,691 | https://github.com/openclaw/openclaw/issues/21691 | https://api.github.com/repos/openclaw/openclaw/issues/21691 | [Feature]: Add support for Sarvam AI models | ### Summary
Add Sarvam AI as a supported LLM provider to enable high-performance Indic language support and sovereign AI models.
### Problem to solve
I can’t use Sarvam AI with OpenClaw right now because their API needs a specific header (api-subscription-key) instead of the usual "Bearer" token.
### Proposed solut... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-20T07:33:44Z | 2026-04-27T05:04:36Z | 2026-04-24T19:05:41Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | dvrajput1910 | 204,345,107 | U_kgDODC4PEw | User | false |
openclaw/openclaw | 3,967,187,660 | I_kwDOQb6kR87sdnrM | 21,704 | https://github.com/openclaw/openclaw/issues/21704 | https://api.github.com/repos/openclaw/openclaw/issues/21704 | [Feature]: Feature Request: Allow configurable file extension override (e.g., .ogg) for OpenAI TTS provider to support native Telegram Voice Notes | ### Summary
Currently, when using the `openai` TTS provider in OpenClaw, the system appears to hardcode the file extension of the outgoing audio as `.mp3`.
### Problem to solve
While this works for standard audio, it breaks native Telegram "Voice Note" functionality (waveforms, consecutive autoplay, and "Raise to ... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-20T08:08:00Z | 2026-04-27T05:04:35Z | 2026-04-24T19:05:44Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ismagarcia | 1,107,432 | MDQ6VXNlcjExMDc0MzI= | User | false |
openclaw/openclaw | 3,968,446,996 | I_kwDOQb6kR87sibIU | 21,844 | https://github.com/openclaw/openclaw/issues/21844 | https://api.github.com/repos/openclaw/openclaw/issues/21844 | [Feature]: Discord Components v2 Interaction Handling | ### Summary
Add support for receiving and responding to Discord Components v2 interactions (button clicks, select menu choices) from messages sent by agents.
### Problem to solve
Currently, agents can send rich interactive messages with buttons and select menus using Components v2, but when users click these buttons... | closed | completed | true | 3 | [
"enhancement"
] | [] | 2026-02-20T12:57:52Z | 2026-04-27T05:04:35Z | 2026-04-24T19:05:53Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | WolfyBlair | 184,862,995 | U_kgDOCwTJEw | User | false |
openclaw/openclaw | 3,968,894,947 | I_kwDOQb6kR87skIfj | 21,924 | https://github.com/openclaw/openclaw/issues/21924 | https://api.github.com/repos/openclaw/openclaw/issues/21924 | [Feishu] Add tool to get group chat members list | ### Summary
Add a Feishu tool to retrieve group chat members list using im/chat/:chat_id/members/get API
### Problem to solve
When the OpenClaw bot is added to a Feishu group chat, there is no way to know who the members are. The message metadata only shows the sender's open_id, but we cannot get a list of all membe... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-20T14:44:52Z | 2026-04-27T05:04:34Z | 2026-04-24T19:05:56Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | samp1eK0nfig | 158,068,302 | U_kgDOCWvuTg | User | false |
openclaw/openclaw | 3,969,426,044 | I_kwDOQb6kR87smKJ8 | 22,014 | https://github.com/openclaw/openclaw/issues/22014 | https://api.github.com/repos/openclaw/openclaw/issues/22014 | [Feature]: Support sending original quality images to Telegram (disable compression) | ### Summary
For a 1080x9000px poster (~2MB), it gets compressed down to 800px + quality 40, resulting in only ~200-300KB.
### Problem to solve
## Problem
When sending images via Telegram, OpenClaw aggressively compresses them resulting in very low quality (200-300KB for a 2MB original).
## Expected Behavior
Suppo... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-20T16:47:23Z | 2026-04-27T05:04:33Z | 2026-04-24T19:05:59Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | godenjan | 5,979,276 | MDQ6VXNlcjU5NzkyNzY= | User | false |
openclaw/openclaw | 3,969,646,259 | I_kwDOQb6kR87sm_6z | 22,073 | https://github.com/openclaw/openclaw/issues/22073 | https://api.github.com/repos/openclaw/openclaw/issues/22073 | [Feature]: Build an SEO-optimized advertising website for OpenClaw | ### Summary
Build an advertising/SEO landing website for OpenClaw.
### Problem to solve
OpenClaw currently lacks a dedicated website optimized for advertising and search engine rankings. This makes it difficult for new users and customers to discover and learn about OpenClaw, limiting outreach and growth.
### Propo... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-20T17:41:42Z | 2026-04-27T05:04:32Z | 2026-04-24T19:06:02Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | jdboatman1 | 69,007,178 | MDQ6VXNlcjY5MDA3MTc4 | User | false |
openclaw/openclaw | 3,970,165,877 | I_kwDOQb6kR87so-x1 | 22,155 | https://github.com/openclaw/openclaw/issues/22155 | https://api.github.com/repos/openclaw/openclaw/issues/22155 | [Feature]: Add CLI-based TTS provider support (similar to Whisper STT) | ### Summary
Enable flexible, platform-independent TTS integration by adding CLI-based provider support, allowing users to integrate local and self-hosted TTS solutions without hardcoded provider restrictions.
### Problem to solve
Currently, OpenClaw's TTS provider configuration is **hardcoded** to only three options... | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-20T20:04:54Z | 2026-04-27T05:04:30Z | 2026-04-24T19:06:05Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | finddabug | 31,019,229 | MDQ6VXNlcjMxMDE5MjI5 | User | false |
openclaw/openclaw | 3,970,760,068 | I_kwDOQb6kR87srP2E | 22,255 | https://github.com/openclaw/openclaw/issues/22255 | https://api.github.com/repos/openclaw/openclaw/issues/22255 | [Feature]: Allow insecure websocket on LAN | ### Summary
As title says, there needs to be an option to allow insecure websocket just like insecure auth on the gateway UI while on LAN.
### Problem to solve
Cannot run on LAN plaintext without adding TLS security.
### Proposed solution
Provide config to allow insecure websocket
### Alternatives considered
_No... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-20T23:16:22Z | 2026-04-27T05:04:29Z | 2026-04-24T19:06:12Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | egaudet | 1,350,777 | MDQ6VXNlcjEzNTA3Nzc= | User | false |
openclaw/openclaw | 3,971,018,840 | I_kwDOQb6kR87ssPBY | 22,328 | https://github.com/openclaw/openclaw/issues/22328 | https://api.github.com/repos/openclaw/openclaw/issues/22328 | [Feature]: autoThread support for MattermosautoThread support for Mattermostt | ### Summary
## Feature Request: autoThread for Mattermost
**Background**
The `autoThread` feature is currently available for Discord (`channels.discord.guilds.<id>.channels.<channelId>.autoThread: true`), which automatically creates a thread on the triggering message and posts the agent reply inside it — keeping the... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-21T01:50:20Z | 2026-04-27T05:04:28Z | 2026-04-24T19:06:15Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | jhezjkp | 769,374 | MDQ6VXNlcjc2OTM3NA== | User | false |
openclaw/openclaw | 3,971,685,263 | I_kwDOQb6kR87suxuP | 22,552 | https://github.com/openclaw/openclaw/issues/22552 | https://api.github.com/repos/openclaw/openclaw/issues/22552 | Security audit should flag open channel access policies (dmPolicy: open with allowFrom: '*') | ## Problem
The `openclaw security audit --deep` command does not check channel-level access policies. A user can have BlueBubbles (iMessage) or other channels configured with:
```json
"dmPolicy": "open",
"allowFrom": ["*"]
```
This means **anyone** can message the assistant and trigger actions on the host machine - ... | closed | completed | true | 2 | [
"security",
"stale"
] | [] | 2026-02-21T08:48:40Z | 2026-04-27T05:04:27Z | 2026-04-24T19:06:19Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ivan516partners | 260,819,466 | U_kgDOD4vKCg | User | false |
openclaw/openclaw | 3,971,808,023 | I_kwDOQb6kR87svPsX | 22,590 | https://github.com/openclaw/openclaw/issues/22590 | https://api.github.com/repos/openclaw/openclaw/issues/22590 | [Feature]: 添加一个手机适配页面（Add a mobile Web UI） | ### Summary
由于中国用户（Chinese users）访问即时通讯工具（IM Tools）较为困难，通过手机（mobile phone）使用 openclaw 不太方便（so difficult），给 Gateway Dashboard 添加（Add）一个手机适配 WebUI，或许是一个方案（solution）。
### Problem to solve
中国用户通过手机使用适配的即时通讯工具访问 openclaw 会存在网络问题。
### Proposed solution
给 Gateway Dashboard 添加（Add）一个手机适配 WebUI，在手机端使用浏览器使用。
### Alternativ... | closed | completed | true | 4 | [
"enhancement"
] | [] | 2026-02-21T10:31:09Z | 2026-04-27T05:04:27Z | 2026-04-24T19:06:22Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ttzc | 58,620,217 | MDQ6VXNlcjU4NjIwMjE3 | User | false |
openclaw/openclaw | 3,971,852,514 | I_kwDOQb6kR87svaji | 22,603 | https://github.com/openclaw/openclaw/issues/22603 | https://api.github.com/repos/openclaw/openclaw/issues/22603 | [Feature]: the Gemini new models from anti-gravity are not in the model section, and I can't use the Gemini 3 models as well, so pls fix the anti-gravity models and update it to the latest. | ### Summary
Basically, the main problem here is I use anti-gravity to connect with the openclaw. Recently, a new update in anti-gravity occurred, and the Gemini 3 models are not working anymore. We had to use the Gemini 3.1 models, and I want you to please update it to Gemini 3.1 because currently I can't even use any... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-21T11:08:05Z | 2026-04-27T05:04:26Z | 2026-04-24T19:06:25Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | ENITITYERROR | 86,696,952 | MDQ6VXNlcjg2Njk2OTUy | User | false |
openclaw/openclaw | 3,971,951,869 | I_kwDOQb6kR87svyz9 | 22,631 | https://github.com/openclaw/openclaw/issues/22631 | https://api.github.com/repos/openclaw/openclaw/issues/22631 | [Feature]: Make Model Configuration Provider-Aware | ### Summary
After selecting a provider, the model selection list should only show models available for that provider instead of all models, to avoid confusion.
### Problem to solve
Currently, all models are displayed regardless of the selected provider, which causes confusion and increases the risk of selecting unsu... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-21T12:32:35Z | 2026-04-27T05:04:25Z | 2026-04-24T19:06:28Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | jainilDesai | 178,654,413 | U_kgDOCqYMzQ | User | false |
openclaw/openclaw | 3,971,971,143 | I_kwDOQb6kR87sv3hH | 22,635 | https://github.com/openclaw/openclaw/issues/22635 | https://api.github.com/repos/openclaw/openclaw/issues/22635 | [Feature]: CLI: add --quiet flag or OPENCLAW_NO_TAGLINE env var to suppress startup taglines | ### Summary
Add --quiet flag or OPENCLAW_NO_TAGLINE env var to suppress CLI taglines
### Problem to solve
Every CLI command outputs a random tagline after the version string, e.g.:
"2026.2.19-2 (45d9b20) — Say "stop" and I'll stop—say "ship" and we'll both learn a lesson."
These are charming initially but become no... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-21T12:47:26Z | 2026-04-27T05:04:25Z | 2026-04-24T19:06:31Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | yellowcar1101 | 90,629,161 | MDQ6VXNlcjkwNjI5MTYx | User | false |
openclaw/openclaw | 3,972,212,473 | I_kwDOQb6kR87swyb5 | 22,724 | https://github.com/openclaw/openclaw/issues/22724 | https://api.github.com/repos/openclaw/openclaw/issues/22724 | [Feature]: Skill usage telemetry & ratings on ClawhHub | ### Summary
Auto-collect skill execution results from agents and display success rate, usage count, and platform compatibility on ClawhHub — like "大众点评" for AI skills, reviewed by AI automatically.
### Problem to solve
ClawhHub lists skills but provides no quality signal. Users have no way to know if a skill actua... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-21T15:24:47Z | 2026-04-27T05:04:24Z | 2026-04-24T19:06:35Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | hxppk | 19,534,180 | MDQ6VXNlcjE5NTM0MTgw | User | false |
openclaw/openclaw | 3,972,756,409 | I_kwDOQb6kR87sy3O5 | 22,927 | https://github.com/openclaw/openclaw/issues/22927 | https://api.github.com/repos/openclaw/openclaw/issues/22927 | [Feature]: Add Postgres/pgvector store driver for memorySearch | ### Summary
The memorySearch config already accepts store.driver as a string field (defaulting to "sqlite"), but only the SQLite implementation exists. It would be valuable to have a postgres driver option backed by pgvector, allowing users who already run Postgres infrastructure to consolidate their vector search th... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-21T20:39:41Z | 2026-04-27T05:04:23Z | 2026-04-24T19:06:38Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | almuller-dev | 199,714,710 | U_kgDOC-dnlg | User | false |
openclaw/openclaw | 3,973,210,409 | I_kwDOQb6kR87s0mEp | 23,097 | https://github.com/openclaw/openclaw/issues/23097 | https://api.github.com/repos/openclaw/openclaw/issues/23097 | [Feature]: [Discord] Agent receives history_count: 0 in Threads - can't see original message or conversation history | ### Summary
Discord Thread context not passed to Agent (history_count: 0), causing inability to see original message, conversation history, or own previous messages.
### Problem to solve
1. When Agent creates a Thread from a message, it cannot see the original message content
2. Agent cannot see its own previous mes... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-22T01:05:22Z | 2026-04-27T05:04:22Z | 2026-04-24T19:06:41Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | INGHRTY | 89,714,580 | MDQ6VXNlcjg5NzE0NTgw | User | false |
openclaw/openclaw | 3,974,220,369 | I_kwDOQb6kR87s4cpR | 23,500 | https://github.com/openclaw/openclaw/issues/23500 | https://api.github.com/repos/openclaw/openclaw/issues/23500 | [Feature]: ISSUE_WHATSAPP_PROXY | ### Summary
```markdown
## Feature Description
Add support for configuring a proxy for WhatsApp connections, similar to the existing proxy support for Telegram.
## Use Case
Users in regions where WhatsApp is blocked (e.g., China mainland) need to use a proxy to connect to WhatsApp Web. Currently, Telegram already s... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-22T11:04:26Z | 2026-04-27T05:04:21Z | 2026-04-24T19:06:47Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | SoHuDrgon | 11,847,516 | MDQ6VXNlcjExODQ3NTE2 | User | false |
openclaw/openclaw | 3,974,767,324 | I_kwDOQb6kR87s6iLc | 23,653 | https://github.com/openclaw/openclaw/issues/23653 | https://api.github.com/repos/openclaw/openclaw/issues/23653 | [Feature Request] Improving Usability and "Software-style" Experience | ### Summary
I am a user from Macau and I really appreciate this project. I would like to suggest making OpenClaw more user-friendly, transitioning it from a technical tool to a more accessible "Software product."
### Problem to solve
Currently, OpenClaw requires significant manual configuration and technical knowled... | closed | completed | true | 12 | [
"enhancement"
] | [] | 2026-02-22T15:10:20Z | 2026-04-27T05:04:20Z | 2026-04-24T19:06:50Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | stanley-1028 | 222,644,232 | U_kgDODUVICA | User | false |
openclaw/openclaw | 3,975,609,405 | I_kwDOQb6kR87s9vw9 | 23,973 | https://github.com/openclaw/openclaw/issues/23973 | https://api.github.com/repos/openclaw/openclaw/issues/23973 | [Feature]: Morning Cake Recipe Generator | ### Summary
Add a "Morning Cake Recipe Generator" skill that automatically surfaces a personalized, AI-generated cake recipe card on the Live Canvas when OpenClaw launches in the morning (6AM–10AM local time), implemented as a TypeScript-first plugin with configurable preferences.
### Problem to solve
Users who enjo... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-22T22:49:56Z | 2026-04-27T05:04:19Z | 2026-04-24T19:06:53Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | kalibrezune11-glitch | 263,222,698 | U_kgDOD7B1qg | User | false |
openclaw/openclaw | 3,975,967,476 | I_kwDOQb6kR87s_HL0 | 24,076 | https://github.com/openclaw/openclaw/issues/24076 | https://api.github.com/repos/openclaw/openclaw/issues/24076 | [Feature]: mobile nodes (ios/android) should expost contacts book | ### Summary
access to call logs and contact book from the ios/android nodes
### Problem to solve
canvas, camera and audio are a great start
next thing I can think of is that ios/android nodes should provide contact book access
few things openclaw will benefit from
- reading call logs, so it can summarise from who... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-23T02:09:55Z | 2026-04-27T04:30:13Z | 2026-04-24T19:06:57Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | championswimmer | 1,327,050 | MDQ6VXNlcjEzMjcwNTA= | User | false |
openclaw/openclaw | 3,976,624,113 | I_kwDOQb6kR87tBnfx | 24,240 | https://github.com/openclaw/openclaw/issues/24240 | https://api.github.com/repos/openclaw/openclaw/issues/24240 | [Feature]: append new content to the old one instead of replace it | ### Summary
In the current version, if there are multiple reply to the same content, the new one will replace the old one. Can we append the new one, or use another way to save the old one so that if the user want to take a look, they still can find it?
### Problem to solve
The new reply will replace the old one
##... | closed | completed | true | 2 | [
"enhancement"
] | [] | 2026-02-23T06:47:22Z | 2026-04-27T05:04:19Z | 2026-04-24T19:07:05Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | TingweiMeng | 69,484,268 | MDQ6VXNlcjY5NDg0MjY4 | User | false |
openclaw/openclaw | 3,978,539,944 | I_kwDOQb6kR87tI7Oo | 24,553 | https://github.com/openclaw/openclaw/issues/24553 | https://api.github.com/repos/openclaw/openclaw/issues/24553 | [Feature]: UX: Improve edit tool error handling and param requirements | ### Summary
The `edit` tool frequently fails when called without the `newText` parameter, resulting in a parameter validation error. This suggests UX friction that could be improved.
## Current Behavior
When the `edit` tool is called without the required `newText` parameter:
- Fails immediately with error: "required... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-23T14:42:45Z | 2026-04-27T05:04:18Z | 2026-04-24T19:07:14Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | apoapostolov | 4,083,812 | MDQ6VXNlcjQwODM4MTI= | User | false |
openclaw/openclaw | 3,978,784,657 | I_kwDOQb6kR87tJ2-R | 24,585 | https://github.com/openclaw/openclaw/issues/24585 | https://api.github.com/repos/openclaw/openclaw/issues/24585 | [Feature]: Expose `isHeartbeat` flag in `PluginHookAgentContext` for plugin hooks | ### Summary
Plugin hooks (`before_agent_start`, `agent_end`, etc.) currently receive agent context via `PluginHookAgentContext`, but this context does not include the `isHeartbeat` flag. This prevents plugins from distinguishing between heartbeat runs and user-initiated interactions, leading to unnecessary API calls a... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-23T15:31:26Z | 2026-04-27T05:04:17Z | 2026-04-24T19:07:17Z | CONTRIBUTOR | null | 20260427T090020Z | 2026-04-27T09:00:20Z | rbutera | 6,047,293 | MDQ6VXNlcjYwNDcyOTM= | User | false |
openclaw/openclaw | 3,979,152,998 | I_kwDOQb6kR87tLQ5m | 24,633 | https://github.com/openclaw/openclaw/issues/24633 | https://api.github.com/repos/openclaw/openclaw/issues/24633 | [Feature]: allowBots config for Telegram provider | ### Summary
Add an allowBots configuration option for the Telegram channel provider, matching existing Slack and Google Chat support.
### Problem to solve
Clawdbot's Telegram provider silently drops all messages from other bots (is_bot: true). This is not configurable. Slack and Google Chat both have an allowBots... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-23T16:49:03Z | 2026-04-27T05:04:16Z | 2026-04-24T19:07:21Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | metacrafttech | 212,340,613 | U_kgDODKgPhQ | User | false |
openclaw/openclaw | 3,979,165,206 | I_kwDOQb6kR87tLT4W | 24,636 | https://github.com/openclaw/openclaw/issues/24636 | https://api.github.com/repos/openclaw/openclaw/issues/24636 | [Feature]: per-session usage tracking for cron and automation jobs | ### Summary
Expose token usage metrics per cron/automation session so operators can monitor cost and optimise job efficiency.
### Problem to solve
When cron jobs fire, there's no visibility into how many tokens each job consumes. Operators running multiple scheduled jobs (heartbeats, monitors, refinement pipelines)... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-23T16:51:50Z | 2026-04-27T05:04:15Z | 2026-04-24T19:07:23Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | metacrafttech | 212,340,613 | U_kgDODKgPhQ | User | false |
openclaw/openclaw | 3,979,427,418 | I_kwDOQb6kR87tMT5a | 24,674 | https://github.com/openclaw/openclaw/issues/24674 | https://api.github.com/repos/openclaw/openclaw/issues/24674 | [Feature]: Multimodal - llama-cpp-python | ### Summary
Unleash the full potential of multimodal models
### Problem to solve
，
### Proposed solution
，
### Alternatives considered
，
### Impact
，
### Evidence/examples
，
### Additional information
， | closed | not_planned | true | 2 | [
"enhancement"
] | [] | 2026-02-23T17:53:40Z | 2026-04-27T05:04:14Z | 2026-04-24T19:07:30Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | hemin5926 | 171,674,340 | U_kgDOCjuK5A | User | false |
openclaw/openclaw | 3,981,016,807 | I_kwDOQb6kR87tSX7n | 24,940 | https://github.com/openclaw/openclaw/issues/24940 | https://api.github.com/repos/openclaw/openclaw/issues/24940 | [Feature]: iMessage - clean up self-send echo artifacts | ### Summary
iMessage self-chat conversations (texting same Apple ID from iMessage) shows duplicate messages for both user and bot turns, making the channel hard to use as a natural conversation interface.
### Problem to solve
When OpenClaw is configured with a single Apple ID, every message appears twice: once when ... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-24T01:43:29Z | 2026-04-27T05:04:13Z | 2026-04-24T19:07:33Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | amahjoor | 74,957,929 | MDQ6VXNlcjc0OTU3OTI5 | User | false |
openclaw/openclaw | 3,982,359,070 | I_kwDOQb6kR87tXfoe | 25,237 | https://github.com/openclaw/openclaw/issues/25237 | https://api.github.com/repos/openclaw/openclaw/issues/25237 | [Feature]: macOS Accessibility API integration for native app automation | ### Summary
Add macOS Accessibility API support to enable AI control of native Mac applications (Calendar, Notes,
Reminders, WeChat, etc.) beyond shell commands.
### Problem to solve
Current macOS node capabilities are limited to:
- `system.run` - shell commands only
- `system.notify` - notifications
- `canv... | closed | not_planned | true | 3 | [
"enhancement"
] | [] | 2026-02-24T08:51:51Z | 2026-04-27T05:04:12Z | 2026-04-24T19:07:44Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | liujingping | 22,542,058 | MDQ6VXNlcjIyNTQyMDU4 | User | false |
openclaw/openclaw | 3,982,532,482 | I_kwDOQb6kR87tYJ-C | 25,252 | https://github.com/openclaw/openclaw/issues/25252 | https://api.github.com/repos/openclaw/openclaw/issues/25252 | [Feature]: adding Universal Tool Calling Protocol | ### Summary
Hi everyone, Recently I came across openclaw and I've seen that there is MCP support. How about using UTCP as well - no wrapper tax.
https://github.com/universal-tool-calling-protocol
### Problem to solve
.
### Proposed solution
.
### Alternatives considered
_No response_
### Impact
.
### Evidenc... | closed | not_planned | true | 1 | [
"enhancement"
] | [] | 2026-02-24T09:31:40Z | 2026-04-27T05:04:12Z | 2026-04-24T19:07:47Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | Raezil | 170,965,471 | U_kgDOCjC53w | User | false |
openclaw/openclaw | 3,983,259,423 | I_kwDOQb6kR87ta7cf | 25,361 | https://github.com/openclaw/openclaw/issues/25361 | https://api.github.com/repos/openclaw/openclaw/issues/25361 | [Feature] Add config flag to disable interleaved thinking mode | ### Summary
Currently, agents using Sonnet with `<thinking_mode>interleaved</thinking_mode>` emit thinking blocks automatically between tool calls, regardless of `/thinking` or `/reasoning` settings.
### Problem to solve
- Thinking blocks clutter conversation history, especially in messaging surfaces (Telegram, Disc... | closed | completed | true | 1 | [
"enhancement"
] | [] | 2026-02-24T11:53:10Z | 2026-04-27T05:04:11Z | 2026-04-24T19:07:49Z | NONE | null | 20260427T090020Z | 2026-04-27T09:00:20Z | borishrncic | 16,835,029 | MDQ6VXNlcjE2ODM1MDI5 | User | false |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.