Pre-submission checklist | 提交前检查

• I have searched existing issues and this hasn't been mentioned before | 我已搜索现有问题，确认此问题尚未被提及
• I have read the project documentation and confirmed this issue doesn't already exist | 我已阅读项目文档并确认此问题尚未存在
• This issue is specific to MemOS and not a general software issue | 该问题是针对 MemOS 的，而不是一般软件问题

Problem Statement | 问题陈述

• I have searched existing issues and this hasn't been mentioned before | 我已搜索现有问题，确认此问题尚未被提及
• I have read the project documentation and confirmed this issue doesn't already exist | 我已阅读项目文档并确认此问题尚未存在
• This issue is specific to MemOS and not a general software issue | 该问题是针对 MemOS 的，而不是一般软件问题

Problem Statement | 问题陈述

memos-local-openclaw currently lacks a complete and robust solution for multi-OpenClaw collaboration, especially when multiple OpenClaw instances are running on the same machine and need to share memories, tasks, and skills through a Hub/Client model.

In practice, several problems appear in this scenario:

• Viewer / Hub ports may conflict across multiple local OpenClaw instances
• agent identity, database, workspace, and session boundaries may not be isolated clearly enough
• client join / pending approval / approval / leave / disable / role-switch flows can become inconsistent
• admin actions such as promote / demote / remove member need clearer safeguards and notifications
• Hub shutdown / restart and client reconnect flows need to be more robust
• sharing-related notifications are incomplete or unclear in some cases
• documentation and landing pages do not fully explain the multi-OpenClaw collaboration architecture and deployment model

As MemOS is increasingly used in local multi-agent and multi-instance workflows, we need first-class support for multi-OpenClaw sharing rather than a single-instance-only experience.

Proposed Feature / Expected Outcome

Add a robust multi-OpenClaw team sharing solution for memos-local-openclaw, including:

• clear Hub / Client collaboration architecture
• dual-instance and multi-instance isolation on the same machine
• reliable state transitions for join / approve / reject / leave / disable / switch role
• conflict-free Viewer / Hub startup behavior
• isolated agent ID / database / session / workspace behavior across instances
• complete admin workflow for approval, member management, and role changes
• clear notifications for sharing, unsharing, removal, role changes, Hub online/offline events
• improved documentation and landing-page guidance for team sharing and multi-instance deployment

Acceptance Criteria

• multiple OpenClaw instances can run on the same machine without Viewer / Hub port conflicts
• each instance keeps its own identity, workspace, session, and storage boundaries
• Hub / Client state remains consistent during pending approval, approval, leave, disable, and role-switch scenarios
• admin operations are guarded and user-visible, including self-removal prevention and role-change notifications
• client instances can handle Hub shutdown / recovery gracefully
• docs and website clearly explain multi-OpenClaw sharing architecture, setup steps, and operational behavior

Additional Context

This feature is intended to support a practical local collaboration model:

• one OpenClaw instance can act as a Hub
• multiple other OpenClaw instances can join as Clients
• private data remains local unless explicitly shared
• shared memories, tasks, and skills can circulate across instances in a controlled way

This would make memos-local-openclaw much more useful for real-world team workflows, personal/work dual-instance setups, and multi-agent collaboration scenarios.

Willingness to Implement | 实现意愿

• I'm willing to implement this myself | 我愿意自己解决

Willingness to Implement | 实现意愿

• I'm willing to implement this myself | 我愿意自己解决
• I would like someone else to implement this | 我希望其他人来解决