Skip to content

Milestones

List view

  • Interactive co-browsing system with shared browser window, instruction panel, and multi-agent tab locking. Enables agents to hand off browser interaction to users when human input is required (OTP, login, CAPTCHA, etc.).

    No due date
    0/6 issues closed
  • Asynchronous CLI command execution tool for agents. Process lifecycle management, output buffering, streaming, and blocked command safety. Prevents agent paralysis from blocking CLI calls.

    No due date
    0/4 issues closed
  • Enhanced agent-to-user notification system with urgency levels, custom sounds, message support, and notification history tracking. Enables AI agents to effectively get user attention when needed.

    No due date
    0/4 issues closed
  • Introduce a structured, versioned, and skill-aware knowledge base system. This includes: - controlled document ingestion (scraping) - versioned documentation storage - skill-linked knowledge retrieval - token-efficient content usage Goal: Improve agent accuracy while maintaining strict control over token usage and memory efficiency.

    No due date
    0/7 issues closed
  • Released: April 2026 (planned) ## Highlights - Reduced token usage across MCP interactions - Optimized memory retrieval and compression - Improved prompt construction efficiency - Reduced redundant context injection - Enhanced multi-agent token sharing ## System Impact - Lower token cost per request - Faster response cycles - Improved scalability for large projects - Maintained reasoning quality with reduced context size ## Focus Areas - memory retrieval optimization - prompt modularization - code context reduction - history compression - tool usage efficiency --- This release focuses on improving system efficiency without reducing intelligence. See full changelog: https://github.com/coderooz/Local-MCP-Memory-Server/blob/main/CHANGELOG.md

    Overdue by 2 month(s)
    Due by April 30, 2026
    0/12 issues closed