stm-design

Design git-friendly multi-user Short-Term Memory systems with session isolation and recovery patterns.

2|2|Updated Jan 20, 2026
One-click install
npx skills add https://github.com/srulyt/srulys-agent-packs --skill stm-design
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: stm-design
Source: https://github.com/srulyt/srulys-agent-packs/tree/main/.roo/skills/stm-design
Command: npx skills add https://github.com/srulyt/srulys-agent-packs --skill stm-design

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This Skill provides comprehensive guidance on designing and implementing Short-Term Memory (STM) systems that are git-friendly, support multi-user workflows, and are recovery-capable, preventing common pitfalls like merge conflicts and data loss.

Core Features & Use Cases

  • STM Fundamentals: Understand the difference between STM and LTM, and when to use STM.
  • Git-Friendly Patterns: Learn to isolate state in session directories to minimize merge conflicts.
  • Multi-User Concurrency: Implement strategies for handling multiple users accessing state simultaneously.
  • Schema Design: Design efficient and maintainable state schemas.
  • Recovery Patterns: Build systems that can gracefully recover from interruptions.
  • Use Case: When developing a multi-agent system that needs to maintain state across complex, multi-step operations, this skill helps you design a robust and collaborative state management system that integrates seamlessly with version control.

Quick Start

Load this skill when designing state management for multi-agent workflows, implementing session isolation patterns, or creating recovery-capable agent systems.

Frequently Asked Questions about stm-design

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I design git-friendly state management for multi-agent workflows?

Git-friendly state management for multi-agent workflows isolates state in session directories to minimize merge conflicts. This approach structures data so multiple agents can operate concurrently without overwriting shared history during version control operations.

What is Short-Term Memory (STM) in agent systems and when do I need it?

Short-Term Memory (STM) in agent systems maintains temporary state across complex, multi-step operations. You need STM when your multi-agent workflow requires session isolation and recovery capabilities without committing transient operational data to Long-Term Memory.

How do I handle multi-user concurrency in agent state schemas?

Multi-user concurrency in agent state schemas requires specific isolation patterns to handle simultaneous access. By designing schemas that separate concurrent sessions into distinct directories, you prevent data loss and race conditions during multi-user operations.

What's the best way to build recovery-capable agent systems that survive interruptions?

Recovery-capable agent systems require dedicated recovery patterns within the state management design. By implementing structured directory layouts and schema mechanisms, the system can gracefully resume multi-step operations and restore state after unexpected interruptions.

Why does my multi-agent state management cause merge conflicts in git?

Multi-agent state management causes git merge conflicts when state is not properly isolated in session directories. Without git-friendly patterns, concurrent agents writing to shared state files create overlapping changes that version control cannot resolve automatically.