memory-management

Decode shorthand and acronyms into a two-tier memory system with CLAUDE.md and memory/ directories.

14|3|Updated Jan 19, 2026

Search Tags:#glossary #acronyms #knowledge-base #memory-management #hot-cache #nicknames

Ready to run

npx skills add https://github.com/kevinlin/cowork-z --skill memory-management-kevinlin

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: memory-management
Source: https://github.com/kevinlin/cowork-z/tree/main/src-tauri/resources/skill-templates/productivity-memory-management
Command: npx skills add https://github.com/kevinlin/cowork-z --skill memory-management-kevinlin

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Memory-management decodes shorthand, acronyms, nicknames, and internal language so Claude can understand requests like a colleague would.

Core Features & Use Cases

Two-tier memory system: CLAUDE.md hot cache and memory/ deep memory
Memory of people, projects, and context to accelerate responses
Progressive memory lookup as needed (CLAUDE.md -> memory/glossary.md -> memory/people/projects)

Quick Start

Create CLAUDE.md hot-cache and memory/ glossary and profiles to enable immediate decoding.

Frequently Asked Questions about memory-management

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

FAQPage Schema

What is a two-tier memory system for managing project context?▼

A two-tier memory system organizes project context using a CLAUDE.md hot cache for immediate retrieval and a memory/ directory for long-term knowledge. This structure supports progressive memory lookup from hot cache to deep memory stores.

How do I set up a knowledge base for Claude to remember people and projects?▼

You set up a knowledge base by creating a CLAUDE.md hot cache and structured memory/ subdirectories for people, projects, and context. This enables progressive memory lookup to accelerate Claude's responses using stored profiles and glossaries.

Can I use a hot cache and deep memory store to scale project context retrieval?▼

Yes, you can scale project context retrieval by using CLAUDE.md as a hot cache and memory/ as a deep memory store. Structured subdirectories like memory/people and memory/projects support scalable decoding and retrieval of internal language.

Why does Claude forget project acronyms and internal shorthand between sessions?▼

Claude forgets project acronyms and internal shorthand without a persistent memory architecture. Implementing a CLAUDE.md hot cache and a memory/ glossary provides progressive lookup, ensuring internal language is decoded consistently across sessions.