claude-memory-architect

Name: claude-memory-architect
Availability: InStock
Author: m2ai-portfolio

Interview users and audit memory repos to design Claude Code memory specs.

4|1|Updated Apr 11, 2026

Search Tags:#claude #retrieval #memory #session-context #ai-memory #memory-spec

Ready to run

npx skills add https://github.com/m2ai-portfolio/m2ai-skills-pack --skill claude-memory-architect

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: claude-memory-architect
Source: https://github.com/m2ai-portfolio/m2ai-skills-pack/tree/main/skills/claude-memory-architect
Command: npx skills add https://github.com/m2ai-portfolio/m2ai-skills-pack --skill claude-memory-architect

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill helps you design a personalized memory system for Claude Code that captures what matters, preserves useful context across sessions, and avoids unwanted carryover, enabling more coherent interactions without manual re-entry.

Core Features & Use Cases

Interview-driven design process that yields a tailored memory spec using building blocks like Decay, Promotion, Salience, Disclosure, Multi-Signal Retrieval, and Compaction.
Supports deployment via CLAUDE.md entries, hooks (SessionStart), or agent-scoped memory to maintain separation and resilience.
Suitable for developers and knowledge workers who want consistent context, reduced repetitive explanations, and safer long-term memory management across projects.

Quick Start

Answer the seven interview questions to begin crafting your personalized memory spec.

Frequently Asked Questions about claude-memory-architect

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

FAQPage Schema

How do I design a persistent memory system for Claude Code sessions?▼

To design a persistent memory system for Claude Code, you complete an interview phase and repo audit to build a tailored memory spec using building blocks like decay, promotion, salience, and compaction. This spec is deployed via CLAUDE.md entries or hooks to preserve useful context across sessions.

What is memory compaction and decay in AI session context?▼

Memory compaction and decay in AI session context are building blocks for managing information lifecycle. Decay reduces the prominence of older context over time, while compaction summarizes and condenses memory to retain essential details without unwanted carryover.

How do I prevent unwanted context carryover when managing Claude memory?▼

You prevent unwanted context carryover in Claude memory by enforcing explicit TTLs, retention rules, and disclosure protocols within a tailored memory spec. These constraints ensure only safe and useful context is promoted across sessions.

Can I use SessionStart hooks to deploy a Claude memory spec?▼

Yes, you can deploy a Claude memory spec using SessionStart hooks. The design workflow supports injection strategies that output memory rules via hooks, CLAUDE.md entries, or agent-scoped files to maintain separation and resilience.

What's the best way to implement multi-signal retrieval for AI memory?▼

The best way to implement multi-signal retrieval for AI memory is to define explicit retrieval signals within a personalized memory spec. The design process audits open-source memory repos to inventory patterns and build a tailored retrieval strategy.

Do I need agent-scoped files to separate Claude memory across projects?▼

You need agent-scoped files to separate Claude memory across projects when you require strict context isolation. The memory spec design process supports agent-scoped memory deployment to maintain separation and resilience for distinct project workflows.