memory-management

Orchestrates a six-tier memory system across agent sessions via boot, learning, and seal phases.

5|3|Updated Feb 21, 2026
One-click install
npx skills add https://github.com/richfrem/agent-plugins-skills --skill memory-management-richfrem
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: memory-management
Source: https://github.com/richfrem/agent-plugins-skills/tree/main/plugins/memory-management/skills/memory-management
Command: npx skills add https://github.com/richfrem/agent-plugins-skills --skill memory-management-richfrem

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes assets (resource) and references (resource) components.

What problem does it solve?

Tiered memory system that preserves contextual continuity across agent sessions, balancing hot cache, deep storage, and external persistence to avoid losing context between runs.

Core Features & Use Cases

  • Six-tier architecture (HOT CACHE, RLM SUMMARY LEDGER, VECTOR STORE, DEEP STORAGE, VAULT, SOUL) for flexible persistence and on-demand loading.
  • Boot, learning, and seal workflows to load critical context, promote/demote knowledge, and record traces across sessions.
  • Use cases include initializing a session with retained context, deciding what to remember, and promoting/demoting knowledge across tiers while maintaining isolation per project.

Quick Start

Boot the memory-management skill to initialize tiered memory storage and begin preserving context across sessions.

Frequently Asked Questions about memory-management

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

FAQPage Schema
How do I maintain cross-session context for an AI agent?

A tiered memory system preserves contextual continuity across agent sessions by orchestrating a six-tier architecture that balances hot cache and deep storage. It loads critical context on startup and records traces to vector stores when a session seals.

What is tiered memory storage for persistent agent context?

Tiered memory storage is a six-tier architecture—comprising hot cache, summary ledger, vector store, deep storage, vault, and soul—that balances hot cache and deep storage. It promotes and demotes knowledge across these tiers during learning and seal phases to avoid losing context.

How do I persist vector store memory traces between agent runs?

To persist vector store memory traces, initiate the seal phase at the end of a session. This workflow records traces to deep storage, vector stores, and vaults as needed, ensuring contextual continuity is preserved for the next boot phase.

Can I configure tiered memory storage for individual project isolation?

Yes, the tiered memory system is configurable per project to maintain isolation. Configuration options include primer, boot digest, boot contract, snapshot paths, domain directories, data directories, and vault integration.

When should I promote or demote knowledge across memory tiers?

You should promote or demote knowledge across memory tiers during the learning phase. This process evaluates hot cache and deep storage levels to decide what to remember, ensuring critical context is elevated while outdated data is archived.