agent-memory-systems

Memory is crucial for intelligent agents. This skill covers agent memory architecture: short-term (context window), long-term (vector stores), and cognitive architectures. The key insight is that memory isn't just storage; it's retrieval. A million stored facts are useless if you can't find the right one.

Chunking, embedding, and retrieval strategies determine whether your agent remembers or forgets. Focus on retrieval to avoid intelligence failures and inconsistent answers. Obsess over chunking, embedding quality, and retrieval methods for optimal memory performance.

This skill provides guidance on memory type architecture, vector store selection, and chunking strategies. It also highlights anti-patterns to avoid, such as storing everything forever or chunking without testing retrieval.

What It Does

Provides the architecture and strategies for implementing and optimizing agent memory, including short-term, long-term, and working memory. Focuses on efficient retrieval of information.

When To Use

When building or optimizing AI agents that require memory capabilities for tasks such as maintaining context, recalling past interactions, or storing and retrieving knowledge.

Installation

Copy SKILL.md to your skills directory