Agent Memory Systems
Design short-term, long-term, and typed memory with chunking and retrieval so agents stay consistent across sessions.
Overview
Agent-memory-systems is an agent skill most often used in Build (also Ship review and Grow lifecycle) that architects short- and long-term agent memory with retrieval-first chunking and vector-store patterns.
Install
npx skills add https://github.com/davila7/claude-code-templates --skill agent-memory-systemsWhat is this skill?
- Covers short-term (context window), long-term (vector stores), and cognitive memory organization
- Emphasizes retrieval over raw storage: chunking, embeddings, and search quality
- Memory type architecture and vector store selection patterns
- Capability map includes episodic, semantic, procedural, working memory, and memory decay
- Frames agent "forgetting" as retrieval failures, not intelligence limits
- 10 named memory capabilities in skill frontmatter (agent-memory through memory-decay)
Adoption & trust: 605 installs on skills.sh; 27.8k GitHub stars; 3/3 security scanners passed (skills.sh audits).
What problem does it solve?
Your agent stores transcripts and embeddings but still gives wrong or repetitive answers because nothing retrieves the right memory at the right moment.
Who is it for?
Builders implementing persistent agent memory, RAG, or session continuity who need explicit retrieval architecture—not a single prompt tweak.
Skip if: One-shot codegen with no persistence, or products that only need a static FAQ page without agent state.
When should I use this skill?
Designing or debugging agent memory, vector retrieval, chunking, and long-term persistence for intelligent agents.
What do I get? / Deliverables
You align memory types, chunking, embeddings, and vector store choice so agents recall context reliably instead of treating every chat as a cold start.
- Memory type map for your agent
- Vector store selection rationale
- Chunking and retrieval strategy notes
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Memory architecture is canonical on Build agent-tooling when you implement persistence and retrieval for Claude Code-style agents. Agent-tooling is where vector stores, episodic vs semantic memory, and retrieval strategy decisions belong—not generic app CRUD alone.
Where it fits
Choose episodic vs semantic stores before wiring your agent’s MCP or custom memory plugin.
Evaluate vector DB APIs and embedding pipelines when connecting a knowledge base to nightly ingest.
Audit whether inconsistent answers trace to chunk size or missing retrieval filters before launch.
Add decay and summarization so returning users get relevant history without exceeding context limits.
How it compares
Architecture and retrieval design for agent memory—not a turnkey hosted memory SaaS integration list.
Common Questions / FAQ
Who is agent-memory-systems for?
Solo developers building Claude Code, Cursor, or Codex agents that must remember users, procedures, and facts across sessions using vectors and structured memory types.
When should I use agent-memory-systems?
During Build when designing agent-tooling and stores; during Ship review when auditing retrieval quality; during Grow lifecycle when scaling personalized history without context overflow.
Is agent-memory-systems safe to install?
Use the Security Audits panel on this Prism page to assess repo risk; memory designs often touch PII—review data retention and secrets separately from the skill text.
SKILL.md
READMESKILL.md - Agent Memory Systems
# Agent Memory Systems You are a cognitive architect who understands that memory makes agents intelligent. You've built memory systems for agents handling millions of interactions. You know that the hard part isn't storing - it's retrieving the right memory at the right time. Your core insight: Memory failures look like intelligence failures. When an agent "forgets" or gives inconsistent answers, it's almost always a retrieval problem, not a storage problem. You obsess over chunking strategies, embedding quality, and ## Capabilities - agent-memory - long-term-memory - short-term-memory - working-memory - episodic-memory - semantic-memory - procedural-memory - memory-retrieval - memory-formation - memory-decay ## Patterns ### Memory Type Architecture Choosing the right memory type for different information ### Vector Store Selection Pattern Choosing the right vector database for your use case ### Chunking Strategy Pattern Breaking documents into retrievable chunks ## Anti-Patterns ### ❌ Store Everything Forever ### ❌ Chunk Without Testing Retrieval ### ❌ Single Memory Type for All Data ## ⚠️ Sharp Edges | Issue | Severity | Solution | |-------|----------|----------| | Issue | critical | ## Contextual Chunking (Anthropic's approach) | | Issue | high | ## Test different sizes | | Issue | high | ## Always filter by metadata first | | Issue | high | ## Add temporal scoring | | Issue | medium | ## Detect conflicts on storage | | Issue | medium | ## Budget tokens for different memory types | | Issue | medium | ## Track embedding model in metadata | ## Related Skills Works well with: `autonomous-agents`, `multi-agent-orchestration`, `llm-architect`, `agent-tool-builder`