skills.rest
RohanRaoCs
RohanRaoCsCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

memory-design

Analyzes a codebase and produces a justified AI memory architecture plan.

Updated May 16, 2026
Search Tags:#embeddings#rag#vector database#ai agents#codebase analysis#mem0#memory architecture
One-click install
npx skills add https://github.com/RohanRaoCs/memory-architect --skill memory-design-rohanraocs
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: memory-design
Source: https://github.com/RohanRaoCs/memory-architect/tree/main/skills/memory-design
Command: npx skills add https://github.com/RohanRaoCs/memory-architect --skill memory-design-rohanraocs

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Choosing the right memory stack for an AI application is hard: frameworks like Mem0, Zep, and Letta change monthly, pricing pages shift, and generic advice ignores your actual codebase. This Skill reads your code, derives memory requirements from real evidence, researches the current tooling landscape with live web search, and produces a tailored architecture document. ## Core Features & Use Cases - Deep Code Comprehension: Reads your codebase to map LLM call sites, user identity handling, existing storage, and any partial or flawed memory implementations, with file:line references. - Clarification and Research Phases: Asks targeted multiple-choice questions about scale, budget, and compliance, then runs live web research on framework health, pricing, and production maturity. - Approval-Gated Architecture Document: Presents a proposed stack table with rejected alternatives and assumptions for your approval, then writes a concrete implementation plan to memory-design/architecture.md. - Use Case: You are building a multi-user support chatbot and need persistent per-user memory. Run the Skill to get a justified recommendation across framework, vector database, embedding model, graph layer, and storage, plus a step-by-step integration plan. ## Quick Start Type /memory-design in your AI assistant and let it analyze your codebase to propose a memory architecture.

Frequently Asked Questions about memory-design

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

FAQPage Schema
How do I design a memory architecture for my AI app?▼

Run the memory-design Skill in your AI assistant. It reads your codebase, asks multiple-choice questions about scale, budget, and compliance, researches current frameworks and vector databases via live web search, then writes an approved architecture plan to memory-design/architecture.md.

Mem0 vs Zep vs Letta: which memory framework should I use?▼

The Skill evaluates Mem0, Zep, Letta, Cognee, Hindsight, and LangMem against your stack, scale, and budget using current GitHub activity and pricing. It recommends one framework per axis with justification rather than listing alternatives.

Does this work with an existing partial memory implementation?▼

Yes. Phase 1 classifies your codebase as greenfield, partial, or complete-but-flawed, audits existing memory components with file:line references, and recommends whether to keep, replace, or augment each piece, including data migration considerations.

Can I get a memory stack recommendation with a zero dollar budget?▼

Yes. A $0 budget is interpreted as $0 for memory infrastructure only, excluding LLM and embedding API costs. The Skill researches current free tiers and self-hosted options like pgvector, Chroma, and LanceDB and picks on capability.

What does the memory architecture document contain?▼

The document includes the final stack table with costs, a current-state audit, prerequisites, setup steps with package and environment variable names, read and write path wiring at specific file:line locations, configuration knobs, an integration checklist, tests, and a cost projection.

Related Skills

Explore More Software Engineering Skills→
pluginagentmarketplacepluginagentmarketplace
2

model-deployment

Deploy LLMs to production with vLLM, TGI, and Ollama configurations.

Community
Advanced
PostHogPostHog
58

instrument-llm-analytics

Instrument LLM calls to emit $ai_generation events to PostHog.

Official
Intermediate
davila7davila7
30.5k

llm-app-patterns

Implements RAG pipelines, agent architectures, prompt management, and LLMOps monitoring patterns.

Community
Advanced
yunseo-kimyunseo-kim
1

llm-memory-patterns

Implement LLM memory patterns using mem0, Qdrant, and Ollama.

Community
Advanced