StrategistKitStrategistKit — home
Get the free skill →

Use case

Build AI agent, RAG & LLM systems

Building reliable AI systems is its own discipline — retrieval that actually grounds answers, agents that coordinate instead of looping, evals that catch regressions before users do. These Claude skills encode the patterns that separate a demo from a system you can run in production: RAG architecture and debugging, multi-agent orchestration, MCP server design and LLM evaluation.

They are for engineers building on top of LLMs who want hard-won architecture, not vibes. Below are the StrategistKit skills for designing and hardening agent, RAG and LLM systems.

8 skills in this collection.

Use case

Frequently asked questions

What do these AI-system skills actually build?

Architecture and review for LLM systems: RAG pipelines, multi-agent orchestration, MCP servers, eval frameworks and production-hardening checklists. They return designs, patterns and critiques, not a black box.

Are these for building agents, or using them?

For building. They are aimed at engineers designing agent, RAG and LLM systems — orchestration patterns, retrieval debugging, evaluation and hardening.

How is a skill better than asking an LLM directly?

A skill bundles the workflow and the hard-won pattern, so you get a consistent, senior-level design pass every time instead of re-deriving it from scratch.

What do I need to run them?

A paid Claude plan. Add the .skill file to your Claude app and invoke it with your system context.

Where are they sold?

On PromptBase and Agensi, which handle payment and delivery. Pricing is per skill.

Related collections