Classify Data with Complex Rules2 skills
Guide to text and data classification using Claude with multi-shot prompting and retrieval-augmented generation.
Engineering Blueprint
A comprehensive collection of copy-paste code examples, tutorials, and best practices covering Claude capabilities, integrations, and agent patterns.
Run this command to deploy the blueprint to your environment.
Provides production-ready code snippets and guides for common Claude use cases, removing the need to discover patterns from scratch. Each guide demonstrates techniques ranging from basic prompt chaining to multi-agent orchestration with external system integration, so you can adapt proven approaches to your own projects.
Browse the table of recipes organized by domain (Capabilities, Tool Use, Integrations, etc.) and navigate to the guide that fits your use case. Each guide is a self-contained Jupyter notebook with executable code, sample data, and evaluation approaches. Run the cells locally or adapt the code into your own project. Output: working implementations and learned patterns you can apply immediately.
Guide to text and data classification using Claude with multi-shot prompting and retrieval-augmented generation.
Techniques for enhancing Claude with domain-specific knowledge using RAG, including contextual embeddings and re-ranking.
Techniques for effective text summarization using multi-shot, domain-based, and chunking approaches.
Techniques for translating natural language questions into SQL queries with self-improvement and RAG.
Patterns for integrating external tools, APIs, and functions to extend Claude's capabilities beyond text.
Guide to Claude's vision capabilities including image analysis, chart interpretation, form extraction, and image generation.
Guide to creating and deploying custom Skills that give Claude specialized capabilities for document generation, data analysis, and business automation.
Reference implementations of fundamental agent patterns including prompt chaining, routing, orchestrators, and evaluator-optimizers.
Go from problem to working solution in hours, not weeks, using reference implementations that already handle edge cases and best practices.
All code examples are primarily Python with concepts transferable to other languages. Capabilities guides use Jupyter notebooks; Agent SDK tutorials use Python SDK. Skills require beta API headers for code execution and file generation. Managed Agents is a hosted service requiring API authentication.
Requires Claude API key for execution. Some notebooks assume Python familiarity. Third-party integrations (Pinecone, Wikipedia, Voyage AI, etc.) may require separate credentials. Agent SDK tutorials include Docker and GitHub token requirements for certain examples.
Capabilities Guides
Six core Claude techniques including classification, RAG, summarization, text-to-SQL, knowledge graphs, and contextual embeddings with evaluation.
Tool Use & Integration
Learn to extend Claude with external tools, APIs, and data sources including customer service agents, SQL queries, and vector databases.
Multimodal Examples
Vision capabilities, chart interpretation, form extraction, and image generation integrations with Stable Diffusion.
Agent Patterns
Reference implementations of prompt chaining, routing, orchestrator-subagent architectures, and evaluator-optimizer workflows.
Claude Agent SDK
Tutorial series building from simple research agents to multi-agent systems with external system integration via MCP.