Skip to content

A central collection of turn-key agentic workflow templates for building AI-powered applications with Claude.

License

Notifications You must be signed in to change notification settings

ChrisSc/agentic-patterns

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

6 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Agentic Patterns

A central collection of turn-key agentic workflow templates for building AI-powered applications with Claude.

What Are Agentic Patterns?

Agentic patterns are reusable, well-architected blueprints for AI workflows. They provide structured approaches to common AI application problems—like breaking down complex tasks into sequential steps or executing independent operations in parallel—with built-in observability, cost tracking, and quality validation.

Key Features

  • Turn-Key Templates: Complete, tested implementations ready to deploy immediately
  • GitHub Repository Templates: Use the "Use this template" button to create a new repository
  • Comprehensive Documentation: Each pattern includes architecture guides, configuration strategies, and examples
  • Observable & Measurable: Token tracking, cost logging, and structured JSON logging included
  • Flexible Configuration: Per-step model selection, validation gates, rate limiting, and more

Quick Start

Prerequisites

  • Python 3.10 or higher
  • Anthropic API key (for Claude access)
  • Git (for cloning repositories)

Using a Template

Option 1: Use GitHub's template button Navigate to prompt-chaining and click "Use this template" to create your own repository.

Option 2: Clone directly

git clone https://github.com/ChrisSc/prompt-chaining.git
cd prompt-chaining

Example: Prompt-Chaining Pattern

The prompt-chaining pattern breaks complex tasks into three sequential steps:

Analyze → Process → Synthesize

Perfect for document summarization, content generation, or data extraction. See the Prompt-Chaining template for a complete working example.

Available Patterns

Prompt-Chaining

Sequential task decomposition with validation gates

A three-step workflow where each step builds on the previous: analyze user intent, process with the appropriate model, and synthesize a polished response. Includes per-step model configuration, validation gates, and comprehensive cost/performance tracking.

  • When to use: Document analysis, content generation, data extraction, decision support
  • Key feature: Validation gates between steps ensure quality
  • Cost: $0.006–$0.018 per request depending on configuration

For detailed pattern information, see PATTERNS.md.

Repository Structure

agentic-patterns/
├── README.md                           # This file
├── PATTERNS.md                         # Pattern catalog and selection guide
├── LICENSE.md                          # MIT License

Pattern templates are complete, standalone repositories:

Pattern Selection

Not sure which pattern to use? Check PATTERNS.md for a decision framework and comparison guide.

License

This project is licensed under the MIT License. See LICENSE.md for details.

About

A central collection of turn-key agentic workflow templates for building AI-powered applications with Claude.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Contributors 2

  •  
  •