agentic-patterns-source

# Agentic Patterns Source

## Overview

[Awesome Agentic Patterns](https://agentic-patterns.com/) is a curated catalog of production-ready AI agent patterns. It bridges the gap between toy demos and production systems by documenting repeatable solutions that teams are actually using.

**Primary URL**: https://agentic-patterns.com/

## When to Use This Skill

| Use this skill when... | Use other skills when... |
|------------------------|-------------------------|
| Designing multi-agent orchestration | Implementing Claude Code-specific hooks (`claude-hooks-configuration`) |
| Solving context window limitations | Writing custom agent definitions (`custom-agent-definitions`) |
| Implementing feedback/self-correction loops | Creating file-based coordination (`agent-file-coordination`) |
| Researching security/sandboxing patterns | Setting up MCP servers (`mcp-management`) |
| Need production-tested patterns with references | Need Claude Code workflow patterns (`multi-agent-workflows`) |

## Pattern Categories

### 1. Context & Memory

Managing limited context windows through curation, caching, and episodic memory.

**Key Patterns:**
- Context Window Auto-Compaction
- Progressive Disclosure for Large Files
- Semantic Context Filtering
- Self-Identity Accumulation

**When to research:** Context limits affecting performance, memory persistence needs, large codebase navigation.

### 2. Feedback Loops

Self-healing mechanisms, CI integration, and iterative refinement.

**Key Patterns:**
- Self-correction loops
- CI integration patterns
- Iterative refinement workflows

**When to research:** Implementing retry logic, self-healing agents, continuous improvement cycles.

### 3. Learning & Adaptation

Reinforcement fine-tuning and skill library evolution.

**Key Patterns:**
- Memory Reinforcement Learning (MemRL)
- Skill Library Evolution

**When to research:** Agents that improve over time, skill accumulation, adaptive behavior.

### 4. Orchestration & Control

Task decomposit