Awesome Agentic Patterns 
A curated catalogue of agentic AI patterns — real‑world tricks, workflows, and mini‑architectures that help autonomous or semi‑autonomous AI agents get useful work done in production.
Why? Tutorials show toy demos. Real products hide the messy bits. This list surfaces the repeatable patterns that bridge the gap so we can all ship smarter, faster agents.
What counts as a pattern?
- Repeatable – more than one team is using it.
- Agent‑centric – improves how an AI agent senses, reasons, or acts.
- Traceable – backed by a public reference: blog post, talk, repo, or paper.
If your link ticks those boxes, it belongs here.
Quick Tour of Categories
| Category | What you'll find |
|---|---|
| Orchestration & Control | Task decomposition, sub‑agent spawning, tool routing |
| Context & Memory | Sliding‑window curation, vector cache, episodic memory |
| Feedback Loops | Compilers, CI, human review, self‑healing retries |
| Tool Use & Environment | Shell, browser, DB, Playwright, sandbox tricks |
| UX & Collaboration | Prompt hand‑offs, staged commits, async background agents |
| Reliability & Eval | Guardrails, eval harnesses, logging, reproducibility |
Categories are fluid — open a PR if you see a better slice!
The tables below are auto‑generated from the patterns/ folder.
Context & Memory
- Context Window Anxiety Management
- Context-Minimization Pattern UPDATED
- Curated Code Context Window
- Curated File Context Window
- Dynamic Context Injection
- Episodic Memory Retrieval & Injection
- Filesystem-Based Agent State
- Layered Configuration Context
- Memory Synthesis from Execution Logs
- Proactive Agent State Externalization
Feedback Loops
- Background Agent with CI Feedback
- Coding Agent CI Feedback Loop
- Dogfooding with Rapid Iteration for Agent Improvement
- Graph of Thoughts (GoT)
- Inference-Healed Code Review Reward
- Reflection Loop
- Rich Feedback Loops > Perfect Prompts
- Self-Critique Evaluator Loop
- Self-Discover: LLM Self-Composed Reasoning Structures
- Spec-As-Test Feedback Loop
- Tool Use Incentivization via Reward Shaping
Learning & Adaptation
- Agent Reinforcement Fine-Tuning (Agent RFT)
- Compounding Engineering Pattern
- Skill Library Evolution
- Variance-Based RL Sample Selection
Orchestration & Control
- Action-Selector Pattern
- Agent-Driven Research
- Autonomous Workflow Agent Architecture
- Conditional Parallel Tool Execution
- Continuous Autonomous Task Loop Pattern
- Discrete Phase Separation UPDATED
- Disposable Scaffolding Over Durable Features
- Distributed Execution with Cloud Workers UPDATED
- Dual LLM Pattern
- Explicit Posterior-Sampling Planner
- Feature List as Immutable Contract
- Inference-Time Scaling
- Initializer-Maintainer Dual Agent Architecture
- Inversion of Control
- Iterative Multi-Agent Brainstorming
- Language Agent Tree Search (LATS)
- LLM Map-Reduce Pattern
- Multi-Model Orchestration for Complex Edits UPDATED
- Opponent Processor / Multi-Agent Debate Pattern
- Oracle and Worker Multi-Model Approach
- Parallel Tool Call Learning
- Plan-Then-Execute Pattern
- Progressive Autonomy with Model Evolution
- Progressive Complexity Escalation UPDATED
- Self-Rewriting Meta-Prompt Loop
- Specification-Driven Agent Development
- Stop Hook Auto-Continue Pattern
- Sub-Agent Spawning UPDATED
- Swarm Migration Pattern
- Three-Stage Perception Architecture
- Tool Capability Compartmentalization
- Tree-of-Thought Reasoning
Reliability & Eval
- Anti-Reward-Hacking Grader Design
- Asynchronous Coding Agent Pipeline
- CriticGPT-Style Code Review
- Extended Coherence Work Sessions
- Lethal Trifecta Threat Model
- Merged Code + Language Skill Model
- No-Token-Limit Magic
- RLAIF (Reinforcement Learning from AI Feedback)
- Structured Output Specification UPDATED
- Versioned Constitution Governance
Security & Safety
Tool Use & Environment
- Agent SDK for Programmatic Control
- Agent-First Tooling and Logging
- Agentic Search Over Vector Embeddings
- CLI-Native Agent Orchestration
- Code Mode MCP Tool Interface Improvement Pattern
- Code-Over-API Pattern
- Code-Then-Execute Pattern
- Dual-Use Tool Design
- Dynamic Code Injection (On-Demand File Fetch)
- Egress Lockdown (No-Exfiltration Channel)
- LLM-Friendly API Design
- Patch Steering via Prompted Tool Selection
- Progressive Tool Discovery
- Shell Command Contextualization
- Subagent Compilation Checker
- Tool Use Steering via Prompting
- Virtual Machine Operator Agent
- Visual AI Multimodal Integration
UX & Collaboration
- Abstracted Code Representation for Review
- Agent-Assisted Scaffolding
- Agent-Friendly Workflow Design
- Chain-of-Thought Monitoring & Interruption UPDATED
- Human-in-the-Loop Approval Framework UPDATED
- Latent Demand Product Discovery
- Seamless Background-to-Foreground Handoff
- Spectrum of Control / Blended Initiative
- Team-Shared Agent Configuration as Code
- Verbose Reasoning Transparency
Contributing in 3 steps
- Fork & branch →
git checkout -b add-my-pattern - Add a pattern file under
patterns/using the template above. - Open a PR titled
Add: my-pattern-name— the README & site will regenerate automatically.
See CONTRIBUTING.md for the fine print.
Inspiration
This project started after the write‑up "What Sourcegraph learned building AI coding agents" (28 May 2025) and the ongoing Raising an Agent video diary. Many first patterns come straight from those lessons — thanks to everyone sharing their journey in the open!
License
Apache‑2.0. See LICENSE.