agent-patterns

A 7×6 framework for agent architecture. 28 patterns, each placed at a coordinate, runnable Python code with verified engineering slices from Claude Code, Aider, OpenHands, DeerFlow. Companion to Designing AI Agents (Manning) by Jia Huang.

Production-grade architecture patterns, decision frameworks, and best practices for building reliable AI agents. Framework-agnostic reference for engineers.

Working code for specific AI agent capabilities. Each tutorial shows the build, the failure mode you will hit, and the fix. Clone, run, adapt.

The definitive catalog of AI agent patterns. One definition, any framework. 27+ agents, 8 patterns, 6 framework adapters.

Comprehensive docs for 21 agentic AI design patterns: mermaid diagrams, ASCII art, pattern discussions, and LangChain code examples

A mined, cited, searchable corpus of implementation patterns from open-source AI-agent and research repos - queryable by coding agents over MCP (gems_query / gems_ground / gems_inspire).

Architectural patterns and best practices for building autonomous AI agents

Educational guide and worked examples for writing Claude Code .claude/agents/*.md subagent definition files: description-as-router, model-tier selection by work shape, the skill/agent content split, restating critical rules inline, and the follow-report-never-override rule for standing decisions.

Implementación de referencia del patrón LLM+Wiki (Taller IA, UAM Azcapotzalco)

Production-oriented patterns for LLM agents with tool use in Python. Guard rails, state, observability.

A blank, fill-in scaffold for the always-loaded memory file of a long-running project. Encodes six mechanisms: a closed status-label taxonomy where labels move only on evidence, numbered standing directives that can be superseded but never silently dropped, supersede-in-place archival of status blocks.

SEO entry point for the Two-Axis framework. The work lives at huangjia2019/agent-design-patterns.

🧬 Battle-tested patterns for building self-evolving AI agents. Production code from 1,000+ autonomous cycles on Claude API. Context engineering, immune systems, multi-agent orchestration.

Production-ready Claude Code agent configurations implementing patterns from Anthropic's Building Effective Agents and Measuring Agent Autonomy in Practice research

Musical groove and swing applied to agent work patterns. Pocket states for autonomy, swing scheduling for contention reduction, syncopation for creative disruption. 14 tests.

The model proposes; the harness disposes. 18+ runnable engineering patterns, framework adapters, and deterministic gates for governed, production AI agents.

116 agent design patterns in 12 categories, each with concept, flow diagram and Python example. 4 rounds of 2026 frontier updates.

Patterns for the persistent layer of agent-assisted work: subagent definitions (per-role), cross-session handoffs (per-task), and an always-loaded project memory file (per-project).