The catalog
Nineteen volumes, 331 patterns.
Grouped into five clusters: the workflow primitives, the infrastructure, context and content, the non-functional concerns, and UX and products. A Fowler-style reference for agentic AI engineering, organized for retrieval, not linear reading. Start anywhere.
Workflow & Primitives
The foundational primitives — what agents do, what they use, what triggers them.
Patterns of AI Agent Workflows
The workflow patterns: Prompt Chaining, Routing, ReAct, Plan-and-Execute, Reflexion, Orchestrator-Workers, Code-as-Action, and long-running agent harnesses. The first volume; established the format.
VOL 02 28 patternsThe Claude Skills Catalog
Skills as the unit of agent capability: Anthropic's skills, community collections, Grid Dynamics Rosetta, obra/superpowers, and how skills ship through CLIs and plugin marketplaces.
VOL 03 36 patternsThe AI Agent Tools Catalog
Function-calling primitives: server tools, MCP servers, sandboxes, browser and web tools, tool design, and Code Mode / programmatic tool calling.
VOL 04 23 patternsThe AI Agent Events & Triggers Catalog
Reactive orchestration: schedule, webhook, file-watcher, queue and message-bus triggers, plus durable execution - the patterns that make agents respond rather than only execute.
VOL 18 14 patternsThe Claude Code Configuration & Workflows Catalog
Configuring and driving Claude Code: CLAUDE.md memory and path-scoped rules, skills and plugins, project subagents, permission rules and auto mode, lifecycle hooks, headless / SDK automation, and spec-driven development.
VOL 19 6 patternsThe Agentic Loop & Autonomy Catalog
The loop-invocation axis: turn-based, goal-gated, interval, local scheduled and cloud scheduled loops, ordered by who holds the stop decision, plus composition. The autonomy companion to Vols 1 and 4.
Infrastructure
The substrate on which agents run — memory, fabric, multi-agent coordination, human-in-the-loop, evaluation.
The AI Agent Fabric Catalog
The infrastructure layer: managed agent runtimes, sandbox isolation tiers, agent gateways, Kubernetes-native agents, deployment and scaling primitives.
VOL 06 17 patternsThe AI Agent Memory Catalog
State and memory: vector stores, dedicated memory layers (Mem0, Letta, Zep, LangMem), managed cloud memory, Postgres+pgvector, temporal knowledge graphs, memory benchmarks.
VOL 07 13 patternsThe Human-in-the-Loop Catalog
Approval gates, declarative per-tool approval, MCP elicitation, observability for agents, intervention mechanisms.
VOL 08 21 patternsThe Evaluation & Guardrails Catalog
Eval frameworks, runtime guardrails, prompt-injection classifiers, red-teaming, agent task benchmarks, the OWASP LLM and Agentic Top 10s.
VOL 09 16 patternsThe Multi-Agent Coordination Catalog
Protocols (MCP, A2A, Agent Client Protocol, agent payments), topologies, frameworks (LangGraph, CrewAI, OpenAI Agents SDK, Google ADK), coordination patterns.
Context & Content
What agents work with — retrieved knowledge, prompts, context windows.
The Retrieval Catalog
RAG patterns at the agent layer: chunking and contextual embedding, retrieval pipelines, managed file search, agentic search, citation.
VOL 15 13 patternsThe Prompting & Context Engineering Catalog
Prompt design, context window engineering, structured output, reasoning effort and compaction: the discipline of structuring agent inputs.
Non-Functional Concerns
Compliance, security, ops, economics — the substrate that makes production agents possible.
The Compliance & Regulatory Catalog
EU AI Act (as amended in 2026), NIST AI RMF, ISO 42001 and 42005, US state AI laws, GDPR-for-AI patterns, the regulated-industry overlay.
VOL 12 16 patternsThe AI Infrastructure Security Catalog
Beyond runtime guardrails: sandboxing, MCP and skill supply-chain scanning, prompt-injection containment, agent identity, secrets and audit infrastructure.
VOL 16 8 patternsThe LLMOps Catalog
Operating LLM-augmented systems: AI gateways, prompt and semantic caching, model deployment, monitoring, fine-tuning, cost management at the model layer.
VOL 17 10 patternsThe AI Agent Economics Catalog
Pricing models, service tiers, FOCUS cost data, ROI and build-vs-buy, pricing the agents you sell: the business-side discipline.
UX & Products
The user-facing surfaces of agents and the product landscape.
The Agent UX Patterns Catalog
Token streaming, intermediate state display, generative UI and agent-to-UI protocols (AG-UI, MCP Apps, A2UI), approval UX, realtime voice agents.
VOL 14 13 patternsThe AI Agent Products Survey
The agent product landscape: coding agents, prompt-to-app builders, voice agent platforms, enterprise agent platforms, frameworks and observability.