Start Here
The introduction defines the book promise: practical agentic system patterns, clear ownership boundaries, and release-quality evidence.
Abrir secciónEdición en español
Agentic Systems Patterns
Un libro práctico para diseñar, evaluar y operar agentic systems: goals, loops, tools, skills, memory, protocols, multi-agent topologies y production runtimes.
Formatos descargables
Lee primero en GitHub Pages.
El reader online es la edición principal. PDF y EPUB se publican como copias de cortesía para lectura offline y anotaciones.
Rutas de lectura
Explora el libro
Pattern Selection and Composition
Architecture Before Autonomy explains why ownership, policy, state, evals, and rollback should come before adding agentic behavior.
Abrir secciónAgent Runtime Foundations
A single agent receives a goal or message, consults its context, and produces an answer or action. This is the smallest useful unit in the catalog.
Abrir secciónEngineering Practice and Frameworks
The lifecycle chapter turns agent work into an engineering process: design, harness, eval, trace, review, rollout, and maintenance.
Abrir secciónEvaluation, Security, and Trust
Evaluation-Driven Agent Development shows how to make evals part of design, implementation, release, and incident learning.
Abrir secciónControl Loops
Planning separates deciding what to do from doing it. The planner creates steps; the executor runs them, reports progress, and handles errors.
Abrir secciónContext, Memory, and Knowledge
Context Budgets and Working Sets explains how to select, label, compress, retrieve, and drop context before it becomes hidden risk.
Abrir secciónTools, Skills, and Protocols
Skills package procedural knowledge as discoverable, versioned folders of instructions, references, scripts, templates, assets, and tests.
Abrir secciónMulti-Agent Systems
Choosing Multi-Agent Topology helps decide when to use supervisors, delegation, parallel agents, consensus, or simpler single-agent designs.
Abrir secciónSystems Architecture
Agentic System Architecture maps how loops, tools, memory, policy, evals, runtime, and product services fit into one system.
Abrir secciónProduction Runtime
Production Runtime Overview defines the control plane that admits work, validates proposals, owns state, enforces policy, and records evidence.
Abrir secciónHands-On Labs
The lab guide explains how to use runnable examples to connect patterns to code, tests, traces, and production gaps.
Abrir secciónCapstone Projects
Capstone Projects combine multiple patterns into reviewable system packets with traces, evals, ADRs, runbooks, and rollout controls.
Abrir secciónHistorical Patterns
Historical Patterns translates older agent terminology into clearer current architecture language and explains why some terms were retired.
Abrir secciónPublishing Appendix
Publishing and Releases documents how to build, verify, package, and publish the online book and courtesy formats.
Abrir sección