Agentic design bible released
A senior Google engineer published a 421‑page Agentic Design Patterns document covering prompt chaining, multi‑agent coordination, guardrails and production code examples — a heavy reference for building agentic systems in production. The depth and code samples make it a practical blueprint for teams moving from copilots to orchestrated agents. (x.com)
The document’s author is Antonio Gullí, listed with Google and credited as the book’s sole author in the Springer release. (link.springer.com) Springer lists the work as a 2025 publication with digital and print ISBNs and a DOI assigned to the title. (link.springer.com) Multiple catalog entries and coverage put the complete manuscript at roughly 420–424 pages and frame it as a collection of 21 discrete agentic design patterns. (aiengineering.beehiiv.com) Publisher and retail descriptions highlight framework-agnostic, runnable examples implemented across LangChain (and LangGraph), Crew AI, and Google’s Agent Developer Kit (ADK). (amazon.com) Community-maintained GitHub mirrors host the full PDF plus chapter notebooks; several repositories report a multi‑dozen count of Jupyter notebooks mapping patterns to concrete implementations. (github.com) The book explicitly includes a chapter on the Model Context Protocol (MCP) alongside chapters covering memory, evaluation, and inter‑agent messaging according to the table‑of‑contents mirrors and repo READMEs. (github.com) Author statements and multiple repo READMEs note that author royalties will be donated to Save the Children, and the manuscript has been widely redistributed by community forks and archives. (github.com)
