liveBook · Manning

Shared state object

As explained in AI Agents in Action, Second Edition

A data object containing information that multiple agents can read and update during collaborative execution.

A shared state object is a structured data object that multiple agents read and update while working toward one goal. Unlike isolated agent work, each contribution becomes available to the rest of the team, allowing later turns to build on, critique, and refine earlier work. section 9.4

Researcher, critic, synthesizer

A collaborative research loop can place three agents around one state object. The researcher adds findings, the critic identifies weaknesses or gaps, and the synthesizer turns the accumulated work into a coherent result. A lightweight turn manager rotates control, while every agent can see the complete shared state. section 9.4

How AI Agents in Action, Second Edition covers it

How deep it goes
Defines typed fields and a consensus property, then shows initialization and loop execution without descending into storage internals.
Examples
Uses a researcher-critic-synthesizer triad, typed contributions, round limits, confidence values, and recent-contribution agreement.
What you can do after
Build it from scratch as a typed collaboration state and connect it to rotating agents and consensus-based stopping.
How it connects
Connects collaborative agent loops with multi-agent systems, cognitive architectures, global workspaces, and consensus.

Related concepts

This explainer was generated by Manning's knowledge graph from the text of AI Agents in Action, Second Edition. The section links point to the parts of the book it draws on.