liveBook · Manning

Task

As explained in Build a Multi-Agent System (from Scratch)

An A2A task is a stateful object representing an agent's execution of a request, including its status and any results.

An A2A Task is a stateful record of a remote agent request. It can contain zero or more Artifacts, and clients can refer to it by task ID in later messages. section 10.1.3

A client starts a task by sending a Message with a Part containing instructions. The remote agent returns a Task after completion or another terminal state; an Artifact carries the result. section 10.1.3

Task status shapes how the exchange proceeds:

  • An input_required Task signals that the remote agent needs clarification; the client passes its task ID back in a reply. section 10.2.2
  • On the server, AgentExecutor handles execution while TaskUpdater publishes status updates and the final Artifact. section 10.3.1

Hailstone request

A client can send the instruction “Compute the Hailstone sequence starting at 6” in a Message Part. The remote agent processes the task and returns its result. section 10.1.3

Figure 10.6 A sequence diagram of the task lifecycle, overlaid with the A2A elements

Follow the client-server exchange from the instruction Message through Task execution and results. section 10.1.3

Figure 10.7 The sequence of interactions for an input-required workflow

Trace the interaction when a remote agent pauses execution to request more input. section 10.2.2

How Build a Multi-Agent System (from Scratch) covers it

How deep it goes
Defines Message–Part–Task–Artifact relationships and executor updates, while noting that its blocking client does not use polling.
Examples
Uses a Hailstone request and an input-required clarification workflow.
What you can do after
Operate A2A task exchanges by sending instructions, tracking task IDs, and consuming final Artifacts.
How it connects
Connects A2A Message, Part, Artifact, AgentExecutor, and TaskUpdater.

Also explained in

Compare how 2 books explain task →

Related concepts

This explainer was generated by Manning's knowledge graph from the text of Build a Multi-Agent System (from Scratch). The section links point to the parts of the book it draws on.