Multi-Agent Research

A coordinator agent spawns parallel research agents, each exploring different aspects, then synthesizes all findings into a recommendation.

agentmcp

Why OSOP matters here

Multi-agent systems are the hardest to debug. OSOP tracks parent-child relationships with spawn indices. When Agent #3 produces a bad result, you can trace exactly what it was asked and what tools it used.

Workflow Steps (5)

Research Coordinator

agent

Code Analysis Agent

agent

Documentation Agent

agent

Test Coverage Agent

agent

Synthesize Findings

agent

Connections (6)

Research Coordinator→Code Analysis Agentspawn

Research Coordinator→Documentation Agentspawn

Research Coordinator→Test Coverage Agentspawn

Code Analysis Agent→Synthesize Findingsparallel

Documentation Agent→Synthesize Findingsparallel

Test Coverage Agent→Synthesize Findingsparallel

Steps

Connections

Node Types

Open in Visual Editor Browse Examples on GitHub

← PreviousClient Health Alert Next →RAG Pipeline with Quality Gates