Narrow Decomposition Risk

Core

Orchestrate multi-agent systems with coordinator-subagent patterns · Difficulty 3/5

decompositioncoveragemulti-agent

Prerequisites

When a coordinator agent decomposes a broad topic into subtasks, narrow decomposition is a common failure mode that produces incomplete results.

Topic: "Impact of AI on creative industries"

Coordinator decomposes into:

AI in digital art creation

AI in graphic design

AI in photography

Result: Report covers only visual arts, completely missing music, writing, and film.

The coordinator's task decomposition was too narrow. All three subtasks focused on visual arts, missing other creative domains.

Explicit decomposition guidelines: "Ensure subtasks cover ALL relevant sub-domains"

Coverage validation: After decomposition, validate that subtasks span the full scope

Domain enumeration: Require the coordinator to enumerate all relevant domains before creating subtasks

Partition by source type: Assign distinct subtopics or source types to each agent to minimize duplication

When subagents all succeed but the final output has gaps, the root cause is usually the coordinator's decomposition, not the subagents' execution.

✓Narrow task decomposition causes incomplete results even when all subtasks succeed
✓Validate that decomposition covers the full scope before executing
✓If subagents succeed but output has gaps, check coordinator decomposition

Use prompt chaining for predictable workflows, dynamic decomposition for open-ended tasks