AgentSpawn: Adaptive Multi-Agent Collaboration Through Dynamic Spawning for Long-Horizon Code Generation

TL;DR

AgentSpawn introduces a dynamic multi-agent system that adapts during long-horizon code generation tasks by spawning new agents based on runtime complexity, improving success rates and reducing memory use.

Contribution

This paper presents AgentSpawn, a novel architecture enabling adaptive agent collaboration with dynamic spawning, memory transfer, and coherence protocols for better long-term code generation.

Findings

34% higher completion rates than static baselines on SWE-bench

42% reduction in memory overhead through selective slicing

Effective handling of unanticipated complexity in code generation

Abstract

Long-horizon code generation requires sustained context and adaptive expertise across domains. Current multi-agent systems use static workflows that cannot adapt when runtime analysis reveals unanticipated complexity. We propose AgentSpawn, an architecture enabling dynamic agent collaboration through: (1) automatic memory transfer during spawning, (2) adaptive spawning policies triggered by runtime complexity metrics, and (3) coherence protocols for concurrent modifications. AgentSpawn addresses five critical gaps in existing research around memory continuity, skill inheritance, task resumption, runtime spawning, and concurrent coherence. Experimental validation demonstrates AgentSpawn achieves 34% higher completion rates than static baselines on benchmarks like SWE-bench while reducing memory overhead by 42% through selective slicing.