Abstract
Large Language Models (LLMs) generate plausible backend code, but a single-pass paradigm provides no guarantee of correctness or runtime reliability. We present ExecuGraph, a multi-agent framework that places execution-based validation at the center of backend code synthesis.
The framework consists of six specialized agents (Planner, Code Generator, Logical Reviewer, Evaluator, Optimizer, and Explainer), coordinated by a typed directed workflow with a bounded retry budget. ExecuGraph is implemented on LangGraph with locally hosted models (Ollama) and an optional retrieval layer for algorithmic technique recall. Every evaluation is guarded by a subprocess-isolated sandbox with a wall-clock timeout.
We evaluate on a curated 30-problem DSA suite (internal-30), HumanEval (n=64), and an APPS-introductory subset, contrasting ExecuGraph against a single-agent one-shot baseline and a single-agent execution-retry baseline (a Reflexion-style ablation that isolates the contribution of multi-agent decomposition).
On internal-30, the three conditions are statistically indistinguishable (n=30; paired Wilcoxon p=0.59 MF vs. SO, p=0.08 SR vs. SO); 95% bootstrap confidence intervals on all pairwise mean differences include zero. On HumanEval, multi-full edges ahead by +3.1 pp. The strongest signal is cross-model: with DeepSeekCoder V2 Lite, graph-category accuracy improves from 57.5% (oneshot) to 80.0% (multi-full), a +22.5 pp jump that supports a scaling hypothesis: the value of multi-agent decomposition grows with base-model capability.
The primary contribution of the framework is methodological: a single codebase that collapses by configuration into one-shot, execution-retry, and per-agent ablation conditions, enabling controlled measurement of each lever's marginal contribution. A per-agent ablation, retry-budget sweep, error-class taxonomy, and test-source audit are reported.
Blogger's Review: ExecuGraph enhances the reliability of backend code generation through multi-agent collaboration, particularly showcasing the substantial advantages of execution validation. This approach not only optimizes generation outcomes but also provides fresh insights and frameworks for future research in code synthesis.