# SEW: Self-Evolving Agentic Workflows for Automated Code Generation

arXiv: https://arxiv.org/abs/2505.18646 · HTML: https://mlsanity.com/paper/2505.18646/

## Key facts

- Cited by 20 papers in the map (data through 14 Oct 2026).
- Placed in Software engineering coding agents, one of 16 directions in LLM agents and tool use.
- Builds on 19 of its 33 references that are in the map.
- First posted on arXiv 24 May 2025.

## Placed in

- LLM agents and tool use / [Software engineering coding agents](https://mlsanity.com/agents/software-engineering-coding-agents/)
- Synthetic data and self-improvement for LLMs / [Self-evolving LLM agents](https://mlsanity.com/synth/self-evolving-llm-agents/)

## Builds on (most cited first, 19 of 19)

- [Program Synthesis with Large Language Models](https://mlsanity.com/paper/2108.07732/) (2021)
- [LiveCodeBench: Holistic and Contamination Free Evaluation of Large Language Models for Code](https://mlsanity.com/paper/2403.07974/) (2024)
- [MetaGPT: Meta Programming for A Multi-Agent Collaborative Framework](https://mlsanity.com/paper/2308.00352/) (2023)
- [Is Your Code Generated by ChatGPT Really Correct? Rigorous Evaluation of Large Language Models for Code Generation](https://mlsanity.com/paper/2305.01210/) (2023)
- [Promptbreeder: Self-Referential Self-Improvement Via Prompt Evolution](https://mlsanity.com/paper/2309.16797/) (2023)
- [TheAgentCompany: Benchmarking LLM Agents on Consequential Real World Tasks](https://mlsanity.com/paper/2412.14161/) (2024)
- [Optimizing Instructions and Demonstrations for Multi-Stage Language Model Programs](https://mlsanity.com/paper/2406.11695/) (2024)
- [MapCoder: Multi-Agent Code Generation for Competitive Problem Solving](https://mlsanity.com/paper/2405.11403/) (2024)
- [Structured Chain-of-Thought Prompting for Code Generation](https://mlsanity.com/paper/2305.06599/) (2023)
- [Large Language Models as Evolutionary Optimizers](https://mlsanity.com/paper/2310.19046/) (2023)
- [AgentScope: A Flexible yet Robust Multi-Agent Platform](https://mlsanity.com/paper/2402.14034/) (2024)
- [EvoAgent: Towards Automatic Multi-Agent Generation via Evolutionary Algorithms](https://mlsanity.com/paper/2406.14228/) (2024)
- [Multi-Agent Design: Optimizing Agents with Better Prompts and Topologies](https://mlsanity.com/paper/2502.02533/) (2025)
- [Symbolic Learning Enables Self-Evolving Agents](https://mlsanity.com/paper/2406.18532/) (2024)
- [TaskWeaver: A Code-First Agent Framework](https://mlsanity.com/paper/2311.17541/) (2023)
- [FlowBench: Revisiting and Benchmarking Workflow-Guided Planning for LLM-based Agents](https://mlsanity.com/paper/2406.14884/) (2024)
- [PyBench: Evaluating LLM Agent on various real-world coding tasks](https://mlsanity.com/paper/2407.16732/) (2024)
- [AIOS Compiler: LLM as Interpreter for Natural Language Programming and Flow Programming of AI Agents](https://mlsanity.com/paper/2405.06907/) (2024)
- [On the effectiveness of Large Language Models for GitHub Workflows](https://mlsanity.com/paper/2403.12446/) (2024)

## Built on by (most cited first, 20 of 20)

- [Agent KB: Leveraging Cross-Domain Experience for Agentic Problem Solving](https://mlsanity.com/paper/2507.06229/) (2025)
- [Code as Agent Harness](https://mlsanity.com/paper/2605.18747/) (2026)
- [EvoAgentX: An Automated Framework for Evolving Agentic Workflows](https://mlsanity.com/paper/2507.03616/) (2025)
- [Recursive Harness Self-Improvement](https://mlsanity.com/paper/2607.15524/) (2026)
- [From Static Templates to Dynamic Runtime Graphs: A Survey of Workflow Optimization for LLM Agents](https://mlsanity.com/paper/2603.22386/) (2026)
- [HarnessDev: Can LLMs Create and Evolve Their Own Agent Harness?](https://mlsanity.com/paper/2609.01437/) (2026)
- [HarnessEvolve: Learning from Reference Trajectories for Reliable Agent Self-Evolution](https://mlsanity.com/paper/2609.00829/) (2026)
- [Do Self-Evolving Agents Forget? Capability Degradation and Preservation in Lifelong LLM Agent Adaptation](https://mlsanity.com/paper/2605.09315/) (2026)
- [AgentStream: How Well Do Self-Evolving LLM Agents Perform Under Streaming Tasks?](https://mlsanity.com/paper/2608.00155/) (2026)
- [Polymath: A Self-Optimizing Agent with Dynamic Hierarchical Workflow](https://mlsanity.com/paper/2508.02959/) (2025)
- [Self-Evolving Coding Agents](https://mlsanity.com/paper/2608.03392/) (2026)
- [GLOW: Graph-Language Co-Encoding for Agentic Workflow Performance Prediction](https://mlsanity.com/paper/2512.15751/) (2025)
- [Knowledge-Guided Multi-Agent Framework for Application-Level Software Code Generation](https://mlsanity.com/paper/2510.19868/) (2025)
- [TRACE: An Experiential Framework for Coherent Multi-hop Knowledge Graph Question Answering](https://mlsanity.com/paper/2604.11193/) (2026)
- [AgentCo-op: Retrieval-Based Synthesis of Interoperable Multi-Agent Workflows](https://mlsanity.com/paper/2605.20425/) (2026)
- [Proxifield: Decentralized Multi-Agent Communication through Semantic Proximity](https://mlsanity.com/paper/2609.20889/) (2026)
- [MoFlow: Multi-Objective Agentic Workflow Generation](https://mlsanity.com/paper/2609.38294/) (2026)
- [Breaking Babel: A Self-Evolving Multi-Agent System for Long-Form Subtitle Translation](https://mlsanity.com/paper/2609.38660/) (2026)
- [Joint Workflow and Prompt Optimization for User Behavior Simulation](https://mlsanity.com/paper/2610.07663/) (2026)
- [CB-Orchestrator: Adaptive Workflow Optimization for LLM Agents via Contextual Bandits](https://mlsanity.com/paper/2610.13064/) (2026)

## Abstract (arXiv)

Large Language Models (LLMs) have demonstrated effectiveness in code generation tasks. To enable LLMs to address more complex coding challenges, existing research has focused on crafting multi-agent systems with agentic workflows, where complex coding tasks are decomposed into sub-tasks, assigned to specialized agents. Despite their effectiveness, current approaches heavily rely on hand-crafted agentic workflows, with both agent topologies and prompts manually designed, which limits their ability to automatically adapt to different types of coding problems. To address these limitations and enable automated workflow design, we propose Self-Evolving Workflow (SEW), a novel self-evolving framework that automatically generates and optimises multi-agent workflows. Extensive experiments on three coding benchmark datasets, including the challenging LiveCodeBench, demonstrate that our SEW can automatically design agentic workflows and optimise them through self-evolution, bringing up to 12% improvement on LiveCodeBench compared to using the backbone LLM only. Furthermore, by investigating different representation schemes of workflow, we provide insights into the optimal way to encode workflow information with text.