AgentDevel: Reframing Self-Evolving LLM Agents as Release Engineering

2026-01-08

arXiv.org

AI Summary

This work addresses the instability, poor auditability, and lack of regression control commonly observed in existing self-evolution methods for large language model (LLM) agents. The authors reframe agent evolution as a release engineering problem and propose a regression-aware, externalized release pipeline. By decoupling fault diagnosis from implementation and incorporating a behavior-flipping gating mechanism, the approach extracts symptom-level quality signals from execution traces to generate auditable engineering specifications. Evaluated on execution-intensive benchmarks, the method achieves consistent performance gains, substantially reduces regression errors, and produces a reproducible, auditable single-version evolutionary trajectory.

Technology Category

• Machine Learning: Evolutionary Learning
• Search and Optimization: Evolutionary Computation
• Multiagent Systems: Adversarial Agents

Application Category

• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Economics, Online Markets and Human Computation: LLM based quality controls for crowd work

Abstract

Recent progress in large language model (LLM) agents has largely focused on embedding self-improvement mechanisms inside the agent or searching over many concurrent variants. While these approaches can raise aggregate scores, they often yield unstable and hard-to-audit improvement trajectories, making it difficult to guarantee non-regression or to reason about failures across versions. We reframe agent improvement as \textbf{release engineering}: agents are treated as shippable artifacts, and improvement is externalized into a regression-aware release pipeline. We introduce \textbf{AgentDevel}, a release engineering pipeline that iteratively runs the current agent, produces implementation-blind, symptom-level quality signals from execution traces, synthesizes a single release candidate (RC) via executable diagnosis, and promotes it under flip-centered gating. AgentDevel features three core designs: (i) an implementation-blind LLM critic that characterizes failure appearances without accessing agent internals, (ii) script-based executable diagnosis that aggregates dominant symptom patterns and produces auditable engineering specifications, and (iii) flip-centered gating that prioritizes pass to fail regressions and fail to pass fixes as first-class evidence. Unlike population-based search or in-agent self-refinement, AgentDevel maintains a single canonical version line and emphasizes non-regression as a primary objective. Experiments on execution-heavy benchmarks demonstrate that AgentDevel yields stable improvements with significantly fewer regressions while producing reproducible, auditable artifacts. Overall, AgentDevel provides a practical development discipline for building, debugging, and releasing LLM agents as software development.

Problem

• LLM agents
• self-evolution
• non-regression
• release engineering
• auditability

Innovation

• release engineering
• implementation-blind critic
• executable diagnosis
• flip-centered gating
• non-regression