AgenticAKM : Enroute to Agentic Architecture Knowledge Management

TL;DR

This paper introduces AgenticAKM, an agent-based system that automates architecture knowledge management by decomposing complex tasks into specialized agents, improving the generation of architecture decision records from code repositories.

Contribution

The paper presents a novel agentic framework for AKM that effectively automates architecture recovery and documentation, overcoming limitations of naive LLM approaches.

Findings

AgenticAKM generates better ADRs than baseline methods.

The approach is practical and promising for automating AKM.

User study confirms improved quality of architecture documentation.

Abstract

Architecture Knowledge Management (AKM) is crucial for maintaining current and comprehensive software Architecture Knowledge (AK) in a software project. However AKM is often a laborious process and is not adopted by developers and architects. While LLMs present an opportunity for automation, a naive, single-prompt approach is often ineffective, constrained by context limits and an inability to grasp the distributed nature of architectural knowledge. To address these limitations, we propose an Agentic approach for AKM, AgenticAKM, where the complex problem of architecture recovery and documentation is decomposed into manageable sub-tasks. Specialized agents for architecture Extraction, Retrieval, Generation, and Validation collaborate in a structured workflow to generate AK. To validate we made an initial instantiation of our approach to generate Architecture Decision Records (ADRs) from code repositories. We validated our approach through a user study with 29 repositories. The results demonstrate that our agentic approach generates better ADRs, and is a promising and practical approach for automating AKM.