Ai coding agents guided by specs improve software development process
A Lean and Spec-Driven AI-Assisted Software Development Lifecycle for Applied AI Education: The AI-SDLC Approach
Software EngineeringArtificial Intelligence
Summary
Software developers often use AI tools that help write code, but these tools usually aren't firmly connected to standard software engineering steps. The authors created a simple, clear process that uses written instructions and special files to guide AI tools through each stage of software building. This system was tested in a college course where students used it to make business software. Their feedback showed the process helped keep AI actions clear, reviewable, and properly tested throughout development.
What this means in practice
- •For software development teams: Incorporate a spec-driven lifecycle to control and audit AI coding agents during software creation, ensuring clear, testable, and reviewable development processes.
- •For software project managers: Use phase-specific AI skill configurations and repository-local guides to coordinate AI tools in business software projects, improving governance and task execution.
Authors
Andreas Martin, Sandro Schwander
Abstract
AI coding agents increasingly support software development beyond code completion, including planning, implementation, testing, and repository-level task execution. Their practical use, however, often remains only weakly connected to established software engineering practices. The aim of this work is to develop and evaluate a lightweight, spec-driven lifecycle for governed agentic software engineering. The lifecycle combines established software engineering practices with repository-local guidance through specifications, AGENTS.md, and phase-specific agent skill files. The approach was developed in the context of the FHNW course AI-assisted Software Development and applied by students to business-oriented software use cases. Its educational and practical applicability is explored through a student survey combining closed rating items with open-ended questions. The contribution of this work is a process-oriented framework that enables AI coding agents to operate with bounded autonomy within an explicit, reviewable, and test-oriented software development lifecycle.