Formalizes Agentic Software Engineering (ASE) as a distinct discipline. Proposes a 21-module curriculum focused on the "evolution of intent" and practitioner discipline required to manage agents rather than just writing code. This work emphasizes that AI Literacy is a developmental capacity requiring structured Scaffolding and Prompt Engineering discipline. It connects to the need for Curriculum Design that targets Metacognition and Agentic AI rather than just syntax mastery.
Key Findings
The paper documents that professional software work increasingly consists of directing agents rather than writing code: Anthropic's Economic Index classifies 79% of Claude Code interactions as automation, Handa and colleagues find AI exposure for Computer Programmer tasks at approximately 75% of the role's distinct activities, and Brynjolfsson and colleagues report a 13% relative decline in employment for workers aged 22 to 25 in occupations most exposed to AI.The academic literature on agentic software engineering converges on the finding that the missing capability is not better models but structured practitioner discipline.ASE-26 is a comprehensive undergraduate curriculum for agentic software engineering as a discipline, deposited as a citable reference on Zenodo under CC BY-ND 4.0.Its central conceptual contribution is the evolutionary spiral as the operational form of the co-evolution of intent and build — the iterative cycle in which a developer frames intent, the agent builds, and the developer judges and revises.The curriculum sets out pedagogical commitments for grading work co-produced with an agent and is designed to outlast the specific capabilities of today's models, teaching skills such as auditability, context engineering, verification, multi-agent workflows, and AgentOps.A motivating vignette captures the shift: a developer writes a paragraph, the agent asks clarifying questions and produces six hundred lines of code with tests, a commit message, and a draft pull request — roughly eleven minutes, where the same task took two hours three years ago.Curriculum Design
The twenty-one-module structure organizes the discipline for teaching, building from the discipline's framing to the practitioner skills the industry currently lacks. Because grading work co-produced with an agent raises novel questions about authorship and assessment, the curriculum includes explicit pedagogical commitments for evaluating student work. And because model capabilities change rapidly, the discipline is framed around durable skills — framing tasks, judging outputs, verifying results, and managing the evolution of intent — rather than around any particular tool.
Implications for AI in Education
ASE-26 reframes AI Literacy for software professionals as the ability to manage agents, not merely to prompt them, and its emphasis on the co-evolution of intent and build aligns with Curriculum Design that targets Metacognition rather than syntax mastery. For higher education, the curriculum is a concrete template for programs that want to teach the discipline of Agentic AI development — including verification and auditability — rather than leaving students to acquire it informally. The labor-market evidence the paper marshals gives urgency to this curriculum agenda, and its grounding in durable practitioner discipline speaks to the broader question of what should be taught when AI can generate code.
Connected Concepts
AI LiteracyScaffoldingPrompt EngineeringCurriculum DesignMetacognitionAgentic AIConnected Articles
Tracing GenAI Literacy Interaction Patterns — Tracing GenAI Literacy: Student-AI Interaction Patterns in Academic WritingGuided LLM Scaffolding Independent Learning — Beyond Access: Guided LLM Scaffolding for Independent Learning in Undergraduate StatisticsAI Adoption Training Public Sector — The Main Barrier to AI Adoption in the Public Sector is Lack of TrainingFinkelstein Principled AI Education 2025 — Principled AI Education FrameworkAgency Gap AI Writing — The agency gap in AI-supported writing: how reactive and proactive agent designs shape multimodal reasoningAgentic AI Education Scoping Review — Agentic AI in Education: A Scoping Review of Research Landscape, Capabilities, and the Frontier Agent ParadigmCitation
Mikael Gorsky (2026). ASE-26: A Curriculum for Agentic Software Engineering as a Discipline. arXiv:2606.01152.