pith. sign in

arxiv: 2605.24447 · v1 · pith:WPYBSGRAnew · submitted 2026-05-23 · 💻 cs.SE

Beyond AI Delegation: A Prompt Pattern Framework for Productive Struggle and Evaluative Judgement in Secure Coding Education

Philipp Haindl , Oliver Eigner , Peter Kieseberg This is my paper

Pith reviewed 2026-06-30 13:36 UTC · model grok-4.3

classification 💻 cs.SE
keywords prompt patternsproductive struggleevaluative judgementsecure coding educationAI in educationGenAI delegationprompt engineeringDELTA framework
0
0 comments X

The pith

Nine prompt patterns mapped to productive struggle and evaluative judgement let instructors integrate AI into secure coding courses while keeping students in the reasoning role.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper develops a framework to stop students from offloading coding work to large language models by structuring AI use around required effort and judgement. It adapts nine prompt patterns drawn from computer science literature and connects them to productive struggle and evaluative judgement. An advanced secure coding course design built with the DELTA framework shows three patterns guiding students through vulnerability discovery and fixes. This approach matters because unchecked AI access can remove the effort that builds lasting skill. The result is a replicable method for AI-augmented classes that still demands student thinking.

Core claim

We apply Design Science Research to synthesise and adapt a taxonomy of nine prompt engineering patterns from established catalogs in the computer science literature, mapped to two pedagogical constructs: Productive Struggle and Evaluative Judgement. A course design for an Advanced Secure Coding module, structured using the DELTA framework, demonstrates the artifact's applicability. Nine prompt patterns, each mapped to a specific pedagogical function, give instructors fine-grained control over how students interact with AI. The secure coding design shows how three patterns (Flipped Interaction, Alternative Approaches, and Cognitive Verifier) scaffold vulnerability discovery and remediation wh

What carries the argument

Nine prompt patterns adapted from CS literature and mapped to Productive Struggle and Evaluative Judgement through the DELTA framework.

If this is right

  • Instructors obtain concrete patterns that control student-AI interactions at the level of individual pedagogical goals.
  • Students remain responsible for reasoning during vulnerability discovery and remediation in secure coding tasks.
  • The framework supplies a documented template that can be copied to other modules or courses.
  • Future empirical studies can test the framework directly in live classroom settings.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • The same patterns could be tested in non-security programming courses where AI delegation is also a risk.
  • Tool builders might incorporate prompts that force alternative approaches or verification steps by default.
  • Long-term use might reduce students' tendency to accept AI output without review once they leave the course.

Load-bearing premise

Adapting a taxonomy of prompt patterns from existing literature and mapping them to productive struggle and evaluative judgement via the DELTA framework will yield a replicable method that preserves student reasoning.

What would settle it

A controlled comparison in a secure coding course that measures whether students using the nine prompt patterns delegate fewer tasks to AI and show stronger independent vulnerability detection than students given unrestricted AI access.

read the original abstract

Large language models make it easy for students to delegate writing, analysis, and problem-solving to automated systems, bypassing the effortful engagement that produces lasting understanding. We introduce a practical framework that helps educators keep GenAI in the course without removing the cognitive demands that make it worthwhile. We apply Design Science Research (DSR) to synthesise and adapt a taxonomy of nine prompt engineering patterns from established catalogs in the computer science literature, mapped to two pedagogical constructs: Productive Struggle and Evaluative Judgement. A course design for an Advanced Secure Coding module, structured using the DELTA framework, demonstrates the artifact's applicability. Nine prompt patterns, each mapped to a specific pedagogical function, give instructors fine-grained control over how students interact with AI. The secure coding design shows how three patterns (Flipped Interaction, Alternative Approaches, and Cognitive Verifier) scaffold vulnerability discovery and remediation while keeping students in the reasoning role. The framework provides a replicable approach to designing AI-augmented learning experiences that preserve student reasoning, and establishes a structured basis for future empirical evaluation in live course settings.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 1 minor

Summary. The manuscript applies Design Science Research to synthesize a taxonomy of nine prompt engineering patterns from the CS literature, mapping them to the pedagogical constructs of productive struggle and evaluative judgement. It demonstrates the resulting framework through a conceptual course design for an Advanced Secure Coding module structured via the DELTA framework, asserting that three specific patterns (Flipped Interaction, Alternative Approaches, and Cognitive Verifier) enable AI-assisted vulnerability discovery and remediation while keeping students in the reasoning role. The work claims to deliver a replicable approach for AI-augmented learning experiences and a basis for future empirical evaluation.

Significance. If the proposed mappings can be shown to function as described in practice, the framework would supply instructors with a structured, literature-grounded method for integrating generative AI into secure coding education without bypassing the cognitive effort required for lasting skill development. The DSR approach provides a systematic foundation for the artifact, and the focus on secure coding addresses a domain where AI tools are rapidly adopted.

major comments (2)
  1. [Abstract] Abstract: The assertion that the secure coding design 'shows how three patterns (Flipped Interaction, Alternative Approaches, and Cognitive Verifier) scaffold vulnerability discovery and remediation while keeping students in the reasoning role' is presented as a demonstration of the artifact, yet the manuscript supplies only the authors' conceptual description of the DELTA-structured course design with no student interaction data, output analysis, or measures of productive struggle/evaluative judgement.
  2. [Framework application section] Framework application section: The central claim of providing a 'replicable approach ... that preserve[s] student reasoning' rests on the untested assumption that the literature-derived pattern mappings will produce the intended cognitive outcomes; the manuscript does not address potential failure modes in which the patterns might still permit delegation or fail to elicit evaluative judgement.
minor comments (1)
  1. The abstract and conclusion could more explicitly qualify the demonstration as conceptual and distinguish it from empirical validation to align reader expectations with the evidence presented.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the constructive comments, which help clarify the scope of our Design Science Research contribution. We address each major comment below and propose targeted revisions to better distinguish the conceptual nature of the artifact from empirical validation.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The assertion that the secure coding design 'shows how three patterns (Flipped Interaction, Alternative Approaches, and Cognitive Verifier) scaffold vulnerability discovery and remediation while keeping students in the reasoning role' is presented as a demonstration of the artifact, yet the manuscript supplies only the authors' conceptual description of the DELTA-structured course design with no student interaction data, output analysis, or measures of productive struggle/evaluative judgement.

    Authors: We agree that the demonstration is conceptual rather than empirical. The manuscript applies DSR to develop the framework artifact, with the secure coding course design serving as an illustrative application of the patterns. We will revise the abstract to change 'shows how' to 'conceptually demonstrates how' and reinforce that the work provides a basis for future empirical evaluation, consistent with the paper's positioning and conclusion. revision: yes

  2. Referee: [Framework application section] Framework application section: The central claim of providing a 'replicable approach ... that preserve[s] student reasoning' rests on the untested assumption that the literature-derived pattern mappings will produce the intended cognitive outcomes; the manuscript does not address potential failure modes in which the patterns might still permit delegation or fail to elicit evaluative judgement.

    Authors: The mappings are synthesized from established CS literature on prompt patterns and pedagogical constructs, offering a structured, replicable approach as claimed. We acknowledge that potential failure modes are not explicitly addressed. We will add a paragraph in the framework application section discussing scenarios where patterns might permit delegation (e.g., superficial student engagement) and how the DELTA structure and evaluative judgement focus can mitigate these, while maintaining the paper's focus on artifact design. revision: yes

Circularity Check

0 steps flagged

No significant circularity: synthesis from external literature

full rationale

The paper's central derivation applies Design Science Research to synthesize and adapt a taxonomy of nine prompt patterns from established external CS literature catalogs, then maps them to the pedagogical constructs of Productive Struggle and Evaluative Judgement via the DELTA framework for a secure coding course design. No equations, fitted parameters, self-definitional reductions, or load-bearing self-citations appear in the provided text; the claims rest on literature-derived mappings and a descriptive demonstration rather than any quantity or premise that reduces to the authors' own prior outputs by construction. This is a standard non-circular synthesis case.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claim rests on domain assumptions about the transferability of prompt patterns to pedagogical constructs and the suitability of DSR for this synthesis, without free parameters or invented entities.

axioms (2)
  • domain assumption Prompt engineering patterns from computer science literature can be meaningfully mapped to pedagogical constructs like productive struggle and evaluative judgement.
    This mapping forms the core of the nine-pattern taxonomy.
  • domain assumption Design Science Research is an appropriate method for developing and demonstrating the applicability of the educational framework.
    The paper applies DSR to synthesise the artifact and show its use in a secure coding module.

pith-pipeline@v0.9.1-grok · 5726 in / 1258 out tokens · 34931 ms · 2026-06-30T13:36:12.577468+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

24 extracted references · 13 canonical work pages

  1. [1]

    Owen K. Terry. I’m a Student. You Have No Idea How Much We’re Using ChatGPT.https://www.chronicle.com/article/ im-a-student-you-have-no-idea-how-much-were-using-chatgpt, May 2023. Last accessed 2026-04-21

    2023

  2. [2]

    James Hiebert and Douglas A. Grouws. The effects of classroom mathematics teach- ing on students’ learning. In Frank K. Lester, editor,Second handbook of research on mathematics teaching and learning, pages 371–404. Information Age Publishing, Charlotte, 2007. ISBN 978-1-59311-589-0

    2007

  3. [3]

    Aras Bozkurt et al. Speculative Futures on ChatGPT and Generative Artificial Intelligence (AI): A Collective Reflection from the Educational Landscape.Asian Journal of Distance Education, 18(1):53–130, 2023. doi: 10.5281/zenodo.7636568

    work page doi:10.5281/zenodo.7636568 2023

  4. [4]

    McCarty, Joel Bren- demuhl, and A

    Jane Southworth, Kati Migliaccio, Joel Glover, Diane Reed, C. McCarty, Joel Bren- demuhl, and A. Thomas. Developing a model for AI Across the curriculum: Trans- forming the higher education landscape via innovation in AI literacy.Computers and Education: Artificial Intelligence, 4:100127, 2023. doi: 10.1016/j.caeai.2023.100127

    work page doi:10.1016/j.caeai.2023.100127 2023

  5. [5]

    Cecilia K. Y. Chan. A comprehensive AI policy education framework for university teaching and learning.International Journal of Educational Technology in Higher Education, 20(1), 2023. doi: 10.1186/s41239-023-00408-3

    work page doi:10.1186/s41239-023-00408-3 2023

  6. [6]

    Jules White, Quchen Fu, Sam Hays, Michael Sandborn, Carlos Olea, Henry Gilbert, Ashraf Elnashar, Jesse Spencer-Smith, and Douglas C. Schmidt. A Prompt Pattern Catalog to Enhance Prompt Engineering with ChatGPT. InProceedings of the 30th Conference on Pattern Languages of Programs, PLoP ’23, USA, 2023. The Hillside Group. ISBN 9781941652190

    2023

  7. [7]

    Enhancing Interaction with Large Language Models: A Cat- alog of Prompt Engineering Techniques

    Aniruddha Girish Pai. Enhancing Interaction with Large Language Models: A Cat- alog of Prompt Engineering Techniques. In2025 International Conference on Com- puting Technologies (ICOCT), pages 1–8, 2025. doi: 10.1109/ICOCT64433.2025. 11118417

    work page doi:10.1109/icoct64433.2025 2025

  8. [8]

    Developing evaluative judgement for a time of generative artificial intelligence.As- sessment & Evaluation in Higher Education, 49(6):893–905, 2024

    Margaret Bearman, Joanna Tai, Phillip Dawson, David Boud, and Rola Ajjawi. Developing evaluative judgement for a time of generative artificial intelligence.As- sessment & Evaluation in Higher Education, 49(6):893–905, 2024. doi: 10.1080/ 02602938.2024.2335321

    work page arXiv 2024

  9. [9]

    Designing Learning, Teaching and Assessment (DeLTA) process.www.sun.ac.za

    Centre for Teaching and Learning. Designing Learning, Teaching and Assessment (DeLTA) process.www.sun.ac.za. URLhttp: 11 //www.sun.ac.za/english/learning-teaching/ctl/t-l-resources/ design-for-learning-teaching-and-assessment-(delta)-cycle

  10. [10]

    Vygotsky.Mind in society: The development of higher psychological processes

    Lev S. Vygotsky.Mind in society: The development of higher psychological processes. Harvard University Press, Cambridge, MA, 1978. ISBN 978-0674576292

    1978

  11. [11]

    University students are using AI to write essays

    Katy Quach. University students are using AI to write essays. Now what?https: //www.theregister.com/2022/12/27/university_ai_essays_students/, 2022. Last accessed 2026-04-21

    2022

  12. [12]

    Devel- oping evaluative judgement: enabling students to make decisions about the quality of work.Higher Education, 76(3):467–481, 2018

    Joanna Tai, Rola Ajjawi, David Boud, Phillip Dawson, and Ernesto Panadero. Devel- oping evaluative judgement: enabling students to make decisions about the quality of work.Higher Education, 76(3):467–481, 2018. doi: 10.1007/s10734-017-0220-3

    work page doi:10.1007/s10734-017-0220-3 2018

  13. [13]

    Lodge, Shadi S

    Jason M. Lodge, Shadi S. Alhadad, and Michael J. Lewis. When the black box is a black box: A statement on the use of AI in education.Australasian Journal of Educational Technology, 39(4):1–8, 2023. doi: 10.14742/ajet.8837

    work page doi:10.14742/ajet.8837 2023

  14. [14]

    Schmidt, Jesse Spencer-Smith, Quchen Fu, and Jules White

    Douglas C. Schmidt, Jesse Spencer-Smith, Quchen Fu, and Jules White. Towards a Catalog of Prompt Patterns to Enhance the Discipline of Prompt Engineering.Ada Lett., 43(2):43–51, June 2024. ISSN 1094-3641. doi: 10.1145/3672359.3672364. URL https://doi.org/10.1145/3672359.3672364

    work page doi:10.1145/3672359.3672364 2024

  15. [15]

    Prentice Hall, New Jersey, NY, 1995

    Erich Gamma, Richard Helm, Ralph Johnson, and John Vlissides.Design patterns: Elements of reusable object-oriented Software. Prentice Hall, New Jersey, NY, 1995. ISBN 978-0201633610

    1995

  16. [16]

    Boyuan Jia, Pu Yan, and Yicheng Liu. What makes a good prompter? insights into prompt literacy across mind, experience, and culture.Proceedings of the Association for Information Science and Technology, 62(1):931–935, October 2025. doi: 10.1002/ pra2.1316. URLhttps://doi.org/10.1002/pra2.1316

    work page doi:10.1002/pra2.1316 2025

  17. [17]

    Hevner, Salvatore T

    Alan R. Hevner, Salvatore T. March, Jinsoo Park, and Sudha Ram. Design science in information systems research.MIS Quarterly, 28(1):75–105, 2004

    2004

  18. [18]

    Springer,

    Paul Johannesson and Erik Perjons.An Introduction to Design Science. Springer,

  19. [19]

    J. Biggs. Enhancing teaching through constructive alignment.Higher Education, 32 (3):347–364, 1996. doi: 10.1007/bf00138871

    work page doi:10.1007/bf00138871 1996

  20. [20]

    The 3R’s (Rethink, Redesign, Reassess): A framework for adapting your teaching practice.https://www.sun.ac.za/english/ learning-teaching/ctl, 2023

    Centre for Teaching and Learning. The 3R’s (Rethink, Redesign, Reassess): A framework for adapting your teaching practice.https://www.sun.ac.za/english/ learning-teaching/ctl, 2023. Last accessed 2026-04-21

    2023

  21. [21]

    Punya Mishra and Matthew J. Koehler. Technological Pedagogical Content Knowl- edge: A Framework for Teacher Knowledge.Teachers College Record, 108(6):1017– 1054, 2006. doi: 10.1111/j.1467-9620.2006.00684.x

    work page doi:10.1111/j.1467-9620.2006.00684.x 2006

  22. [22]

    Assigning AI: Seven Approaches for Students, with Prompts

    Ethan Mollick and Lilach Mollick. Assigning AI: Seven Approaches for Students, with Prompts. The Wharton School Research Paper, 2023. 12

    2023

  23. [23]

    Owura Asare, Meiyappan Nagappan, and N. Asokan. Is GitHub’s Copilot as bad as humans at introducing vulnerabilities in code?Empirical Software Engineering, 28(6), September 2023. ISSN 1382-3256. doi: 10.1007/s10664-023-10380-1. URL https://doi.org/10.1007/s10664-023-10380-1

    work page doi:10.1007/s10664-023-10380-1 2023

  24. [24]

    Asleep at the keyboard? assessing the security of github copilot’s code contributions,

    Hammond Pearce, Baleegh Ahmad, Benjamin Tan, Brendan Dolan-Gavitt, and Ramesh Karri. Asleep at the Keyboard? Assessing the Security of GitHub Copi- lot’s Code Contributions.Communications of the ACM, 68(2):96–105, January 2025. ISSN 0001-0782. doi: 10.1145/3610721. URLhttps://doi.org/10.1145/3610721. 13

    work page doi:10.1145/3610721 2025