Structured autonomy: bounding generative agent behavior through layered pedagogical and domain constraints¶
Design · All designs
Evidence · 2 claims (2 for) · 1 study (1 design),q1· 0 of 1 report an effect size · 2 claims rest on one study
Description¶
CyberAGENTS is a layered, ontology-guided multi-agent framework for gamified cybersecurity learning built on the principle of structured autonomy. The article states it "enables structured autonomy through ontology-guided validation, schema-governed behavioral control, and competency-based progression." Three layers regulate the system: a competency graph initializes gameplay, behavioral schemas bound each agent's instructional role, and a cybersecurity ontology validates all generated content before display. The design principle is that "generative flexibility should be bounded through structured autonomy rather than eliminated through static scripting."
Design Implications¶
Context¶
Requirements¶
- A central orchestrator maintaining shared state across turns, including topic, difficulty, hint usage, feedback history, and XP trajectory
- Behavioral schemas encoding operational modes, trigger conditions, progression logic, and evaluation criteria for each agent
- A domain ontology defining valid entity types and permissible relations for pre-display validation
Constraints¶
- Evaluated only in cybersecurity education at novice difficulty with short 5-15 minute sessions
- The authors state the small-scale deployment and preliminary ablation limit broader claims about long-term learning outcomes
Target Learners¶
- Undergraduate cybersecurity students
Learning Goals¶
- Cybersecurity reasoning about attacks, defenses, vulnerabilities, and tools
- Sustained engagement and iterative skill-building in practice-driven technical domains
Claims¶
- Ablating expert-informed structural components lowered LLM-judge scores on all dimensions, with largest drops in challenge quality, intent alignment, and domain grounding [+W]
- LLM-as-judge evaluation scored CyberAGENTS highest on domain grounding (4.46) and agent role fidelity (4.35) [+W]
Related Designs¶
- AISecKG ontology as a pre-display validation layer for agent-generated content
- Competency-mapping module organizing curriculum topics by difficulty, prerequisites, and concept clusters
Examples¶
-
Key Sources¶
- Hornung, I., Marasinghe Arachchige, D., Kumarage, T., Agrawal, G., Deng, Y., Chen, Y.-C., & Liu, H. (2026). CyberAGENTS: Structured Autonomy for Agentic Gamified Learning in Cybersecurity. Preprint. https://arxiv.org/abs/2608.07965. doi:10.48550/arxiv.2608.07965
Page metadata
| Field | Value |
|---|---|
type |
design |
id |
structured-autonomy-agentic-learning-framework |
title |
Structured autonomy: bounding generative agent behavior through layered pedagogical and domain constraints |
description |
CyberAGENTS is a layered, ontology-guided multi-agent framework for gamified cybersecurity learning built on the principle of structured autonomy. |
status |
review |
generated |
by process:settle-candidates on 2026-10-10 |
trust |
unverified — no one has confirmed this page's content |
sources |
1 — hornung-2026 |
Frontmatter is the page's machine-readable record — see the Schema & Guide.