Structural Quality Gaps in Practitioner AI Governance Prompts: An Empirical Study Using a Five-Principle Evaluation Framework

TL;DR

This study introduces a five-principle framework for evaluating the structural completeness of AI governance prompts, revealing significant gaps in existing practitioner-authored prompts and suggesting automated analysis solutions.

Contribution

It presents a novel, theory-grounded evaluation framework for governance prompts and applies it to real-world GitHub files, uncovering common structural deficiencies.

Findings

37% of prompts scored below the structural completeness threshold

Data classification and assessment rubrics are often missing

Automated static analysis could improve prompt quality

Abstract

AI governance programmes increasingly rely on natural language prompts to constrain and direct AI agent behaviour. These prompts function as executable specifications: they define the agent's mandate, scope, and quality criteria. Despite this role, no systematic framework exists for evaluating whether a governance prompt is structurally complete. We introduce a five-principle evaluation framework grounded in computability theory, proof theory, and Bayesian epistemology, and apply it to an empirical corpus of 34 publicly available AGENTS.md governance files sourced from GitHub. Our evaluation reveals that 37% of evaluated file-model pairs score below the structural completeness threshold, with data classification and assessment rubric criteria most frequently absent. These results suggest that practitioner-authored governance prompts exhibit consistent structural patterns that automated static analysis could detect and remediate. We discuss implications for requirements engineering practice in AI-assisted development contexts, identify a previously undocumented artefact classification gap in the AGENTS.md convention, and propose directions for tool support.