Appendix A. Prompt Library

General explanation

Explain [topic] for [audience]. Start with a plain-language definition, then explain the key mechanisms, give a worked example, identify common misconceptions, and list points that require verification. Do not invent citations.

Compare tools

Compare [tool A], [tool B], and [tool C] for [task]. Use current official documentation where available. Compare capabilities, limitations, privacy considerations, cost factors, accessibility, and learning curve. Separate verified facts from your assessment and flag anything that may have changed.

Research question refinement

Review this draft research question: [question]. Identify ambiguous terms, hidden assumptions, scope problems, and possible constructs. Offer three alternatives and explain the trade-offs. Do not invent evidence or claim that a gap exists without a literature search.

Paper extraction

Using only the attached paper, extract the research aim, theoretical framework, design, sample, data collection, analysis, findings, limitations, and relevance to [topic]. If a detail is absent, write ‘not reported’. Provide page references for quotations.

Literature synthesis

Compare the supplied sources on their research questions, methods, contexts, findings, and limitations. Identify convergence, disagreement, and gaps. Cite each claim to the supplied source and distinguish authors’ findings from your interpretation.

Academic editing

Edit the following text for clarity, coherence, grammar, and academic tone. Preserve the meaning, data, quotations, and technical terminology. Do not add claims or references. After the revision, list substantive changes and uncertainties.

Doctoral chapter review

Act as a critical doctoral-level reviewer. Assess alignment between the research questions, theory, methods, findings, and conclusions. Identify the most important weaknesses with section-specific evidence and propose a prioritised revision plan. Do not rewrite the entire chapter.

Data analysis plan

Given this research question, design, sample, and variable dictionary [insert], suggest suitable analyses and assumptions to check. Explain what each analysis can and cannot establish. Do not assume causality or invent results.

Code review

Review this code for correctness, security, maintainability, and edge cases. Explain each proposed change, identify assumptions, suggest tests, and do not execute destructive actions. Treat all generated code as untrusted until reviewed.

Business strategy

Develop a preliminary strategy analysis for [organisation/context] using the evidence supplied. Separate facts, assumptions, hypotheses, and recommendations. Identify missing information, counterarguments, risks, and measurable indicators. Do not fabricate market statistics.

Lesson plan

Create a lesson plan for [learner group] on [topic], aligned with these learning outcomes [outcomes]. Include sequence, activities, formative checks, accessibility adaptations, and assessment criteria. Flag factual claims for teacher verification.

Meeting summary

From the supplied notes, separate decisions, action items, owners, deadlines, open questions, and background. Do not infer agreement where none is recorded. Mark unclear names or dates for confirmation.

Translation review

Translate the following text into [language] for [audience and context]. Preserve technical terms and register. Flag ambiguous idioms or phrases where multiple translations are possible. This is a draft for human review, not certified translation.

Risk review

Evaluate this proposed AI use case. Identify data sensitivity, potential harms, affected people, bias risks, security risks, reversibility, oversight requirements, and verification steps. Recommend controls proportionate to the risk.

Source verification

List every factual claim and reference in the supplied draft that needs verification. For each item, state what evidence would confirm it and where an authoritative source might be found. Do not claim verification unless you have actually inspected the source.