Agent/Skill for Literature Reviews (github.com)

A new initiative has introduced a set of high-precision, PRISMA-compliant prompts designed to enhance the utility of Large Language Models (LLMs) as collaborative tools for conducting literature reviews. These specialized prompts aim to circumvent "The Architect’s Paradox," a common issue where LLMs generate vague summaries instead of detailed methodological frameworks. With prompts focusing on various aspects of the literature review process—such as search strategies, theoretical mapping, quality appraisal, and manuscript preparation—this tool provides researchers with the structured support necessary to ensure their findings are rigorous, defensible, and ready for publication. Significantly, this approach addresses four critical flaws in AI-assisted research by tailoring prompts to specific research databases, distinguishing between procedural guidance and execution, and ensuring outputs comply with international reporting standards. The modular design encourages a stepwise application of methods, starting from foundational elements to advanced synthesis strategies. Researchers can also customize the prompts to align with their unique needs, promoting a more practical interaction with AI in academic research. By streamlining the literature review process, these prompts not only increase efficiency but also improve the quality and transparency of scholarly communication in the AI/ML community.