What is critically-reviewing-literature?

This repository is a skill package for AI coding assistants such as Cursor, designed to help researchers write critical literature reviews. It is based on a method from a 2019 academic paper by Fernandez that teaches new researchers how to review a body of academic work rather than just summarizing papers one by one. The skill targets common problems in weak literature reviews: long lists of citations that read like a phone book, summaries stacked together without synthesis, description without critique, citing papers secondhand without reading them, claiming contributions or gaps without evidence, giving up too early when a topic seems unstudied instead of adapting nearby research, and research questions that are too broad to actually answer with a feasible study design. To fix these problems, the skill walks an AI agent through four required stages: assemble the relevant literature, arrange it into categories and comparison tables, assess it critically against the standards of its academic field, and only then write the review text. The process builds in five kinds of outcomes a good review should achieve, such as showing what is already known, identifying real gaps, and generating an answerable research question. Each stage has required outputs, and the agent is not allowed to skip ahead, for example it cannot write the review body before completing the comparison tables, and it cannot claim a contribution or gap without finishing the assessment step first. This is meant for writing or revising literature review sections of a thesis, dissertation, or standalone review paper, for finding real gaps in existing research, and for turning an old, list like review into one organized around concepts. It is not meant for simple grammar polishing, general research methods coaching unrelated to literature, or downloading full text from paid databases. Installation means copying the skill folder into a personal skills directory used by Cursor or a similar agent client, or uploading the packaged zip file directly if the client supports it. Once installed, mentioning a literature review task in a new conversation triggers the skill automatically. The package includes a main instructions file, a template file with search plans and writing patterns, and this README, and it credits the original Fernandez 2019 paper without copying its copyrighted text.