Karpathy LLM Wiki vs Literature Review — Which AI Skill Should You Use?

The short version

By ClawHub Community

Build a persistent knowledge base that gets smarter every time you add a source or ask a question.

By Mohit Agarwal (PM Skills)

Turn a pile of research papers into a structured, critical review with themes and gaps.

Reads your sources and creates structured markdown wiki pages with summaries, key facts, and cross-references
Updates existing pages when new sources add information, and flags contradictions between sources
Answers questions by synthesizing across your wiki pages with inline citations
Runs health checks to find contradictions, orphaned pages, missing cross-references, and gaps in your knowledge base
Maintains a master index and operation log so you always know what's in your wiki and how it evolved

Organizes research evidence by theme instead of summarizing one paper at a time
Synthesizes findings across multiple studies and highlights where they agree or conflict
Critically analyzes study quality, methodology, and bias for key sources
Identifies gaps in the research and what the field still needs to explore
Formats output for academic journals, dissertations, grant proposals, or policy briefs

You're researching a topic over weeks or months and want to compound your learning instead of re-reading
You need to track information from multiple sources and spot where they agree or contradict
You're building expertise in a domain and want a searchable, structured knowledge base that grows with you
You want your AI to remember what you've already processed and build on it, not start from scratch each time

You're writing a dissertation or thesis and need a structured background chapter
You're preparing a grant proposal and need to show what's known and what's missing
You're building the research foundation for a new product or policy initiative
You have a stack of research papers and need to make sense of what they say collectively