jinleiphys

Research LLM Wiki — AI skill for Claude Code

AI community

Agent skills that turn your scientific reading and research history into persistent LLM-maintained markdown wikis, not one-shot RAG.

How to install Research LLM Wiki

This entry records only its repository, not the path inside it, so there is no exact command to give. Open jinleiphys/research_LLM_wiki and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Research LLM Wiki does

Agent skills that turn your scientific reading and research history into persistent LLM-maintained markdown wikis, not one-shot RAG. For Claude Code + Codex CLI

Alternatives in AI

WeKnora — Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, an 26.9k ★
ARIS Agent Guide — For AI agents reading this repo 6.2k ★
Open Knowledge — Beautiful, AI-native markdown IDE and LLM wiki 3.7k ★

README

research_LLM_wiki

**English** | [中文](./README.zh.md)

A pair of cross-harness agent skills (**Claude Code** + **Codex CLI**) that turn your scientific reading and your own research history into **persistent, structured, LLM-maintained markdown wikis** instead of one-shot RAG retrievals.

Two skills, one idea. Use either alone, both together for full effect.

Skill	What it tracks	Auto-loads?
literature-wiki	Field-level knowledge: every paper you read, with cross-cutting entity / method / system / observable / debate pages	No — invoked when ingesting / querying / linting
research-profile	Your own research portfolio: projects, papers, ideas, failures, methods you own, collaborators, talks, reviews, funding	Yes — `profile.md` auto-loads into every new session

The idea

Most LLM-document workflows are RAG: every question re-discovers knowledge from raw PDFs. Nothing accumulates. After reading 50 papers on a topic over six months, your assistant still answers the next question by reading fragments of three random PDFs and stitching them together. The synthesis you spent months building lives only in your head.

These skills implement a different pattern, articulated by [Andrej Karpathy in his "LLM Wiki" gist](https://gist.github.com/karpathy/442a6bf555914893e9891c11519de94f):

Instead of just retrieving from raw documents at query time, the LLM **incrementally builds and maintains a persistent wiki** — a structured, interlinked collection of markdown files that sits between you and the raw sources.

You curate sources and ask questions. The LLM does the bookkeeping: summarizing, cross-referencing, filing, contradiction-spotting. The wiki is the compiled artifact, kept current as new sources arrive, never rediscovered from scratch.

This repository applies that pattern to scientific research, in two complementary scopes.