Repolode
제출하기
Egonex-AI

Egonex-AI/Understand-Anything

A Claude Code plugin and multi-platform skill set that turns codebases and LLM wikis into interactive knowledge graphs you can explore and query.

GitHub에서 보기웹사이트
스타
86,937
포크
7,331
열린 이슈
330
마지막 커밋
7일 전

어떤 프로젝트인가요?

Understand Anything analyzes a project with a multi-agent pipeline and builds a knowledge graph of its files, functions, classes, and dependencies. The graph is shown in an interactive web dashboard where nodes can be clicked, searched, and explained in plain English, and guided tours order the architecture by dependency.

Structural facts such as imports, exports, definitions, and call sites are extracted deterministically with Tree-sitter, while an LLM adds summaries, tags, architectural layers, and business-domain mapping. Later runs are incremental, re-analyzing only changed files. Commands cover chat questions about the code, diff impact analysis, file-level explanations, onboarding guides, and business domain extraction.

The same pipeline can analyze a Karpathy-style LLM wiki, extracting wikilinks, entities, and claims into a force-directed graph. The project supports Claude Code natively and installs into Codex, Cursor, VS Code Copilot, Gemini CLI, OpenCode, and several other agent platforms. Graph output can be committed to a repository and viewed later with a read-only local viewer that makes no LLM calls.

핵심 기능

  • 01Interactive graph of files, functions, classes, and dependencies with search and node details.
  • 02Guided tours that walk through architecture in dependency order.
  • 03Diff impact analysis showing which parts of the system a change affects.
  • 04Layer grouping (API, Service, Data, UI, Utility) with a color-coded legend.
  • 05Incremental re-analysis so later runs only process changed files.
  • 06Analysis of Karpathy-pattern LLM wiki knowledge bases into a navigable graph.
  • 07Output localized to several languages, including Simplified Chinese.

활용 사례

  • Getting oriented in a large unfamiliar codebase during onboarding.
  • Reviewing which modules a pending change touches before committing.
  • Generating a shared, committed architecture map for a team to browse.
  • Turning a personal Markdown or wiki knowledge base into a linked graph.

토픽

#knowledge-graph#codebase-analysis#claude-code#code-understanding#developer-tools#llm-agents#visualization#onboarding#antigravity-skills#business-knowledge

자주 묻는 질문

Understand-Anything은(는) 무엇인가요?
A Claude Code plugin and multi-platform skill set that turns codebases and LLM wikis into interactive knowledge graphs you can explore and query. Understand Anything analyzes a project with a multi-agent pipeline and builds a knowledge graph of its files, functions, classes, and dependencies. The graph is shown in an interactive web dashboard where nodes can be clicked, searched, and explained in plain English, and guided tours order the architecture by dependency.
Understand-Anything은(는) 무료로 사용할 수 있나요? 라이선스는 무엇인가요?
Understand-Anything은(는) MIT 라이선스의 오픈소스입니다. 상업용 제품에 사용하기 전에 라이선스 조건을 확인하세요.
Understand-Anything은(는) 아직 유지보수되고 있나요?
활발히 유지보수 중. Understand-Anything의 마지막 커밋은 2026년 10월 12일입니다.
Understand-Anything은(는) 어디에 사용할 수 있나요?
Getting oriented in a large unfamiliar codebase during onboarding. Reviewing which modules a pending change touches before committing. Generating a shared, committed architecture map for a team to browse. Turning a personal Markdown or wiki knowledge base into a linked graph.
Understand-Anything은(는) 어디서 받을 수 있나요?
소스 코드와 설치 방법은 GitHub에 있습니다: https://github.com/Egonex-AI/Understand-Anything

요약은 프로젝트 README를 바탕으로 AI가 작성했습니다. GitHub 통계는 매일 갱신됩니다.

관련 프로젝트