Repo Analyzer skill
A skill that dissects an open-source project's architecture and returns one Markdown report with Mermaid diagrams and a comparison with rivals
- 499 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- The skill clones repositories and saves reports in the home folder
- It goes online for reviews, comparisons and the project's website
- Parallel subagents start automatically on large projects
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/repo-analyzerDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- The skill clones repositories and saves reports in the home folder
- It goes online for reviews, comparisons and the project's website
- Parallel subagents start automatically on large projects
This is third-party code. Review the repository files before installing.
What it does
Repo Analyzer is a skill for dissecting the architecture of open-source projects. Ask the agent to analyze `owner/repo`, a GitHub, GitLab or Gitee URL or a local path, and the skill assembles a report on how the project is built. It concentrates on the reasons behind decisions: which problem the project solves, how it differs from alternatives and what trade-offs sit in its modules. The work runs in stages. The skill clones the repository and counts effective lines per module, searches the web for reviews and comparisons, asks targeted questions about the project's traits and designs the report chapters itself. For large codebases it then launches subagents on the core modules, cross-checks conclusions between modules and verifies how much code was read. You choose depth from three modes: Quick covers at least 30% of core modules, Standard (the default) at least 60%, Deep at least 90%. The result is saved to `~/repo-analyses/{project-name}-{date}/ANALYSIS_REPORT.md` with Mermaid diagrams. The report is written in Chinese by default and follows your language when you ask in another.
Who it is for. Team leads and engineers who study someone else's codebase before choosing a library or framework, and authors of technical reviews.
Good fit when
- Before choosing between two frameworks you want a breakdown of their design decisions, and the skill accepts a request to compare two projects
- You are about to read a large repository such as `ollama/ollama`, and a report with diagrams explains its structure before you dive into code
- An author prepares an article on how an open-source project works and gathers material with an assessment of strengths and weaknesses
- A quick overview is enough: Quick mode covers at least 30% of core modules
Not a fit when
- You need to chase a single bug or debug a function: the skill works at the architecture level
- You want a review of one file or a pull request
- You plan a code change that does not touch the system's structure
- A Russian or English request does not trigger it: the auto-trigger keywords in its description are Chinese, so invoke it by name
Example request
Analyze the architecture of astral-sh/ruff in Standard mode and save the report with diagramsLimitations
The last commit in the repository dates from April 2026. The report comes out in Chinese by default, and the auto-trigger keywords in the skill's description are Chinese, so from a Russian or English request it is safer to call the skill by name. Parallel subagents in Standard and Deep modes noticeably increase the number of model calls. For richer web research the author suggests Tavily MCP or Brave Search MCP; without them the skill uses the agent's built-in tools.
How to disable. Delete the `~/.claude/skills/repo-analyzer` folder, or run `/plugin uninstall repo-analyzer@repo-analyzer` if you installed through the plugin marketplace. Finished reports stay in `~/repo-analyses/`.
FAQ
How do I install Repo Analyzer in Claude Code?
Run `npx skills add yzddmr6/repo-analyzer`. The plugin marketplace route also works: `/plugin marketplace add yzddmr6/repo-analyzer`, then `/plugin install repo-analyzer@repo-analyzer`. For a manual install clone the repository into `~/.claude/skills/repo-analyzer`.
Can it analyze a private repository?
The skill accepts a local path. Clone the private repository yourself and pass the folder path, so no network access to the code is needed. Searches for reviews and comparisons still go to public sources.
What does the finished report contain?
The problem context, positioning among alternatives, a project overview, a deep module analysis with trade-offs, an assessment of strengths and weaknesses, and Mermaid diagrams. The chapters change per project; the report has no fixed template.
Does the skill paste project code into the report?
Not by default. It describes flows, logic and ideas at the design level and pastes raw code only when needed, preferring diagrams and tables.
How do I compare two projects?
Ask for the comparison in one request, for example `compare express and fastify`. The skill analyzes both and highlights differences in design philosophy instead of lining up feature lists.
Discussion
No questions or reviews yet.
Related
SkillsGives the agent an understanding of your repository: it indexes the code and answers questions about it in Claude Code, Cursor, Codex, Windsurf and more
kyy-logs code-review-skill
code-review-skill
A code review skill with four independent probes, evidence-backed findings and a merge gate derived from JSON
A code-review CLI and plugin for agents: reads a git diff and writes line-level comments with context
A Claude Code plugin that extracts behavioral specs, test vectors and acceptance criteria from a codebase, with a provenance trail
Indexes a codebase into a knowledge graph and gives the agent call chains, change blast radius and code search over MCP
A Codex skill for reviewing AI-written code: explains blocks, names files and lines, checks goals against evidence.
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Authentication | Repository stars | Main difference |
|---|---|---|---|---|
| Repo Analyzer (this card) | no | 499 | ||
| Repomix | no | not required | 29,161 | packs a repository into one file for a model but writes no analysis itself |
| Deepwiki MCP | no | not required | 1,389 | fetches ready-made deepwiki.com pages as Markdown, while Repo Analyzer reads the code and writes a report to the chosen depth |
README in short
The description gives three install routes, a feature list, sample requests in Chinese, a table of depth modes and eight working steps from cloning to the final report. It also shows the report's structure, the repository's file tree and the optional search MCP servers.
README badge
Original SKILL.md
---
name: repo-analyzer
description: Use when the user mentions "分析项目"、"分析仓库"、"分析 GitHub"、"项目分析"、"源码分析"、"架构分析"、"代码分析"、"学习这个项目"、"研究这个框架"、"看看这个库怎么实现的"、"对比两个项目"、"项目评测"、"框架评测"
---
# Git 项目深度分析技能
深度分析开源项目并生成专业架构报告。报告是有深度洞察的技术研究，读完后读者能理解业务问题、掌握架构设计、产生自己的思考。
## When to Use
- 分析开源项目的架构和设计
- 对比两个同类项目的设计差异
- 深入研究一个框架或库的实现思路
## When NOT to Use
- 简单的代码问题或调试
- 单文件分析或代码审查
- 不涉及架构层面的代码修改
## 输出语言
默认中文。如果用户使用其他语言提问，则跟随用户语言。
## 核心原则
### 1. 业务视角优先
从"这个项目解决什么问题"出发，不是"这个文件里有什么函数"。
| 不要 | 要 |
|------|-----|
| `handleRequest(ctx)` 函数接收一个 Context 参数... | 请求进来后，系统会经过鉴权、限流、路由分发三个阶段... |
| `interface MessageQueue { push(); pop() }` | 模块间通过消息队列解耦，生产者只管投递，消费者按优先级拉取 |
### 2. 抽象层次把控：不贴代码，讲设计
默认在设计模式和架构层面描述，**非必要情况下不贴原始代码**。重点突出流程、逻辑、设计思想，用架构图（Mermaid）、流程图、表格来表达，而非代码片段。只有设计特别精妙、项目自创独特概念、或实现是核心卖点时才展示代码，且必须先用自然语言解释。