Project Analysis Skill

Perform comprehensive analysis of a project codebase to understand its structure, architecture, key files, and dependencies.

When to Use

Activate when:

• User asks "analyze this project" or "what does this codebase look like?"
• Starting work on an unfamiliar codebase
• Reviewing project architecture before making changes
• Generating documentation about project structure
• Understanding dependency landscape

When NOT to use: For questions about a single file or feature, just read the relevant file directly — a full analysis is overkill.

Analysis Workflow

Step 1: Survey the Codebase

Build a high-level overview using standard filesystem operations:

1. List the directory tree with ls or Glob to see top-level structure
2. Read manifest files to identify the stack: package.json, go.mod, pyproject.toml, Cargo.toml, Gemfile, pom.xml, etc.
3. If useful, count files by extension to gauge language distribution, e.g. find . -name '*.ts' -not -path '*/node_modules/*' | wc -l

Step 2: Understand Project Type

Based on what the survey found, identify the project type:

Step 3: Read Key Files

Examine critical files:

Always check:

• README.md - Project purpose and documentation
• package.json / Cargo.toml / pyproject.toml - Dependencies and metadata

Check if they exist:

• tsconfig.json / vite.config.ts / webpack.config.js - Build configuration
• Dockerfile / docker-compose.yml - Container setup
• .env.example - Environment variables
• CONTRIBUTING.md - Development workflow

Step 4: Explore Architecture

List key directories (src/, lib/, app/, etc.) to understand component organization, then identify architectural patterns:

Step 5: Analyze Dependencies

Read the dependency file to understand:

• Runtime dependencies: Core libraries the project relies on
• Dev dependencies: Build tools, test frameworks, linters
• Package manager: npm, yarn, pnpm, cargo, pip, etc.
• Dependency count: Overall complexity indicator

Analysis Depth

Scale depth to the request — quick = structure + project type only; deep = all steps plus reading entry points and core modules.

Output Format

Present analysis in this structure:

## Project Analysis: [Name]

### Overview
- **Type**: [Web app / API / CLI / Library / etc.]
- **Primary Language**: [language] (X%)
- **Total Files**: X

### Architecture
- **Pattern**: [MVC / Layered / Feature-based / etc.]
- **Key Directories**:
  - `src/commands/` - CLI command implementations
  - `src/services/` - Business logic layer

### Key Files
| File | Purpose |
|------|---------|
| `index.ts` | Application entry point |
| `package.json` | Dependencies and scripts |

### Dependencies
- **Package Manager**: [npm/yarn/pnpm]
- **Runtime** (X packages): [key ones listed]
- **Development** (Y packages): [key ones listed]

### Observations
- [Notable patterns or concerns]
- [Suggested improvements]

Best Practices

1. Be specific: Name actual files and directories, don't generalize
2. Note concerns: Flag large files, missing docs, or unusual patterns

Common Mistakes

• Don't guess architecture from file names alone — read key files before concluding
• Don't include vendored/generated directories (node_modules, dist, target) in counts or structure