AGENT SKILL

codebase-analysis

Patterns and techniques for analyzing brownfield codebases. Use when onboarding to unfamiliar code, preparing for refactoring, conducting architecture reviews, or identifying technical debt.

License 许可
MIT
最近更新
2026-06-24
Stars
★ 66
Forks
24

01 前言 Front Matter

name
codebase-analysis
description
Patterns and techniques for analyzing brownfield codebases. Use when onboarding to unfamiliar code, preparing for refactoring, conducting architecture reviews, or identifying technical debt.

02 资源文件 6

03 SKILL.md 正文

Codebase Analysis Skill

Systematically explore and understand existing codebases, including entry point discovery, dependency tracing, pattern detection, and technical debt identification.

Variables

Variable Default Description
ANALYSIS_DEPTH standard quick (5min), standard (30min), deep (1hr+)
OUTPUT_FORMAT markdown markdown, json, c4
INCLUDE_DEBT true Include technical debt assessment
INCLUDE_DIAGRAMS true Generate C4 diagrams

Instructions

MANDATORY - Follow the Workflow steps below in order. Do not skip steps.

  • Start with entry points, not random files
  • Trace dependencies before analyzing patterns
  • Document findings as you go

Red Flags - STOP and Reconsider

If you're about to:

  • Analyze random files without finding entry points first
  • Make architectural claims without tracing dependencies
  • Skip pattern detection before identifying anti-patterns
  • Report technical debt without severity classification

STOP -> Read the appropriate cookbook file -> Follow the workflow -> Then proceed

Workflow

  1. [ ] Quick Assessment: Project size, structure, config files
  2. [ ] CHECKPOINT: Identify entry points (read cookbook/entry-point-discovery.md)
  3. [ ] Trace dependencies from entry points
  4. [ ] Detect architectural patterns
  5. [ ] CHECKPOINT: Verify pattern detection before anti-pattern search
  6. [ ] Identify anti-patterns and technical debt
  7. [ ] Document findings with C4 diagrams (if INCLUDE_DIAGRAMS)

Cookbook

Entry Point Discovery

  • IF: Starting analysis of unfamiliar codebase
  • THEN: Read cookbook/entry-point-discovery.md
  • RESULT: List of main entry points by framework

Dependency Tracing

  • IF: Need to understand how modules connect
  • THEN: Read cookbook/dependency-tracing.md
  • RESULT: Import chains, dependency graph

Pattern Detection

  • IF: Need to identify architectural style
  • THEN: Read cookbook/pattern-detection.md
  • RESULT: Identified patterns (MVC, Layered, etc.)

Technical Debt Identification

  • IF: Assessing code quality and debt
  • THEN: Read cookbook/debt-identification.md
  • RESULT: Prioritized debt inventory

Quick Reference

Entry Points by Stack

Stack Look For
Node.js package.json main/bin, index.js, server.js
TypeScript tsconfig.json, src/index.ts
Python __main__.py, app.py, wsgi.py
Go main.go, cmd/*/main.go
React src/index.tsx, pages/_app.tsx

See reference/framework-patterns.md for complete list.

Common Patterns

Pattern Directory Signs
MVC models/, views/, controllers/
Layered presentation/, business/, data/
Hexagonal domain/, ports/, adapters/
Microservices services/*/, packages/*/

Output

Standard Report Format

# Codebase Analysis: [Project Name]

## Overview
- **Size**: X files, Y lines
- **Primary Language**: TypeScript
- **Framework**: Next.js
- **Architecture**: [Pattern]

## Entry Points
1. `src/pages/_app.tsx` - Application root
2. `src/api/` - API routes

## Key Dependencies
- [External deps list]

## Architecture Diagram
[C4 diagram if INCLUDE_DIAGRAMS]

## Technical Debt (if INCLUDE_DEBT)
| Issue | Severity | Location |
|-------|----------|----------|
| ... | High | ... |

04 同仓库更多 Skills