SuperAPI AI Directory

source-code-analysis

Coding & development

View repository

Updated 10-15 20:04

About this skill

Source Code Analysis

A skill that systematically analyzes a software project Outside-In.

Core Principles

  1. Outside-In — Metadata/dependencies first → structure → internal logic order
  2. Follow the Data — Trace the main input→output data pipeline
  3. Ask Before Diving — After Phase 2 is complete, always confirm the focus area with the user // !!!

Mode Selection

Mode Time Required Phase Purpose
Quick Scan 5-10분 1-2 Understanding the project overview
Standard 20-30분 1-5 Preparing for rebuild/contribution
Full Audit 60분+ 1-7 Including security/quality audit

Default: Standard. If the user does not specify otherwise, Standard applies.

7 Phase Workflow

Phase 1: Surface Scan

Identify project metadata, dependencies, entry points, and build system.

읽을 파일: package.json, tsconfig.json, Cargo.toml, pyproject.toml, go.mod, Dockerfile 등

Deliverable: Project Identity Card (language, framework, build, scale, entry point)

Detailed checklist per Phase: phase-checklists.md

Phase 2: Architecture Mapping

Identify the directory→module graph, design patterns, and DI structure.

1. 디렉토리 트리 스캔 (src/ 기준 2단계)
2. 주요 모듈/레이어 식별
3. DI/IoC 패턴 확인
4. Mermaid 아키텍처 다이어그램 생성

Deliverable: Architecture Diagram (Mermaid) + layer description

// !!! STOP after Phase 2 is complete — Request the user to select a focus area. "I have understood the architecture. Which area would you like to analyze in depth?" Present options: Core pipeline / Specific module / External integration / Security

Phase 3: Core Pipeline Tracing

Trace the main data flow (input→processing→output).

1. 엔트리포인트에서 시작
2. 요청/데이터가 흐르는 경로 추적
3. 미들웨어, 서비스, DB 호출 순서 매핑
4. Mermaid 시퀀스/플로우 다이어그램 생성

Parallel Analysis: If the pipeline has multiple branches, deploy Explore subagents per branch.

Agent Role
A Main pipeline (most frequent path)
B Auxiliary pipeline or special handling
C Error/fallback path

Deliverable: Core Data Flow Diagram (Mermaid)

Phase 4: Deep Dive

Analyze the core components in detail.

각 컴포넌트별:
- 책임 (단일 문장)
- 주요 메서드/함수 시그니처
- 상태 관리 방식
- 에러 처리 전략
- 설계 패턴 (Strategy, Factory, Observer 등)

Deliverable: Component Cards (table)

Phase 5: Dependencies & Integration

Map external services and 3rd party libraries.

분류:
- 필수 vs 선택
- 대체 가능 vs 불가
- 유료 vs 무료
- 자체 구축 가능 여부

Deliverable: Dependency Map + replacement options (when rebuilding)

Phase 6: Security & Quality (Full Audit만)

Scan for anti-patterns, vulnerabilities, and technical debt.

- OWASP Top 10 체크
- 하드코딩된 시크릿
- SQL/명령어 인젝션 가능성
- 인증/인가 로직 검증
- 기술 부채 패턴 (TODO, FIXME, 복잡도)

Deliverable: Security & Quality Report

Phase 7: Output Generation

Generate the analysis results as a learning note.

note-writer 스킬 연동:
1. 분석 결과를 note-writer 템플릿에 매핑
2. 00-notes/technology/{project-name}-analysis/ 에 생성
3. references/ 하위에 상세 문서 분리
4. 관련 기존 노트와 wikilink 연결

Note structure: output-template.md

Cross-Skill Integration

Situation Integrated Skill
Decomposing complex architecture decompose-recompose
Verifying design assumptions first-principles
Generating the final note note-writer

Language-Specific Reference

Analysis heuristics per language/framework: language-patterns.md