Advanced codebase analysis with HTML reports, git-aware diffs, trend tracking, SVG badges, CSV export, and CI/CD integration
Coding
Codebase Radar
Try itScan codebase for dependency graph, tech debt hotspots, and module health scores
What it does
Scan codebase for dependency graph, tech debt hotspots, and module health scores
The skill document
Codebase Radar
Full-spectrum static analysis for any codebase: dependency graphs, module health scoring, tech debt detection, dead code, and visualization in Mermaid/D2.
Workflow
- Project scan — Tree view, line counts, language distribution, file size outliers.
- Dependency graph — Resolve imports/requires across modules; detect circular dependencies; list external deps with versions.
- Cohesion & coupling — Score each module: how focused its internals are (cohesion) and how tightly bound to others (coupling).
- Tech debt signals — Scan for TODO/FIXME/HACK density, outdated dependency alerts, low comment-to-code ratio.
- Dead code detection — Find unreferenced functions, classes, variables; flag large commented-out blocks.
- Health scorecard — 5-dimension score per module: Cohesion, Coupling, Maintainability, Coverage estimate, Tech debt.
- Visualization — Generate dependency diagram in Mermaid (
graph TD) or D2 format. - Report — Produce project health summary with ranked refactoring priorities.
Sample Prompts
codebase-radar scan --path ./my-project --format json > report.jsoncodebase-radar scan --path ./my-project --visual mermaid --output deps.mdcodebase-radar health --path ./my-project --score-only --jsoncodebase-radar deadcode --path ./src --exclude tests/
Safety
- Read-only scanner; never modifies source files.
- Language-agnostic for import detection; language-specific analyzers are opt-in.
- Large repos may require
--max-fileslimit to avoid memory issues.
Related skills
Generate visual HTML pages for code review (diff + risk tags), code walkthrough (call chains + trust boundaries), and architecture analysis (module dependencies + tech debt). Three modes with shared dark-theme template.
Study unfamiliar codebases and produce evidence-backed knowledge artifacts. Use for repository orientation, architecture mapping, subsystem tracing, onboarding, or codebase documentation.
Scan AI-generated code for bugs before deploying — 8 static analysis checks from critical (hardcoded secrets, unsafe eval) to low (unused imports). Productio...
代码库索引与理解系统。扫描项目目录,提取所有符号定义(类、函数、变量、导入),构建可搜索的 JSON 索引,支持按名称/类型/文件查询。受 Claude Code 的 codebase indexing 启发,100% 原创实现,使用 Python ast + ripgrep。
Privacy-first Spring Boot code health diagnosis. Detects security vulnerabilities, performance anti-patterns, code quality issues, and dependency risks acros...