Analyzing Codebases
Runs real static-analysis tooling and ranks refactor hotspots by churn x complexity.
Funciona con configuración
Qué hace
Detects a project's languages and monorepo shape, runs the matching static-analysis toolchain (dependency graph, duplication, cognitive complexity, semantic patterns), and writes a machine-parseable refactor map ranking the top 20 hotspots by git churn times complexity. Triggers on /aref or an explicit ask to analyze a codebase before refactoring.
Informe de la prueba
Ran the real toolchain on a throwaway TS repo: jscpd correctly caught an exact-clone duplication cluster and git-churn math correctly ranked the touched-and-duplicated file above its untouched twin (score 96 vs 16, both cognitive complexity 16). But the reference doc's literal commands for dependency-cruiser and eslint-plugin-sonarjs both failed out of the box (needed --no-config plus --include-only, and a manually added @typescript-eslint/parser) — friction not mentioned anywhere in the skill.
Probado el: 2026-07-15 · Claude Code 2.x (agent harness)
Instalación
git clone https://github.com/wayne930242/Reflexive-Claude-Code mkdir -p ~/.claude/skills cp -r Reflexive-Claude-Code/plugins/aref/skills/analyzing-codebases ~/.claude/skills/analyzing-codebases
Comandos y prompts de ejemplo
/analyzing-codebasesRuns real static-analysis tooling and ranks refactor hotspots by churn x complexity.
Los skills se activan con peticiones en lenguaje natural, sin comandos que memorizar. Tras instalarlo, prompts como estos lo activan (en inglés):
Analyze this codebase and tell me what needs refactoring firstGive me a hotspot map of the messiest, most complex files before we refactorWhich files here have the highest git churn combined with high complexity?