# Simplify > Use when a file or module has grown too complex — removes unnecessary abstractions, dead code, and over-engineering without changing behavior. Invoke after implementing a feature when the code feels over-engineered. - Skill: `latuconsinafr/simplify` (Agent Skill) - Install (CLI): `npx skillmds@latest add latuconsinafr/simplify` - Raw SKILL.md: https://api.skillmd.com/api/skills/latuconsinafr/simplify/raw - Safety review: pending - Works with: Claude Code, Claude.ai, OpenAI Codex - Category: Coding & Dev Tools - Author: latuconsinafr (https://skillmd.com/u/latuconsinafr) - Updated: 2026-09-17 - Page: https://skillmd.com/skills/latuconsinafr/simplify --- # Simplify Simplify the code in `$ARGUMENTS` (file or module path). If no argument — simplify the most recently edited file. ## Guiding principle > The best code is code that doesn't exist. The second best is code that's obvious. ## Spawn the `refactoring` agent The `refactoring` agent handles the full simplification workflow: test baseline verification, blast radius checks, identifying complexity sources, proposing and applying changes one at a time, and re-running tests after each step. Spawn it with the following task: ``` Target: Goal: simplify — remove unnecessary abstractions, dead code, and over-engineering without changing behavior Constraint: behavior must be identical before and after each change ``` The agent will: 1. Verify the test suite is green before starting 2. Warn if on `main`/`master` branch 3. Check blast radius for every symbol it considers removing 4. Propose changes grouped by risk (safe / needs verification / out of scope) 5. Ask for confirmation before applying 6. Apply and re-run tests after each change, reverting on failure ## After the agent completes Review the agent's summary: - All changes made and tests verified - Any blocked changes (failed tests on attempt) — note these for a separate follow-up - Deferred items (bugs or features noticed but not touched during refactor)