Skip to content
Conduit

skill

codex-simplify-skill

150 lines that make Codex a senior refactoring engineer. Process + guardrails, not knowledge — because LLMs already know how to refactor.

tavaresgmg/codex-simplify-skillcommunitylistedunreviewed
Use skill

Set up

Skills are installed by placing their folder where your agent loads skills.

Download ZIP from GitHub Open the repository

The download comes from github.com/tavaresgmg/codex-simplify-skill itself, not from our server, so what you get is exactly what its author published.

Install instructions come from the README

Skills do not use an MCP config. Follow the steps in the README below.

README

synced Oct 22, 2026

$simplify — Code Simplifier Skill for OpenAI Codex

A lean skill that gives Codex a structured refactoring process — not a knowledge dump, but guardrails that keep a smart model honest.

Philosophy

Modern LLMs already know what cyclomatic complexity is, how to extract guard clauses, and what constitutes a code smell. Teaching them refactoring theory is like teaching a fish to swim. What they don't have is:

  • A structured process (7 phases from scope detection to impact report)
  • Guardrails that prevent rationalized bad changes (anti-rationalization checklist)
  • Decision gates that know when to auto-apply vs. ask the user
  • Parallelism awareness (when to spawn agents vs. go sequential)

This skill provides process, not knowledge. 145 lines, 2 files.

Quick start

# Clone into your Codex skills directory
git clone https://github.com/tavaresgmg/codex-simplify-skill.git ~/.codex/skills/simplify

Enable skills in ~/.codex/config.toml:

[features]
skills = true
multi_agent = true  # optional, enables parallel analysis

Use

$simplify                          # Recent changes (git diff)
$simplify src/auth/                # Specific directory
$simplify src/utils/format.ts      # Specific file

Natural language also works:

Simplify the entire repository
Audit src/api/ — find problems but don't fix anything

Modes

ModeTriggerBehavior
Refactor (default)$simplify, "simplify", "refactor", "clean up"Find problems AND fix them
Audit"find problems", "audit", "scan", "what's wrong with"Report only — zero edits

7-Phase Process

  1. Scope — detect target files (user-specified, directory, repo, or git diff)
  2. Context — batch-read project standards, target files, and dependencies
  3. Analysis — structural complexity, code smells, convention consistency
  4. Plan — confidence scoring (0-100) with decision gates
  5. Refactoring — atomic edits in priority order with stop-loss on failures
  6. Verification — mandatory lint + typecheck + tests gate
  7. Report — structured summary with metrics and flagged future items

The 5 Iron Laws

  1. FUNCTIONALITY IS SACRED — Zero behavior changes
  2. CLARITY OVER BREVITY — Explicit beats clever
  3. CONSISTENCY OVER PERFECTION — Follow project patterns, migrate all or none
  4. ATOMIC CHANGES — One concern per edit
  5. EVIDENCE OVER OPINION — Every change needs concrete justification

Skill structure

simplify/
├── SKILL.md              # Process + guardrails (145 lines)
└── agents/
    └── openai.yaml       # UI metadata for Codex app (7 lines)

Requirements

  • OpenAI Codex CLI with skills = true
  • Git (for scope detection)
  • Optional: multi_agent = true for parallel analysis

License

MIT

↑↓ navigate openesc close