Center Code

center-code is a Codex skill for high-risk or ambiguous code changes. It forces a file-backed center plan before implementation, requires three root-level approaches, two outside reviews, fixed TDD slices, data-contract hardening, and post-change center-audit proof.

Install

Clone or copy this directory into a skills folder recognized by your agent runtime:

git clone https://github.com/VerbalChainsaw/center-code.git $env:USERPROFILE\.codex\skills\center-code

For cross-runtime discovery on systems that support ~/.agents/skills, mirror or link it there:

New-Item -ItemType Junction -Path $env:USERPROFILE\.agents\skills\center-code -Target $env:USERPROFILE\.codex\skills\center-code

Restart or reload the agent session after installation so the skill list refreshes.

Use

Ask for center-code, or use it when a code change is risky, release-sensitive, persistence-affecting, UI-flow-affecting, or when the user asks for plan-first work, no hacks, outside review, durable planning artifacts, or implementation slices.

Contents

• SKILL.md - the skill entrypoint.
• references/plan-template.md - durable plan template.
• agents/openai.yaml - optional agent interface metadata.