PolyArchPolyArchOfficialยท6 Agent Skills Included

humanize

Iterative AI coding loops with independent automated code review

Runs iterative development loops where one AI implements your plan while Codex independently reviews every round. Generates structured implementation plans from rough drafts, complete with acceptance criteria, scope boundaries, and milestones. Eliminates blind spots and one-shot coding errors by feeding review findings back until all criteria pass. Includes one-shot expert consultations with Codex and Gemini for research-backed answers.
npx skills add PolyArch/humanize --all -g -y
Available:

Tells the AI agent how to start and cancel RLCR review loops and enforces project-wide rules such as English-only output and strict version numbering.

All Skills in This Repository (6)

Pure Emerald Level Indicators

Frequently Asked Questions

FAQPage Schema
How to install Humanize?โ–ผ

Run `npx skills add PolyArch/humanize --all -g -y` in your terminal to install all skills in this suite globally.

What is the RLCR loop in Humanize?โ–ผ

RLCR (Ralph-Loop with Codex Review) is an iterative cycle where Claude implements your plan and Codex independently reviews each round, repeating until all acceptance criteria pass.

How does Humanize improve AI-generated code quality?โ–ผ

It catches issues early through continuous external review with severity markers, feeding problems back into implementation instead of trusting one-shot output.

Does Humanize work with Claude Code and Codex?โ–ผ

Yes. It runs as a Claude Code plugin and uses the Codex CLI for independent reviews, with setup guides for Codex and Kimi runtimes.

Can Humanize turn rough notes into a coding plan?โ–ผ

Yes. The gen-idea and gen-plan skills convert loose thoughts or drafts into structured plans with acceptance criteria, scope boundaries, and milestones.

Related Repositories in Software Engineering

View All in Software Engineeringโ†’