humanize

Coordinate iterative AI development with independent code review until acceptance criteria are met.

1.4k|123|Updated Jan 12, 2026
One-click install
npx skills add https://github.com/PolyArch/humanize --skill humanize-polyarch
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: humanize
Source: https://github.com/PolyArch/humanize/tree/main/skills/humanize
Command: npx skills add https://github.com/PolyArch/humanize --skill humanize-polyarch

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Humanize addresses the challenge of delivering complex software through iterative collaboration between AI agents, enabling one AI to implement plans while another AI provides independent reviews to surface issues early.

Core Features & Use Cases

  • RLCR Loop: Two-phase cycle of Implementation and Code Review that iterates until acceptance criteria are met.
  • Plan-to-Code Workflow: Start from a rough idea or draft, generate a structured plan, implement it, and undergo independent review.
  • Stop-and-Gate Quality: Automatic gating and finalization based on Codex review feedback and automated summaries.

Quick Start

Run the RLCR loop setup with a plan file to start the iterative development and review process.

Frequently Asked Questions about humanize

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I set up iterative AI development with independent code review?

Iterative AI development with independent code review starts by providing a plan file to initiate the RLCR loop, which coordinates one AI for implementation and another for review. This cycle continues until acceptance criteria are met.

Can I generate a structured software development plan from a rough draft?

Generating a structured software development plan from a rough draft is supported directly. The workflow enforces plan generation from initial ideas before entering the phase-based implementation and code review steps.

What is the best way to automate software quality gating during AI implementation?

Automating software quality gating during AI implementation is handled through stop-and-gate finalization. The system automatically gates and finalizes phases based on independent review feedback and automated summaries.

Does end-to-end software project management work with iterative AI agents?

End-to-end software project management works with iterative AI agents by applying a structured RLCR loop. This approach guides multiple implementation and review iterations specifically for complex projects requiring rigorous execution.

How does the RLCR loop handle defect detection in generated code?

The RLCR loop handles defect detection by using a second, independent AI to review the work of the implementing AI. This two-phase cycle of implementation and code review surfaces issues early and repeats until acceptance criteria are satisfied.