improve

Analyze targets for issues and guide evidence-based resolution.

Updated Apr 3, 2026
One-click install
npx skills add https://github.com/robertdp/claude-code-skills --skill improve-robertdp
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: improve
Source: https://github.com/robertdp/claude-code-skills/tree/main/skills/improve
Command: npx skills add https://github.com/robertdp/claude-code-skills --skill improve-robertdp

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve?

This Skill helps you critique and improve a target by finding real issues, explaining why they matter, and walking through fixes one at a time so you stay in control.

Core Features & Use Cases

  • Multi-frame critique: Reviews a target through adversarial perspectives to surface correctness, complexity, and missed edge cases.
  • Guided issue resolution: Presents findings individually with evidence, tradeoffs, and implementation choices.
  • Interactive fixes: Applies approved changes and verifies the result, making it useful for code, plans, docs, or architecture reviews.

Quick Start

Use the improve skill to critique and fix the target you want reviewed.

Frequently Asked Questions about improve

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

FAQPage Schema
How do I review and fix hidden issues in my architecture or code?

Adversarial analysis reviews a target through multiple perspectives to surface hidden assumptions, edge cases, and complexity. It guides resolution by presenting findings individually with evidence, tradeoffs, and implementation choices.

What is the best way to critique a plan and explore fixes step by step?

Guided issue resolution critiques a plan by walking through fixes one at a time. It presents findings individually with evidence and tradeoffs, keeping you in control of each implementation choice.

Can I use adversarial analysis for documents and plans, not just code?

Yes, adversarial analysis applies to code, plans, documents, and architecture reviews. It surfaces correctness issues, unnecessary complexity, and missed edge cases across any target requiring structured critique.

How does multi-frame critique work for finding edge cases?

Multi-frame critique reviews a target through adversarial perspectives to surface hidden assumptions and missed edge cases. It fact-checks findings and uses stepwise decision making to produce actionable improvements.

Does interactive issue resolution apply changes automatically?

Interactive issue resolution applies approved changes and verifies the result. It presents findings individually with implementation choices, ensuring you stay in control before any fix is applied.

When should I not use stepwise decision making for code review?

Stepwise decision making is less suited for rapid, superficial formatting passes. It is designed for complex targets where hidden assumptions, edge cases, and tradeoffs require evidence-based, controlled implementation.