skills.rest
rmans
rmansCOMMUNITY
Explore all skills by this author →
GitHub

scaffold-iterate

Review design documents via adversarial two-loop LLM feedback.

Updated Aug 27, 2026
Search Tags:#code review#llm#scaffold#document review#adversarial testing#design authority
One-click install
npx skills add https://github.com/rmans/ClaudeScaffold --skill scaffold-iterate
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: scaffold-iterate
Source: https://github.com/rmans/ClaudeScaffold/tree/main/Install/.claude/skills/scaffold-iterate
Command: npx skills add https://github.com/rmans/ClaudeScaffold --skill scaffold-iterate

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires python, and includes scripts (resource) and references (resource) components.

What problem does it solve?

This Skill addresses the critical need for rigorous, unbiased review of design documents, catching blind spots that self-review or standard peer review might miss.

Core Features & Use Cases

  • Adversarial LLM Review: Utilizes an external LLM to act as a critical reviewer, identifying potential issues in your documents.
  • Iterative Feedback Loop: Engages in a back-and-forth dialogue between Claude and the external reviewer to resolve issues, allowing Claude to push back or accept feedback.
  • Document Authority Enforcement: Ensures that proposed changes align with the established project's document authority hierarchy.
  • Use Case: Before finalizing a critical system design document, run this Skill to have an AI reviewer challenge its assumptions and identify potential conflicts with higher-level design principles.

Quick Start

Use the scaffold-iterate skill to review the document located at scaffold/design/systems/SYS-001.md.

Frequently Asked Questions about scaffold-iterate

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

FAQPage Schema
How do I run an adversarial review on a design document using an LLM?▼

Adversarial document review uses an external LLM to critically evaluate your design documents, engaging in an iterative two-loop feedback process with Claude to identify and resolve structural issues.

What is the best way to enforce design authority hierarchy during spec review?▼

Adversarial spec review enforces document authority by validating proposed changes against established project hierarchy, ensuring lower-level system documents align with higher-level design principles during the reviewer dialogue.

Does this adversarial review approach support different document criticality levels?▼

Yes, adversarial document review supports tiered review intensity based on criticality, handling design, system, spec, and task files to match the depth of scrutiny required for each document type.

Can I use Python scripts to automate LLM adversarial testing for system design files?▼

Python scripts automate the adversarial testing workflow, logging the review process and facilitating structured discussion between Claude and the external LLM reviewer across various system document formats.

How do I push back against LLM reviewer feedback during document iteration?▼

During document iteration, the iterative feedback loop allows Claude to either accept the external reviewer's feedback or push back, logging the discussion and resolution to ensure robust design outcomes.

Related Skills

Explore More Software Engineering Skills→
ABLER-Digital-ServicesABLER-Digital-Services

req-fn-app-actor-definition

Define system actors and create an actor definition document for authorization design.

Official
Intermediate
RBozydarRBozydar

compound-docs

Automates structured markdown documentation of resolved problems with YAML frontmatter.

Community
Advanced
ShmurdocShmurdoc

codex

Wrap the OpenAI Codex CLI for code review, adversarial testing, and consultation.

Community
Advanced
fredmilhomefredmilhome

devils-advocate

Generate adversarial pedagogical questions from lecture documents to review slide decks.

Community
Basic