brown-mm

Audit Claude output across four passes and produce a corrected version with an error report.

6|5|Updated Mar 9, 2026

Search Tags:#audit #review #quality-assurance #multi-pass #error-detection #fact-verification #internal-consistency

Ready to run

npx skills add https://github.com/kategage/progressive-ai-skills --skill brown-mm

💬 Prompt for ChatGPT / Claude Chat▼

Please help me install this Agent Skill.
Skill: brown-mm
Source: https://github.com/kategage/progressive-ai-skills/tree/main/skills/brown-mm
Command: npx skills add https://github.com/kategage/progressive-ai-skills --skill brown-mm

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

This skill helps users uncover and fix hidden errors in Claude's outputs by performing a disciplined, multi-pass audit that surfaces inconsistencies, factual errors, and overlooked details.

Core Features & Use Cases

Four-pass audit workflow: fact verification, internal consistency, logical structure, and fresh-eyes review to produce a corrected version and a transparent error log.
Universal applicability: works across emails, code, documents, research, and plans.
Outcome-focused: delivers a clean, validated output with an explicit accounting of found issues.

Quick Start

Drop the SKILL.md into your skills directory and invoke the audit when you spot an error to receive a corrected output with a transparent findings report.

Frequently Asked Questions about brown-mm

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

FAQPage Schema

How do I audit AI output for factual errors and internal consistency?▼

To audit AI output for factual errors and internal consistency, run a four-pass review process that verifies facts, checks logical structure, and re-reads with fresh eyes to produce a corrected version and a transparent error report.

What is a multi-pass error detection workflow for generated documents?▼

A multi-pass error detection workflow systematically scans generated content across four stages: fact verification, internal consistency checks, logical structure review, and a fresh-eyes re-read to surface and fix hidden inconsistencies.

Can I use a post-output audit skill to review emails, code, and research plans?▼

Yes, a post-output audit skill applies universally across emails, code, documents, research, and plans, performing fact verification and logical structure checks to deliver a clean, validated output with an explicit accounting of found issues.

How do I fix overlooked details and inconsistencies in generated text?▼

To fix overlooked details and inconsistencies in generated text, apply a disciplined multi-pass audit that enforces fact verification and internal consistency checks, replacing the flawed output with a corrected version and a transparent findings log.

What's the best way to perform quality assurance on AI-generated content?▼

The best way to perform quality assurance on AI-generated content is a four-pass audit workflow covering fact verification, internal consistency, logical structure, and fresh-eyes review, yielding a corrected document and an explicit error log.

When should I invoke an error detection skill for Claude's output?▼

You should invoke an error detection skill when you spot an error in Claude's output or explicitly need a thorough post-output audit, triggering a four-pass review that produces a corrected version and a transparent error report.