publish-step6

Classify madamisu-publish step6 review issues into four tiers and apply safe edits.

Updated May 20, 2026
One-click install
npx skills add https://github.com/kuro0107/madamisu-skills --skill publish-step6
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: publish-step6
Source: https://github.com/kuro0107/madamisu-skills/tree/main/plugins/madamisu/skills/publish-step6
Command: npx skills add https://github.com/kuro0107/madamisu-skills --skill publish-step6

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It resolves inconsistencies found during the madamisu-publish review by classifying issues by severity and automatically applying safe fixes while pausing for user-confirmation on higher-risk problems.

Core Features & Use Cases

  • Issue classification for publish review: Splits review problems into four tiers—light (auto-fix), reference/info-level (auto-fix), design-level (user confirmation), and premise-level (stop).
  • Controlled auto-fixing: Applies edits only for “light” and “reference/info-level” issues, and records every change with review annotations.
  • Safety guardrails with gated progression: Stops on premise-level issues and loops back for design-level problems based on explicit user decisions, then performs a limited-impact diff review.

Quick Start

Run publish-step6 by providing the input directory and the target publish version, so it reads the review report, classifies issues, auto-applies safe fixes, and produces a completion log.

Frequently Asked Questions about publish-step6

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

FAQPage Schema
How do I auto-fix content review issues while preventing unsafe edits?

Classifying review problems into severity tiers enables safe auto-fixing by applying edits only to light and reference-level issues, while design and premise tiers require user confirmation or stop execution.

What is the best way to classify review problems before applying automated fixes?

Classifying review problems involves splitting them into light, reference, design, and premise tiers. This categorization determines whether issues are auto-fixed, require user confirmation, or halt the process, ensuring controlled and safe content modifications.

How do I generate a diff review report after fixing markdown review artifacts?

Generating a diff review report after auto-fixing markdown artifacts requires running a classified workflow that logs all safe edits. The system processes review reports, applies tiered fixes, and saves working logs under a designated directory.

Can I automate versioned publishing workflows without manual intervention for every issue?

Automating versioned publishing workflows is possible by classifying issues into light and reference tiers for automatic fixes. Higher-severity design and premise issues trigger user confirmation gates or stop execution, ensuring necessary manual oversight.

Why does my content review workflow stop when encountering premise-level issues?

Content review workflows stop when encountering premise-level issues to enforce safety guardrails. This gated progression ensures foundational problems are resolved manually before proceeding, preventing unsafe automatic edits to critical artifacts.