What is the difference between Safe and Risky fixes?

Safe fixes are non-destructive changes like syntax correction or header casing that are applied automatically. Risky fixes involve structural changes or logic removal and always require user confirmation.

Does it support batch processing?

Yes, by setting the scope to 'marketplace', Plugin Doctor can perform a full health check and batch analysis of all components in your project directory.

How does it ensure fixes are successful?

Every workflow includes a mandatory 'Verify and Report' phase that executes validation scripts and provides a git-status summary of all changes made.

What components can Plugin Doctor analyze?

Plugin Doctor can diagnose and fix issues across Agents, Commands, Skills, Metadata (plugin.json), and Scripts within the Claude Code ecosystem.

How does Plugin Doctor handle component bloat?

The skill monitors line counts against specific thresholds (e.g., <200 for commands, <800 for skills) and flags components as 'Bloated' or 'Critical' when they exceed best-practice limits.

Plugin Doctor

Name: Plugin Doctor
Author: cuioss

bycuioss

0•

開発者ツール

Diagnoses and repairs quality issues in Claude Code marketplace components through automated safe fixes and guided refactoring workflows.

Plugin Doctor is a specialized diagnostic utility designed for the Plan Marshall orchestration layer to maintain the integrity and performance of Claude Code components. It follows a rigorous Diagnose-Fix-Verify cycle to ensure agents, commands, and skills adhere to standardized architectural patterns. By automatically resolving low-risk issues like metadata errors and prompting for complex refactoring—such as thin-wrapper enforcement or bloat reduction—this skill streamlines the development of reliable, high-quality AI toolsets.

主な機能

01Component bloat analysis with specific line-count threshold monitoring

02Automated 'Safe Fix' application for YAML syntax and frontmatter errors

03Integrated verification reporting using custom diagnostic scripts

04Rule enforcement for architectural patterns including Rule 6 and Pattern 22

050 GitHub stars

06Context-efficient progressive disclosure for minimal token usage

ユースケース

01Refactoring bloated command implementations into thin wrappers with skill delegation

02Performing full marketplace health checks to ensure metadata consistency

03Standardizing foundation skill loading across distributed agent architectures

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add cuioss/plan-marshall plugin-doctor

For use in Claude.ai and ChatGPT

Download Skill