Which version of the Claude Code guidelines are supported?

This skill is built to validate against Architecture Guidelines v0.5 and current Plugin Storage Conventions.

What specifically does the Claude Skill Reviewer validate?

It validates three primary areas: metadata/storage (YAML and persistence), implementation (JSON fencing and hooks), and architecture (two-tier separation and naming conventions).

What happens if one of the review agents fails?

The system provides partial results from successful agents and reports specifically on which reviewer failed and why, ensuring you still receive valuable feedback.

Can I review specific files or just entire packages?

You can review specific agents, individual skills, or entire packages by providing the relevant path to the review command.

Does this tool modify my source code?

No, the skill is strictly read-only. It analyzes your files and provides a detailed report with suggested actions for you to implement.

Claude Skill Reviewer

Name: Claude Skill Reviewer
Author: randlee

byrandlee

•

Security & Testing

Validates and audits Claude Code skills and agents against architectural, implementation, and storage standards.

The Claude Skill Reviewer is a specialized validation suite designed for developers building within the Claude Code ecosystem. It utilizes a three-agent orchestration pattern to concurrently analyze metadata persistence, code mechanics (security and hooks), and architectural integrity such as two-tier separation. By providing actionable, structured reports categorized by severity, this skill ensures that custom agents and skills meet the latest v0.5 guidelines, facilitating higher quality standards and seamless deployment for AI-driven workflows.

Key Features

01Parallel orchestration of metadata, implementation, and architecture review agents

02Architectural compliance checks for two-tier skill/agent design patterns

03Security-focused code reviews targeting JSON fencing and hook implementation

041 GitHub stars

05Automated YAML frontmatter and registry.yaml consistency validation

06Actionable reporting with grouped errors, warnings, and optimization tips

Use Cases

01Debugging agent tool-use hooks and dependency declarations during development

02Ensuring legacy agent packages comply with modern Plugin Storage Conventions

03Auditing new Claude Code skills before production deployment

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add randlee/synaptic-canvas skill-reviewing

For use in Claude.ai and ChatGPT

Download Skill