What does the skill review command specifically audit?

It audits Claude Code skills across eight categories, including structure, frontmatter validity, architecture hierarchy, cross-reference integrity, and convention compliance.

How does the parallel review mode work?

The parallel mode (/review-doc-run) spawns background agents that process multi-item documents incrementally, allowing for much faster reviews compared to sequential processing.

What types of documents can this skill review?

It supports a wide range of design and development documents, including vision statements, architecture specs, roadmaps, milestone specs, task specifications, and implementation plans.

Can this skill automatically fix the issues it finds?

Yes, the sequential doc-reviewer agent has access to full tools and can apply fixes directly, while the parallel mode uses an orchestrator to write findings and apply edits based on agent reports.

Document & Skill Reviewer

Name: Document & Skill Reviewer
Author: creational-ai-legacy

bycreational-ai-legacy

0•

セキュリティとテスト

Automates quality assurance for technical documents and Claude Code skills by auditing structural integrity, logical consistency, and convention compliance.

The Review skill provides a unified framework for ensuring the high quality of technical documentation and the structural correctness of Claude Code skills. By employing specialized agents for document verification and skill auditing, it identifies logical flaws, broken dependency chains, and architectural inconsistencies. It supports both sequential and high-speed parallel processing, allowing developers to maintain rigorous standards for design specifications, implementation plans, and custom skill developments without manual overhead.

主な機能

01Automated auditing of Claude Code skill structure and conventions

020 GitHub stars

03Parallel review mode for high-speed analysis of complex, multi-item documents

04Deep logical consistency checks for dependency chains and contradictions

05Unified document verification for design and development specifications

06Automatic tracking and persistence of review history to dedicated files

ユースケース

01Validating software architecture and roadmap documents before starting implementation

02Auditing custom Claude Code skills for frontmatter validity and structural compliance

03Checking complex task specifications for hidden dependencies and circular references

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add creational-ai-legacy/anvil review

For use in Claude.ai and ChatGPT