What specifically does the Rules Evaluator check?

It validates YAML frontmatter syntax, checks glob pattern correctness, assesses content quality for conciseness, audits directory organization, and measures token efficiency.

Can I use this skill to write new rules from scratch?

No, this skill is designed for auditing and validating existing rules. For creating new rules, you should use standard rule-authoring guides or dedicated authoring skills.

How do I get a more comprehensive report than the basic scan?

You can use the /rules-eval --detailed command to generate a full analysis with specific improvement recommendations for every rule file found.

Does it detect conflicts between different rule files?

Yes, part of the content quality assessment includes checking for non-conflicting instructions to ensure the LLM receives clear, consistent guidance.

What is considered a passing score for a rule file?

Rules scoring 76-90 are considered 'Good' with minor improvements possible, while scores 91-100 are considered 'Excellent' and production-ready.

Rules Evaluator

Name: Rules Evaluator
Author: athola

byathola

•

230

•

Developer Tools

Audits and validates Claude Code rules to ensure high-quality frontmatter, glob pattern accuracy, and token efficiency.

The Rules Evaluator (rules-eval) is a specialized diagnostic tool designed to maintain the integrity and performance of Claude Code rule sets. It performs deep analysis on rule files within the .claude/rules/ directory, identifying YAML syntax errors, unquoted glob patterns, and overly broad path definitions that can degrade LLM performance. By applying a rigorous scoring system across five key dimensions—frontmatter, glob quality, content, organization, and token efficiency—it ensures that your rules are production-ready, non-conflicting, and optimized for minimal context window consumption.

Key Features

01230 GitHub stars

02Automated YAML frontmatter validation for syntax and required field compliance

03Token efficiency measurement to minimize rule redundancy and context overhead

04Organizational structure checks including naming conventions and symlink usage

05Content quality auditing to ensure rules are actionable, concise, and non-conflicting

06Advanced glob pattern analysis to identify syntax errors and path-scoping issues

Use Cases

01Optimizing LLM performance by identifying and condensing verbose or redundant rules

02Debugging why specific rules are not triggering due to invalid glob pattern definitions

03Auditing project-specific rules for quality and syntax before a production deployment

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add athola/claude-night-market rules-eval

For use in Claude.ai and ChatGPT

Key Features

01230 GitHub stars

02Automated YAML frontmatter validation for syntax and required field compliance

03Token efficiency measurement to minimize rule redundancy and context overhead

04Organizational structure checks including naming conventions and symlink usage

05Content quality auditing to ensure rules are actionable, concise, and non-conflicting

06Advanced glob pattern analysis to identify syntax errors and path-scoping issues

Use Cases

01Optimizing LLM performance by identifying and condensing verbose or redundant rules

02Debugging why specific rules are not triggering due to invalid glob pattern definitions

03Auditing project-specific rules for quality and syntax before a production deployment