/rules-eval

name: rules-eval
description: Evaluate Claude Code rules in .claude/rules/ directories for quality
usage: /rules-eval [rules-path] [options]

Rules-Eval

Evaluate and validate Claude Code rules in `.claude/rules/` directories. Checks YAML frontmatter, glob patterns, content quality, and directory organization.

When To Use

Use this command when you need to:

• Validate rule files before deployment
• Audit frontmatter for YAML errors or Cursor-specific fields
• Check glob patterns for syntax and specificity issues
• Assess overall rules organization and naming

When NOT To Use

Avoid this command if:

• Evaluating skills - use /skills-eval instead
• Evaluating hooks - use /hooks-eval instead
• Validating full plugin structure - use /validate-plugin instead

Usage

# Evaluate rules in current project
/rules-eval

# Evaluate specific directory
/rules-eval .claude/rules/

# Detailed analysis with per-file breakdown
/rules-eval --detailed

# Evaluate a plugin's rules
/rules-eval plugins/conserve/.claude/rules/

Scoring System (100 points)

| Category | Points | Focus | |----------|--------|-------| | Frontmatter Validity | 25 | YAML syntax, correct fields, no Cursor-specific fields | | Glob Pattern Quality | 20 | Syntax, specificity, quoting | | Content Quality | 25 | Actionable, concise, focused | | Organization | 15 | Naming conventions, directory structure | | Token Efficiency | 15 | Rule size, redundancy |

Quality Levels

• **91-100**: Excellent - Production-ready
• **76-90**: Good - Minor improvements possible
• **51-75**: Basic - Needs optimization
• **26-50**: Below Standards - Significant issues
• **0-25**: Critical - Invalid or broken rules

Key Validations

Frontmatter

• Valid YAML syntax
• No Cursor-specific fields (`globs`, `alwaysApply`)
• `paths` is a list of quoted glob patterns
• Only known fields (`paths`, `description`)

Glob Patterns

• Valid glob syntax
• Not overly broad (`**/*`)
• Properly quoted special characters

Content

• Actionable guidance (not just descriptions)
• Concise (< 500 tokens per file)
• Single focused topic per file

Organization

• Descriptive kebab-case filenames
• No generic names (`rules1.md`, `misc.md`)
• Subdirectories for large rule sets
• No broken symlinks

Implementation

python3 ${CLAUDE_PLUGIN_ROOT}/scripts/rules_validator.py \
  "${1:-.claude/rules}" \
  ${2:+--detailed}

Related Commands

• `/skills-eval` - Skill quality evaluation
• `/hooks-eval` - Hook quality evaluation
• `/validate-plugin` - Complete plugin structure validation

Related Skills

• **rules-eval skill** (`skills/rules-eval/SKILL.md`) - Detailed evaluation framework
• `modules/frontmatter-validation.md` - YAML/frontmatter checks
• `modules/glob-pattern-analysis.md` - Glob pattern validation
• `modules/content-quality-metrics.md` - Content quality assessment
• `modules/organization-patterns.md` - Directory organization