SuperClaude/.claude/commands/explain.md

Legend

Symbol	Meaning		Abbrev	Meaning
→	leads to		MCP	model context protocol
&	and/with		impl	implementation
w/	with		perf	performance

Execute immediately. Add --plan flag if user wants to see plan first.

Provide comprehensive explanations of concepts, code, or systems in $ARGUMENTS.

Thinking flags (optional):

• --think→multi-concept explanation w/ connections
• --think-hard→complex system explanations w/ deep analysis
• --ultrathink→comprehensive knowledge synthesis & teaching

MCP flags (optional):

• --c7→fetch official lib/framework docs | --seq→step-by-step concept breakdown
• --no-mcp→explain using only codebase context

Doc flags:

• --ultracompressed or --uc→7 0% token reduction in explanations

Examples:

• /user:explain --c7 "React hooks" --think - React hooks w/ contextual connections
• /user:explain --seq "OAuth flow" --think-hard - Deep OAuth analysis w/ security implications
• /user:explain --depth advanced --c7 --ultrathink - Comprehensive system explanation

--depth flag levels:

• ELI5: Simple analogies, avoid jargon, focus→concepts
• Beginner: Cover basics, step-by-step guidance, simple examples
• Intermediate: Tech details, best practices, common patterns
• Expert: Deep impl details, edge cases, perf considerations

When --visual flag is present:

• Create diagrams using Mermaid syntax
• Use ASCII art for simple visualizations
• Include flowcharts for processes
• Add sequence diagrams for interactions

Structure the explanation:

1. Start with a high-level overview
2. Break down into logical components
3. Explain each component with appropriate depth
4. Show relationships and interactions
5. Provide concrete examples

Include relevant formats:

• Code examples that can be run and tested
• Comparison tables for alternatives
• Pros and cons lists for decisions
• Links to authoritative resources

For technical concepts:

• Explain the "what" before the "how"
• Connect to familiar concepts when possible
• Build understanding incrementally
• Address common misconceptions

For code explanations:

• Walk through logic step by step
• Explain design decisions
• Highlight important patterns
• Point out potential gotchas

UltraCompressed Mode

When --ultracompressed flag is present:

• Apply shared/ultracompressed.yml rules
• Generate legend for symbols/abbreviations used
• Compress explanations to ~30% of normal size
• Use telegram-style: remove articles, use symbols
• Format: lists>paragraphs, tables>prose
• Example: "fn takes 3 params→bool" instead of "The function takes three parameters and returns a boolean value"

Research requirements for explanations:

• Technical concepts → C7 documentation for authoritative explanations and examples
• Framework patterns → Must verify with official guides and documentation
• Best practices → WebSearch for current industry standards and expert opinions
• Code examples → Research working patterns from official documentation
• Never explain without verification - always cite authoritative sources
• All explanations must include sources: // Source: [documentation reference]

Report Output:

• Explanation summaries: .claudedocs/summaries/explanation-<topic>-<timestamp>.md
• Learning resources: .claudedocs/reports/learning-resources-<timestamp>.md
• Ensure directory exists: mkdir -p .claudedocs/summaries/ .claudedocs/reports/
• Include report location in output: "📚 Explanation saved to: [path]"

Deliverables: Clear explanation at the appropriate depth level, working examples, visual aids if requested, curated resources for further learning, and comprehensive source references.