majiayu000 Instruction Writer Skill
Expertise in writing effective instructions for Claude Code using latest prompt engineering best practices.
When to Use This Skill
MANDATORY - Automatically invoked when:
- Creating or modifying ANY CLAUDE.md file (project, global, or package-level)
- Creating or modifying SKILL.md files
- Creating or modifying slash command files
- Creating or modifying any instruction/documentation files meant for Claude
- Optimizing prompts for clarity or token efficiency
- User mentions "write instructions", "create a skill/command", or "optimize this prompt"
Core Principles
Apply all principles from @best-practices.md when writing or optimizing instructions. Key focus areas: clarity & specificity, token efficiency, canonical examples, and actionable steps.
Workflow
For New Instructions
- Understand the purpose and trigger conditions
- Draft clear, specific description (if skill) or title/purpose (if command)
- Structure content with headers/XML tags
- Add canonical examples
- Review for token efficiency
For Optimization
- Read target file first to identify all linked files (@ references)
- Read all linked files from target
- Read @best-practices.md for comparison
- Identify issues: verbosity, unclear structure, weak examples, cross-file redundancy
- Report findings with before/after examples
- Wait for approval before applying changes
- Apply optimizations preserving all salient information
Supporting Files
- @best-practices.md: Comprehensive prompt engineering guidelines for Claude 4.x