Why should I avoid negative instructions like 'Never do X'?

Models often attend to key nouns; 'Never use bare python' still activates the concept of bare python. Positive framing like 'Use uv run instead' provides a clear, actionable path.

What is the benefit of adding 'Motivation' to instructions?

Claude generalizes better when it understands the 'Why.' Including a reason helps the model apply the rule correctly even in novel or edge-case situations.

How does this skill help with Claude 4.5?

It implements specific optimizations for Claude 4.5, such as using direct action language and structuring instructions to take advantage of parallel tool execution.

Can I use this to improve my custom Skill files?

Absolutely. It provides templates for description fields, tool restrictions, and progressive disclosure to ensure your custom skills are discovered and used correctly by Claude.

Claude Code Prompt Optimizer

Name: Claude Code Prompt Optimizer
Author: Jamie-BitFlight

byJamie-BitFlight

•

開発者ツール

Refines CLAUDE.md and Skill files by applying Anthropic's official prompt engineering best practices for the Claude Code CLI.

概要

This skill automates the optimization of CLAUDE.md files and custom Agent Skills by converting restrictive negative instructions into actionable positive patterns. It enhances model performance by adding technical context, specific code standards, and concrete examples. By specifically leveraging Claude 4.5's capabilities—such as parallel tool usage and direct action language—this skill ensures that your CLI interactions are more reliable, concise, and aligned with official Anthropic implementation patterns.

主な機能

Implements multi-shot example patterns to improve model accuracy and consistency
Converts negative 'don't' prohibitions into actionable positive behavior patterns
Optimizes instructions for Claude 4.5 direct action language and parallel tool usage
Structures memory files with Markdown headings for better cognitive attention
10 GitHub stars
Applies official Anthropic best practices for specificity and technical motivation

ユースケース

Auditing existing CLAUDE.md files to improve tool usage and communication style
Creating new custom Skills with properly formatted metadata and tool restrictions
Troubleshooting brittle instructions that the AI model frequently ignores

概要

主な機能

Implements multi-shot example patterns to improve model accuracy and consistency
Converts negative 'don't' prohibitions into actionable positive behavior patterns
Optimizes instructions for Claude 4.5 direct action language and parallel tool usage
Structures memory files with Markdown headings for better cognitive attention
10 GitHub stars
Applies official Anthropic best practices for specificity and technical motivation

ユースケース

Auditing existing CLAUDE.md files to improve tool usage and communication style
Creating new custom Skills with properly formatted metadata and tool restrictions
Troubleshooting brittle instructions that the AI model frequently ignores