Optimizes and refines LLM prompts using modern engineering principles for high-performance reasoning models.
Prompt Master is a specialized Claude Code skill designed to transform vague or inefficient AI instructions into high-precision prompts. Based on the latest 2025-2026 guidelines from Anthropic and OpenAI, it helps users apply sophisticated techniques like XML tagging, prompt chaining, and motivational framing to reduce token overhead while increasing output accuracy. It is particularly effective for debugging failing prompts, structuring complex data workflows, and maximizing the reasoning capabilities of advanced LLMs.
主な機能
01Context-aware prompt refinement based on current LLM best practices
02Conversion of negative prohibitions into high-signal positive framing
03Instruction compression to maximize context window efficiency
04XML-tagging implementation for clean separation of instructions and data
0512 GitHub stars
06Transition strategies from mega-prompts to reliable prompt chains
ユースケース
01Systematizing prompt engineering patterns across large-scale AI projects
02Optimizing complex developer workflows that require high-reasoning accuracy
03Refactoring legacy prompts to improve consistency and reduce format errors
