nidhinjs/prompt-master

Metadata

• **Stars:** 4,992
• **License:** MIT
• **Last pushed:** 2026-03-31
• **Description:** A Claude skill that writes the accurate prompts for any AI tool. Zero tokens or credits wasted. Full context and memory retention.

Archetype: Specialization Process (prompt-engineering domain)

Structure

• SKILL.md — Complete prompt engineering skill with tool-routing logic
• references/ — templates.md and patterns.md for category-specific templates

Extractable Value

As a Babysitter Process: `specializations/shared/prompt-engineering`

The SKILL.md contains a well-structured prompt engineering methodology:

1. **9-dimension intent extraction framework** — Task, Target tool, Output format, Constraints, Input, Context, Audience, Success criteria, Examples. Each dimension marked as critical/conditional. 2. **Tool-specific routing logic** — Detailed prompting guidance for 16+ tools/models: - Claude (4.x), ChatGPT/GPT-5.x, o3/o4-mini reasoning models, Gemini 2.x/3, Qwen 2.5/3, Ollama, Llama/Mistral, DeepSeek-R1, MiniMax M2.7/M2.5, Claude Code, Antigravity, Cursor/Windsurf, Cline, GitHub Copilot 3. **Anti-pattern enforcement** — Hard rules against techniques that cause fabrication in single-prompt execution (Mixture of Experts, Tree of Thought, Graph of Thought, etc.) 4. **Model-specific constraints** — e.g., "NEVER add CoT to reasoning-native models", "Claude Opus 4.x over-engineers by default"

Key Differentiators

• The tool-routing approach is unique — maps target tool to specific prompting strategy
• Anti-fabrication rules are concrete and actionable (not vague "be careful" advice)
• The 9-dimension intent extraction is a reusable analysis framework
• Output format is strictly enforced: copyable prompt block + target + optimization note

Process Mapping

Would map well to a babysitter process with:

• Breakpoint for tool identification when ambiguous
• Task for intent extraction across 9 dimensions
• Tool-routing logic as branching in the process
• Output formatting as final stage

Processes

1. Advanced Prompt Engineering Methodology

• **Source**: SKILL.md 9-dimension intent extraction framework and tool-routing logic
• **Placement**: specializations/shared/advanced-prompt-engineering.js
• **Description**: Comprehensive prompt engineering process: extract intent across 9 dimensions (task, target tool, output format, constraints, input, context, audience, success criteria, examples) → apply tool-specific routing logic for 16+ AI models → enforce anti-fabrication rules → generate optimized prompt with target tool and optimization notes.

Plugin Ideas

• **Prompt Engineering Suite**: Babysitter marketplace plugin that provides advanced prompt engineering capabilities with tool-routing logic and anti-pattern enforcement for optimal AI model communication.

Library Mapping

Plugin Marketplace Mapping

Classification Rationale

Cross-domain (works for any AI tool/model), so fits in specializations/shared/. The tool-routing logic and anti-fabrication rules are unique contributions not found in the current process library.