hos-json-prompts

name	HOS-JSON-Prompts
description	Structured prompt engineering framework with JSON templates for generating standardized, reusable AI prompts across diverse scenarios including articles, exams, projects, security, and operations.
version	1.0.0
author	HOS
tags	["prompt-engineering","json-templates","structured-output","ai-writing"]
category	prompt-engineering
risk-level	low
confidence	0.85

🎯 JSON Structured Prompts Framework

Overview

A comprehensive prompt engineering framework that uses JSON templates to generate standardized, reusable AI prompts. This system provides 15 specialized templates covering diverse scenarios from article writing to security auditing, enabling consistent and high-quality AI output across different use cases.

Core Concepts

1. Adaptive Instructions

The framework employs adaptive instruction methodology that dynamically adjusts prompt structure based on user requirements:

• Dynamic Restructuring: Templates serve as reference frameworks, not rigid output structures
• Semantic Reconstruction: Prioritizes reorganizing content based on user input rather than copying template fields
• Flexible Module Management: Allows deletion, reordering, or replacement of modules to fit specific scenarios
• User-Centric Priority: When templates conflict with user needs, user requirements take highest priority

2. Template Structure

Each JSON template follows a standardized structure:

{
  "adaptive_instruction": [...],
  "prompt": {
    "title": "...",
    "author": "...",
    "role": "...",
    "core_constraints": [...],
    "generation_strategy": {...},
    "writing_standards": {...},
    "file_management": {...},
    "final_goal": "..."
  }
}

Key components:

• Role Definition: Clear AI role and core capabilities
• Core Constraints: Input source limitations and originality requirements
• Generation Strategy: Planning and execution steps
• Writing Standards: Style, structure, and formatting guidelines
• File Management: Output organization and versioning rules

3. Writing Standards

All templates enforce consistent writing standards:

• Audience: AI engineers, prompt developers, product managers
• Style: Objective, concise, zero redundancy, logical progression
• Structure: Standardized sections (Role, Constraints, Strategy, Output Standards, Content Requirements, File Management)
• Length: Moderate (500-2000 words)
• Elements: ≥2 example code blocks or tables
• Format: JSON nested or Markdown compatible

File Structure

W-02-HOS-JSON-Prompts/
├── AAA-develop-template.json          # Master template with adaptive instructions
├── develop-article.json               # Article writing template
├── develop-article-code.json          # Article with code examples template
├── develop-auto-skill.json            # Automated skill generation template
├── develop-exam.json                  # General exam/quiz template
├── develop-exam-CT.json               # CT (Capture The Flag) exam template
├── develop-exam-SG.json               # SG exam template
├── develop-merged.json                # Merged multi-purpose template
├── develop-model.json                 # Model development template
├── develop-operations.json            # Operations workflow template
├── develop-operations-report.json     # Operations report template
├── develop-project.json               # Project management template
├── develop-security.json              # Security audit template
├── develop-test.json                  # Testing template
└── develop-token-save.json            # Token optimization template

Usage Guide

Basic Usage

1. Select Template: Choose the appropriate JSON template based on your use case
2. Customize Input: Provide your scenario description as the primary input source
3. Generate Prompt: Use the template structure to generate a standardized prompt
4. Iterate: Refine based on output quality and specific requirements

Template Selection Guide

Template	Use Case	Key Features
AAA-develop-template.json	Master reference	Complete framework with all components
develop-article.json	Content writing	Article structure, SEO optimization
develop-article-code.json	Technical articles	Code examples integration
develop-auto-skill.json	Skill automation	Automated workflow generation
develop-exam.json	General assessments	Quiz/exam generation
develop-exam-CT.json	CTF challenges	Security competition format
develop-exam-SG.json	SG assessments	Specialized exam format
develop-merged.json	Multi-purpose	Combined template for flexible use
develop-model.json	Model development	AI/ML model training prompts
develop-operations.json	Operations	Operational workflow automation
develop-operations-report.json	Reporting	Operations report generation
develop-project.json	Project management	Project planning and tracking
develop-security.json	Security	Security audit and assessment
develop-test.json	Testing	Test case generation
develop-token-save.json	Optimization	Token-efficient prompt design

Best Practices

1. Start with Master Template: Use AAA-develop-template.json as reference for understanding the framework
2. Adapt, Don't Copy: Treat templates as frameworks, not rigid structures
3. Focus on User Input: Primary content should come from user scenario descriptions
4. Maintain Standards: Follow writing standards for consistency
5. Iterate and Refine: Continuously improve templates based on output quality

Integration with HOS Ecosystem

This framework integrates with other HOS components:

• HOS-Sec-Engine: Security validation for generated prompts
• HOS-Vibe-Guard: Quality assessment and template trap detection
• HOS-Silly-Mock: Prevention of fake data patterns in generated content

License

MIT