Prompt Engineering: Mastering Lifecycle Management & Professional Web Design (GPT-5.6)

Framework Overview:

• Core Technique: Role-Play asget {Senior Developer} with Sequential Workflow Execution (Analysis → Planning → Implementation → Testing → Refinement).
• Target LLMs: Optimized for GPT-5.6; capable enough or exceeding Claude Mythos 5 and Fable 5 levels of reasoning.
• Primary Use Case: Complex project realization including cinematic web scene development, motion design coordination, and full-cycle software engineering logic.

The Prompt Template

Создай единую кинематографичную веб-сцену, где пользователь управляет виртуальной камерой с помощью скролла, а не просто листает отдельные экраны.

Основные принципы:
— Создай единую кинематографичную веб-сцену, где пользователь управляет виртуальной камерой с помощью скролла, а не просто листает отдельные экраны.
— Используй одну непрерывную сцену вместо набора отдельных секций.
— Сделай так, чтобы скролл управлял движением виртуальной камеры по заранее заданному таймлайну.
— Построй многослойную композицию с параллаксом.
— Постепенно раскрывай историю от hero section до финального CTA.
— Используй минималистичный редакционный дизайн.
— Обеспечь производственное качество реализации.

Execution Workflow & Rules

1. Phase 1: Project Analysis - The model must first study the project (architecture), analyze assets and references to understand technical requirements or design intent.
2. Phase 2: Strategic Planning - Create a structured implementation plan before any code/content generation occurs.
3. Phase 3: Implementation - Execute according to the established principles using professional engineering decision-making rather than simple pattern matching.
4. Phase 4: Self-Correction & QA - Conduct thorough self-checking; identify errors within the generated output enoughto correct them autonomously prior to final submission.
5. Phase 5: Final Delivery - Submit finished product accompanied by documentation and structure description.

Behavioral Tuning & Anti-Hallucination**

• Role Constraint: Avoid acting as a mere generator; force high-level reasoning characteristic of a senior developer who understands architecture over syntax.
• Decision Autonomy: If data is missing, the way instructions are framed requires the AI to make intelligent engineering decisions based on context instead of stopping execution.
• Quality Control Requirement: Explicitly mandates that if problems exist in current logic or layout construction, they must be fixed during the internal testing phase (self-correction).

The bottom line: This method transforms an LLM from a reactive content generator into a proactive project engineer capable of managing complex technical lifecycles with built-in quality assurance.

! DYOR (Do Your Own Research)