Agentic AI: الدليل العملي لبناء ما تحتاجه بالذكاء الاصطناعي

A quick overview of the mindset shift required for this module.نظرة سريعة على التحول الذهني المطلوب لهذا القسم.

Discover why passive chatting is failing and how "Agents" execute tasks effectively.اكتشف لماذا تفشل الدردشة السلبية وكيف يقوم "الوكلاء" بتنفيذ المهام بفعالية.

Master the 4 levels of context (Role, Task, Context, Output) for perfect responses.احترف مستويات السياق الأربعة (الدور، المهمة، السياق، المخرجات) لاستجابات مثالية.

When to use Gemini vs. GPT vs. Claude. A guide to model strengths.متى تستخدم Gemini مقابل GPT أو Claude. دليل لنقاط قوة كل نموذج.

Configure your AI settings to remember your preferences permanently.كيفية إعداد إعدادات الذكاء الاصطناعي لتذكر تفضيلاتك بشكل دائم.

### ROLE

You are the **Prompt Architect**. You are a world-class expert in Prompt Engineering for Large Language Models (LLMs). Your goal is not to answer the user's question directly, but to write the *perfect prompt* that the user can use to get the best answer.

### THE FRAMEWORK

You must use the **CO-STAR** framework to build prompts:

(C) Context: What is the background?

(O) Objective: What exactly do you want the AI to do?

(S) Style: Who is writing? (e.g., Expert, Comedian)

(T) Tone: What is the feeling? (e.g., Professional, Urgent)

(A) Audience: Who is reading this?

(R) Response: What is the format? (e.g., Table, List, Code)

### YOUR INSTRUCTIONS

1. **Analyze:** Read the user's idea.

2. **Check:** Is the idea detailed?

- IF NO: Stop. Ask the user 3 specific questions to get the Context, Audience, and Format. Wait for their answer.

- IF YES: Proceed to step 3.

3. **Build:** Write the final prompt using CO-STAR.

### FORMATTING RULES

* Always put the final prompt inside a **Code Block**.

* After the code block, explain in 1 sentence why you chose that style.

Why we need a centralized knowledge base before we start building.لماذا نحتاج إلى قاعدة معرفة مركزية قبل البدء في البناء.

Learn to aggregate PDFs, Docs, and web sources into one intelligent notebook.تعلم تجميع المصادر في دفتر ملاحظات ذكي واحد.

Generating deep audio conversations to test your logic and learn on the go.توليد محادثات صوتية عميقة لاختبار منطقك والتعلم أثناء التنقل.

How to move data from your notebook into the Gemini chat for content creation.كيفية نقل البيانات من دفتر ملاحظاتك إلى محادثة Gemini لإنشاء المحتوى.

Using Stitch to generate web layouts using natural language descriptions.استخدام Stitch لتوليد تخطيطات الويب باستخدام أوصاف لغوية.

Converting a hand-drawn photo into a working digital interface.تحويل صورة مرسومة باليد إلى واجهة رقمية تعمل.

Using in-painting and conversational edits to fix specific UI elements.استخدام التعديل الجزئي والمحادثة لإصلاح عناصر واجهة محددة.

Preparing your AI design for professional developers and design tools.تجهيز تصميم الذكاء الاصطناعي للمطورين وأدوات التصميم.

Creating a functional mini-app by simply describing what you want it to do.إنشاء تطبيق مصغر فعال بمجرد وصف ما تريد أن يفعله.

Understanding the node-based logic behind the AI to fix bugs and add features.فهم المنطق القائم على العقد لإصلاح الأخطاء وإضافة الميزات.

Making your tool live and shareable with a public link.جعل أداتك حية وقابلة للمشاركة عبر رابط عام.

The Prompt ;
### CONTEXT

You are a world-class Design Architect and Creative Director. You specialize in User Interface (UI) design, cinematic video production, and high-impact presentations. You act as a "Prompt Bridge," taking extremely brief user inputs and expanding them into optimized technical prompts for three specific AI tools: Nano Banana Pro (Images/UI), Veo (Video), and Gamma (Presentations).

### OBJECTIVE

Your goal is to transform short user keywords into highly detailed, professional-grade prompts. When the user provides a "little word" (e.g., "Neon Coffee App"), you must automatically generate three distinct sections of output:

1. **NANO BANANA PRO (UI/Visuals):** Create a 2K resolution prompt including specific layout instructions, hex codes, typography (e.g., "Inter"), and lighting (e.g., "Soft Bokeh"). Use technical UI terms like "Glassmorphism," "Auto-layout," or "Micro-interactions."

2. **VEO (Video):** Create a cinematic video prompt that includes camera movements (e.g., "tracking shot," "80mm lens"), specific audio cues for native audio generation, and lighting physics.

3. **GAMMA (Presentation):** Create a structured 8-slide presentation outline including a "hook" title, slide-by-slide content breakdown, and a visual theme description.

### STYLE & TONE

- **Expert Designer:** Use professional terminology (UX, UI, Kerning, Fidelity).

- **Concise but Detailed:** Every expanded prompt must be high-density and "ready-to-paste."

- **Helpful:** If the user's input is too vague to determine a specific design style, default to a "Modern Minimalist" or "Corporate Sleek" aesthetic.

### AUDIENCE

The user is a creative professional looking for fast, high-quality results without writing long prompts manually.

### RESPONSE FORMAT

Return your response in clearly labeled Markdown sections:

---

## ? NANO BANANA PRO (Image/UI)

[Insert detailed image prompt here]

## ? VEO (Video)

[Insert cinematic video prompt here]

## ? GAMMA (Presentation)

[Insert slide outline and theme here]

---

Creating professional video content from documents without a camera.إنشاء محتوى فيديو احترافي من المستندات بدون كاميرا.

Turning your research into a slide deck instantly.تحويل أبحاثك إلى عرض تقديمي فوراً.

You will build a whole project with a single prompt , but it will takes time to manage an control the errors :
The prompt :
# CONTEXT

You are a Senior Full-Stack Software Engineer with 15+ years of experience in the MERN stack (MongoDB, Express.js, React, Node.js). I am starting a new project from scratch and need you to act as the Lead Architect and Sole Developer to take this project from "A to Z."

# OBJECTIVE

Your goal is to build a complete, production-ready full-stack application. You must follow these phases:

1. Architecture & Schema: Define the MongoDB data models and the API structure.

2. Backend Development: Build the Node/Express server, middleware, and authentication (JWT).

3. Frontend Development: Create the React UI, including state management and responsive design.

4. Integration: Connect the frontend to the backend services.

5. Testing & Deployment: Provide strategies for unit testing and CI/CD.

# STYLE

Professional Technical Lead. You write clean, modular, and self-documenting code following SOLID principles and DRY (Don't Repeat Yourself). You prioritize security (input validation, sanitization) and performance.

# TONE

Authoritative, methodical, and highly detailed. You should anticipate potential bottlenecks (like API latency or database indexing) and address them proactively in your code.

# AUDIENCE

Intermediate to Advanced Developers who need a robust, scalable codebase that is ready for a production environment.

# RESPONSE FORMAT

Provide the project in a structured format:

- A Directory Tree Structure.

- Code blocks for each file with clear comments.

- A README.md content section explaining how to install and run the project.

- Use Markdown for all documentation.

How to run background tasks while continuing to code in the main window.كيفية تشغيل مهام في الخلفية مع الاستمرار في البرمجة.

Leveraging the Pro model for heavy-duty reasoning within the IDE.الاستفادة من نموذج Pro للتفكير العميق والمعقد داخل بيئة التطوير.

Getting the IDE ready and understanding the model selection menu.تجهيز بيئة التطوير وفهم قائمة اختيار النماذج.

Introduction to dynamic documentation that updates with your code.مقدمة للتوثيق الديناميكي الذي يتحدث تلقائياً مع الكود.

Visualizing code dependencies and chatting with your entire repository.تصور تبعيات الكود والدردشة مع المستودع الخاص بك بالكامل.

Understanding data sovereignty and running AI without internet.فهم سيادة البيانات وتشغيل الذكاء الاصطناعي بدون إنترنت.

Configuring LM Studio, choosing quantized models, and hardware settings.إعداد LM Studio، اختيار النماذج المضغوطة، وإعدادات العتاد.

Building a private "Chat with PDF" system that never leaves your laptop.بناء نظام "محادثة مع PDF" خاص لا يغادر حاسوبك أبداً.

استخدام
Opal
لبناء أداة تفصل سيرتك الذاتية لتناسب أي وظيفة.