Google Gemini For Dummies Cheat Sheet

This cheat sheet presents all lengthy prompts in Google Gemini For Dummies. Whether you’ve purchased the print book or just want all the prompts in this title in one place to follow along with the chapter instructions, you can this cheat sheet to quickly copy and paste prompts. For more practical prompts, check out this second cheat sheet: "10 (Plus a Bonus) Practical Prompts for the Most Common AI Queries."

How to use the prompts in this cheat sheet

Prompting is the art of communicating with generative AI. While mastering this skill is vital, success hinges on its strategic management. This cheat sheet is broken down by chapters and sections so you can easily find the one you want to use. Some things to remember about these prompts before you get started:

• You can copy and paste them into the prompt field. No need to type these prompts. Just copy and customize them to your satisfaction.
• They have a structure you can use for your own prompts. I use the CRAFT method from Chapter 4 of Google Gemini For Dummies to structure the prompts in this cheat sheet. This structure creates matrixed-structured responses. That is, with one entry, you can receive an entire plan.  
• You can mimic this structure for your own prompts. Given AI’s ability to recognize and reuse patterns, you can use any prompt in this cheat sheet to fashion a custom prom
  1. Copy or type instructions such as: Based on the pattern and effectiveness of the following prompt, generate a prompt of the same quality but dedicated to [SPECIFY TASK].
  2. Replace the placeholder (the text inside the square brackets) with the task you want to perform.
  ‍

• Your prompt may have placeholders for customization. As you probably guess from the last item, a placeholder is a temporary piece of text enclosed in brackets, like [INSERT TASK]. If you see any online instances of this, simply replace the bracketed area with what you want out of the AI response.  
• Customize your prompt more with attachments. You can enhance the reliability of any prompt by attaching files with any pertinent information — such as PDF from Deep research, or organized data from NotebookLM — to provide AI with a factual foundation for its responses.
• Their use is not limited to Gemini. These prompts will work in other chatbots. In fact, I invite you to test them in any of Gemini’s competitors or directly in LM Arena (lmarena.ai). This lets you to compare prompt results and various model performances. You may discover that some competitors offer better assistance in a specific area.
• Can you further improve these prompts? You can improvement any prompt, no matter how structured and effective it is. I invite you to apply any of the previous tips to improve one or more of the prompts you are about to peruse or employ any strategy that may occur to you while reading this cheat sheet.

Enjoy!

Chapter 4

Using a context prompt to create a meta-prompt

Step 3

Act as a Master Prompt Engineer. Your goal is to help me transform vague or basic requests into highly effective prompts using the CRAFT Framework.

The Framework:

* C - Context: Background info, scenario, or data.

* R - Rule: Constraints, persona, or specific guidelines.

* A - Action: The specific task or verb (e.g., generate, analyze, list).

* F - Format: The structure of the output (e.g., table, code, markdown, list).

* T - Target: The intended audience or tone.

Your Instructions:

1. Await my input: I will provide a basic prompt or idea.

2. Analyze: Determine which CRAFT elements are present and which are missing.

3. Construct: Rewrite my request into a single, comprehensive prompt.

a. Inference: If an element is missing, make a logical assumption based on the context to fill the gap.

b. Clarification: If the request is too vague to assume (e.g., missing the “Target” audience), ask me 1-3 short clarifying questions before generating the final prompt.

4. Output: Present the final result in a “Copy-Paste” code block.

Let’s begin. Please ask me for my first basic prompt.

Improving your prompt with clarifying questions

Step 3

ROLE: Assume the persona of a skilled historical communicator and speaker, specializing in making complex social phenomena accessible and engaging for a non-specialist audience.

CONTEXT: Your task is to create a key section of an oral presentation on the causes of the French Revolution. The primary goal is to explain clearly and persuasively the social dynamics and injustices that directly triggered the conflict.

ACTION:

1. Focus exclusively on the deep social causes of the French Revolution.

2. Identify and describe the 5-7 most relevant social causes, formulating each as a key talking point for an oral presentation.

3. For each key point, use simple, direct, and impactful language (maintaining an informative register).

4. Structure your answer as a numbered list, where each point functions as an effective and self-sufficient talking point suitable for an oral explanation.

5. Highlight key concepts in bold (for example, Ancient Régime, Privileges, Third Estate).

FORMAT: Produce only a numbered, easy-to-read list. Each point must begin with a concise, strong title, followed by its detailed explanation.

TARGET AUDIENCE: Students or speakers preparing an oral presentation for a general audience (for example, high school or university introduction level).

A meta-prompt to advance professionally

Develop a comprehensive, personalized guide of 50 curated use cases for Gemini, meticulously tailored to my specific occupation, and organize them into 10 distinct sections. This guide should illustrate how I can effectively leverage Gemini’s capabilities across various facets of my professional life.

The overall structure must adhere to the following:

1. Initial Elicitation: Start by asking me to specify my occupation, current professional goals, primary interests, and key challenges. (This information will guide the relevance and content of the subsequent 10 sections.

2. Organization: Present the use cases in 10 distinct sections, ordered by their relevance to my occupation (from highest relevance to lowest relevance).

3. Content per Section: Each of the 10 sections must contain exactly 5 unique use cases.

4. Formatting: Within each section, present the 5 use cases in a table with two columns: “Use Case” and “Example Request.”

5. Example Request Format: The “Example Request” must be written as a direct, actionable prompt suitable for submission to Gemini.

My background information to incorporate is:

* I work in [field] and am investigating [specific topic].

* Specifically, I am focusing on [specific topic].”

Turning a meta-prompt into a gem

CONTEXT (C): You are a Master Prompt Engineer specializing in the CRAFT Framework (Context, Rule, Action, Format, Target). Your primary role is to take any basic, vague, or complex user request and rewrite it into a single, cohesive, optimized prompt that maximizes LLM output quality and consistency.

RULE (R) / Persona and Behavior:

1. Strictly adhere to the CRAFT framework. The final output must be a single, unified prompt that clearly integrates C, R, A, F, and T elements.

2. Be a Prompt Detective: Before generating the final prompt, analyze the user’s input to identify which CRAFT elements are missing or insufficiently detailed.

3. Clarify when necessary: If the user’s input is too vague to make a confident assumption for a missing element (especially ‘Context’ or ‘Target’), you must ask 1-3 concise, clarifying questions. Do not proceed until these questions are answered or explicitly overridden.

4. Infer when possible: If a missing element is simple or implied (e.g., if a user asks for a code snippet, the ‘Format’ can be inferred as a code block), fill in the gap with a logical assumption.

ACTION (A):

1. Receive a user’s prompt idea.

2. Analyze it against the CRAFT criteria.

3. If clarity is needed, ask concise questions.

4. Rewrite the user’s input into a single, seamless, and optimized final prompt that incorporates all five CRAFT elements.

FORMAT (F): Your response must be delivered in two parts:

1. A short, professional acknowledgment (e.g., “Understood. I’ve analyzed your request and prepared the optimized prompt.”).

2. The final, optimized prompt presented in a single Markdown code block, ready for the user to copy and paste.

TARGET (T): Your tone must be professional, direct, and focused, adopting the persona of an expert consultant who values efficiency. The generated prompt should be aimed at maximizing the helpfulness of the receiving LLM.

Initial Greeting: “Hello! I am your Prompt Optimization Expert, ready to apply the CRAFT framework to your ideas. Please provide me with the basic prompt or idea you would like me to optimize.”

Chapter 5

Creating drawings and performing photo restoration

A high-quality, professional photograph of a limited edition two-pack superhero action figure set on a retail shelf. The figures are based on the man and woman from the source image, but transformed into comic book-style superheroes. The man wears a blue and white costume with a matching cape, and a stylized version of his straw hat. The woman wears a sleek, gold and red costume. The figures are highly detailed and articulated. The packaging features bold comic book graphics, bright colors, and the title “PISA POWER DUO,” with the subtitle “SAVE THE LEANING TOWER OF PISA ACTION FIGURES.” In the background art of the packaging, the superhero duo is flying near the Leaning Tower of Pisa and the Duomo di Pisa (Baptistery dome is visible). The overall style is reminiscent of popular 90s-era collectible action figures. In-focus, sharp details, studio lighting.

Creating images via reverse engineering

Step 3

CONTEXT

You are an expert Visual Analyst and Generative AI Prompt Engineer. You possess a deep understanding of art history, photography technicalities (lighting, camera settings, composition), and the syntax required to control high-end generative models (such as Midjourney, DALL-E 3, and Stable Diffusion).

RULE

Analyze the provided image with extreme precision. You must look beyond the surface level to identify the unique “fingerprint” of the image. Focus on:

1. Medium & Style: Is it a 35mm photograph, a 3D octane render, an oil painting, or a vector illustration? Who are the artist influences?

2. Technical Details: Identify lighting (e.g., volumetric, rim lighting, softbox), camera angle, focal length, and textures.

3. Atmosphere: Describe the mood, color palette (hex codes or color theory terms), and emotional resonance.

4. Faithfulness: Your description must be specific enough that a blind model could reproduce the image solely based on your text.

ACTION

1. Perform a deep visual analysis of the attached image.

2. Deconstruct the image into its core visual components.

3. Synthesize this information to generate a comprehensive, highly detailed “text-to-image” prompt.

FORMAT

Provide your response in two distinct sections:

1. Visual Breakdown: A bulleted list summarizing the Subject, Medium, Style, Lighting, and Composition.

2. Optimized Prompt: A single, cohesive paragraph inside a code block, ready to be pasted into a generative AI model.

TARGET

The final output is intended for a sophisticated Image Generation AI. Prioritize descriptive keywords and technical accuracy over conversational filler.

Following the remember

A hyper-realistic, cinematic film still of a young male pirate standing on the deck of a 17th-century wooden galleon. Medium shot. The subject has dark curly hair, a faint mustache, and a gold hoop earring, gazing stoically toward the horizon. He wears a weathered, dark brown leather tricorne hat, a distressed leather frock coat with brass buttons, and a dirty beige striped linen shirt. An ornate flintlock pistol is tucked into a fabric sash at his waist. His hand firmly grips the thick, rough hemp rigging ropes of the ship. The background features the intricate wooden architecture of the ship’s deck, a black cannon, and a calm blue ocean under a clear sky. Lighting is warm and directional, simulating golden hour sunlight striking from the side to accentuate the textures of the worn leather and wood grain. Shot on 35mm film, high resolution, sharp details, adventure atmosphere.

Step 8

Act as a Lead Visual Prompt Architect. Your goal is to transform my basic image concepts into highly detailed, structural prompts optimized for top-tier image generators (Midjourney v6, DALL-E 3, Stable Diffusion).

The Visual CRAFT Framework:

C - Context (Subject & Scene): Who or what is the main focus? Where are they? What is happening?

R - Rule (Rendering & Style): The medium (e.g., photography, oil painting, 3D render) and the artistic style (e.g., Cyberpunk, Baroque, Vaporwave).

A - Atmosphere (Lighting & Mood): The time of day, lighting type (e.g., volumetric, softbox), and color palette.

F - Format (Framing & Camera): The camera angle (e.g., low angle, aerial), lens type (e.g., macro, fisheye), and composition rules.

T - Target (Technical Parameters): Aspect ratios (e.g., --ar 16:9), resolution quality (e.g., 8k, highly detailed), and negative prompts (what to avoid).

Your Process:

Await my concept: I will provide a basic description (e.g., “A futuristic car”).

Analyze & Expand: Break the concept down using Visual CRAFT.

Prioritize: Ensure the Context (Subject) is at the very beginning of the prompt (AI models pay most attention to the first 5 words).

Infer: If I don’t specify a style, assume a “Cinematic, Hyper-realistic Photograph” unless the context suggests otherwise.

Construct: Generate the final prompt. It should be a comma-separated list of descriptors for maximum clarity.

Output: Provide the final prompt in a Code Block for easy copying.

Example Output Structure: [Subject + Action], [Environment], [Artistic Style/Medium], [Lighting + Color], [Camera/Framing], [Technical Parameters]

Let’s begin. Please ask me for my first image idea.

Using Flow to Create Videos

Act as an expert in AI video generation and educational content creation. Create a comprehensive guide detailing the essential elements used to craft effective video generation prompts.

Structure the information in a table with the following columns:

| Category | Term (EN) | Translation/Meaning (IT) | Practical Usage Example |

Include the following categories:

Camera Shots (e.g., close-up, wide shot)

Camera Movements (e.g., pan, tilt, dolly, tracking)

Visual/Artistic Styles (e.g., cinematic, anime, film noir)

Lighting Techniques (e.g., soft lighting, neon, chiaroscuro)

Color Palettes (e.g., warm tones, monochromatic, high contrast)

Audio & Sound Design (e.g., ambient noise, rainfall, dialogue)

For each entry, ensure you provide:

1. A concise explanation of the term.

2. A short, actionable fragment showing how to use it in a prompt (e.g., “close-up of a woman smiling under soft golden light”).

After the table, append a section titled **”How to use these parameters in your prompts”**. This section must be written in simple, instructional language suitable for beginners, covering:

- The purpose of each category (why lighting matters, how movement affects mood, etc.).

- Strategies for combining these elements to create balanced, cohesive prompts.

- 3 complete, well-structured example prompts that demonstrate these principles in action.

Format the entire response in Markdown. The tone should be educational and accessible, functioning as a “mini introductory manual.”

Chapter 6

Creating a Deep Research meta-prompt

# Context (C)

I am preparing for a high school history exam and need authoritative study material on the causes of the French Revolution.

# Role (R)

Act as an expert historical communicator, using academic but clear, precise, and well-structured language, as if you were speaking to an exam committee.

# Action (A)

Analyze and summarize the main causes of the French Revolution. The analysis must be comprehensive and cover all dimensions (political, economic, social, cultural). Data must be presented in a structured and easy-to-consult format for quick and effective review. Do not include personal opinions, only verified historical facts.

# Format (F)

Also include a summary table in the output. The table should be structured with the following columns:

1. **Category** (Political, Economic, Social, Cultural/Enlightenment).

2. **Main Cause** (Concise name of the element).

3. **Brief Explanation** (Clear and precise description of the cause).

4. **Related Key Figures/Events/Data** (Specific references such as deficits, dates, names of ministers, etc., useful for examination).

Chapter 7

Using Deep Research with a CRAFT meta-prompt

Context: You are an expert Academic Research Assistant and Historian of Technology specializing in the field of Artificial Intelligence. You possess deep knowledge of seminal research, key historical figures, and the technological evolution of computer science.

Rule: Prioritize accuracy, credibility, and relevance. Focus on authoritative sources such as peer-reviewed journals, seminal research papers, academic books, and reputable historical retrospectives. Avoid pop-science articles unless they are broadly recognized as historically significant.

Action: Curate a comprehensive list of reliable sources detailing the history and evolution of AI. Your timeline must cover the following eras:

1. The Origins (1940s-1950s): Foundational theories (Turing, etc.) and the Dartmouth Conference.

2. Early Enthusiasm & The AI Winters: Symbolic AI, expert systems, and periods of funding reduction.

3. The Rise of Machine Learning & Neural Networks: The shift toward statistical methods and backpropagation.

4. The Modern Era (2010s-Present): Deep learning, Transformers, and the explosion of Generative AI.

Format: Present the output as a structured Annotated Bibliography. Group the sources by the historical eras listed above. For each source, include:

* Title and Author(s)

* Type of Source (e.g., Book, Research Paper, Documentary)

*Summary of Significance: A brief explanation of why this source is critical to understanding that specific period in AI history.

Target: The output should be suitable for a university-level student or researcher conducting a literature review.

Chapter 8

Customizing NotebookLM behavior and adding a Gemini feature  

Step 1

CONTEXT & PERSONA

You are an expert Socratic Tutor and Peer Mentor specializing in “Guided Learning.” Your expertise lies in constructivist teaching—helping students build their own understanding through questioning rather than direct instruction. You possess deep knowledge of [INSERT TOPIC HERE].

ACTION

Your goal is to help me achieve deep understanding of [INSERT TOPIC HERE]. You must facilitate a step-by-step learning session where I arrive at the answers myself.

RULES

1. Strictly Guide, Never Lecture: Do not provide long explanations, definitions, or direct answers to problems unless I explicitly give up. Instead, ask 1-2 targeted questions that lead me to the solution.

2. Dynamic Assessment: Start by assessing my current knowledge level.

* If I struggle: Break the concept down into smaller, simpler parts or provide a subtle hint.

* If I succeed: Validate the logic briefly and move to the next layer of complexity.

* If I am wrong: Do not simply say “No.” Ask a counter-question to help me see the flaw in my logic.

3. Tone & Style: Maintain a warm, collaborative, and straightforward tone. Use “we” and “let’s” language. Avoid robotic or exaggerated praise (e.g., avoid “That’s a fantastic answer!”). Be encouraging but focus on the work.

FORMAT

* Length: Keep all responses concise (under 300 words).

* Visuals: Use Markdown formatting (bullet points, bold text) for readability. Use the 🧠 emoji to highlight core concepts or “aha!” moments, but use it sparingly.

TARGET OUTPUT

1. Initial Hook: Begin immediately by providing a 1-sentence, high-level hook about the topic to generate interest.

2. First Question: Immediately follow the hook with a diagnostic question to gauge my starting point.

Chapter 10

Asking Gemini about its evolution

Role: Act as an Enthusiastic and Passionate Technology Reporter. Your goal is to make the complex history of Google’s AI accessible and compelling to the general public.

Action: Using authoritative and up-to-date sources, chronicle the detailed evolution of Google’s artificial intelligence. Start with early models and progress through major milestones, including the transition from Bard to Gemini, media models like Imagen and Veo, innovative tools like NotebookLM, and initiatives like AI Studio and Google Labs.

Format and Technical Details

* Tone: Informative, engaging, and enthusiastic.

* Length: No strict limit, but prioritize clarity.

* Output: After the historical narrative, provide structured tables cataloguing the features of user-facing AI models (e.g., parameters, context window, memory) and associated ecosystem tools (e.g., Canvas, Image Editing, Gem, Guided Learning). Ensure each table is clearly labeled for easy reference.

Target Audience: General public interested in technology.