We put together a checklist for you check your mastery of prompting AI tools. We broke it down into three sections:

  • General prompting

  • Image prompting

  • Video prompting

    Of course, the AI landscape is constantly changing so these checklists will get updated from time-to-time. Email us if you have any recommendations or changes you think we should add: team@projectcazimi.xyz

    (Updated October 2025)

General AI prompting

  • Know how to create system prompts

    • Error-Prevention

      • 1. Reality Anchoring - Tie responses to verifiable information 

      • 2. Self-Verification - Ask the AI to check its own work 

      • 3. Explicit Constraints - Define what NOT to do 

      • 4. Fallback Protocols - Define how to handle uncertainty 

      • 5. Fact Separation - Distinguish facts from opinions/analysis

    • Bias Mitigation Strategies: 

      • 1. Perspective Rotation - Explicitly request multiple viewpoints 

      • 2. Neutral Framing - Present questions without leading language 

      • 3. Demographic Balancing - Request inclusive examples 

      • 4. Counter-Stereotype Prompting - Ask for examples that break stereotypes 

      • 5. Evidence Balancing - Request pros/cons or supporting/ opposing evidence

  • Know how to create metaprompts

  • Few shot vs zero shot prompting

  • Know what Chain of Thought Reasoning is

  • Output formats

    • Format Control Techniques: 

    • 1. Explicit Structure Definitions - Detail exactly how information should be organized 

    • 2. Format Examples - Show examples of desired formatting 

    • 3. Markdown/HTML Specifications - Request specific formatting syntax 

    • 4. Tables and Lists Parameters - Define how data should be arranged 

    • 5. Custom Delimiter Systems - Create your own organization system

Image AI prompting

  • Make images in ChatGPT, NanoBanana and Midjourney

  • Understand differences between each model

    • know the limitations and biases of different AI models

  • Consistent characters/visual aesthetic

    • create reproducible results through systematic approaches

  • Workflow to transfer image styles from one model to another

  • Token efficiency to prioritize the most important elements first

  • Understand negative prompts

  • recognize when a concept is beyond current AI capabilities

  • Uses creative workarounds for difficult concepts

  • Can reverse-engineer effective prompts from existing images

  • Knows how to use camera/lens terminology (50mm, wide angle, macro, bokeh)

  • Can specify materials and textures (matte, glossy, weathered, crystalline)

  • Understands how to describe atmosphere and mood

  • Uses color theory concepts (complementary colors, monochromatic, warm/cool palettes)

Video AI prompting

  • Create a veo3 video

  • Create a sora video

  • Create a consistent video from two different video outputs

  • Build a JSON style prompt from Metaprompting

  • Create a character and use in various scenes

  • Create a character from an image tool

  • Dialogue that makes sense

  • Understand differences in models like Kling and Veo3

  • Make a Music Video