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