KERNEL
Fundamentals
Level 1 · FoundationsA simple 5-field starter framework — Task, Input, Constraints, Output, Verify — that builds a solid prompt in one guided pass.
KERNEL is a beginner-friendly framework built around five fixed fields: what you want done (Task), what you're giving the model to work with (Input), what it must respect (Constraints), what shape the answer should take (Output), and how to check the answer is right (Verify). It's not a technique to combine with others — it's a complete, simple structure on its own, meant as the first thing to reach for before learning any other technique. Filling in even just Task alone already produces a usable prompt; the other four fields are there to fill in as you have the information, not required all at once.
When to use it
Use it as your default starting point, especially if you're new to structuring prompts.
When to avoid it
Move beyond KERNEL once your task needs a specific reasoning strategy, a persona, or a verification loop — at that point, look for a dedicated technique that matches what you actually need.
Advantages
- Nothing to learn beyond five plain-English fields
- Produces a reasonably complete prompt even with only Task filled in
- A single, consistent structure to fall back on for almost any task
Disadvantages
- Doesn't give you a reasoning strategy, persona, or verification loop — those need a dedicated technique
- Five generic fields can't capture everything a more specialized technique's own vocabulary would
Examples
Task: Write a two-paragraph product announcement. Input: new feature is faster search, launching next Monday. Constraints: keep it under 120 words, no jargon. Output: plain text, no headings. Verify: does it mention the launch date and the benefit clearly?
Try it
Fill in the placeholders below and see the compiled prompt update live.
Preview
Task: {{TASK}}
Input: {{INPUT}}
Constraints: {{CONSTRAINTS}}
Output: {{OUTPUT}}
Verify: {{VERIFY}}