What is the outcome?
Ask for the thing you need to use, not a broad topic. “Draft a 45-second UGC script” is clearer than “help with my video.”
Prompting, explained
Better prompting starts before you type
AI gives weak answers when the task is vague, the context is missing or the standard for success lives only in your head. Learn to define the outcome, supply relevant context, set boundaries, show examples and ask for a result you can actually check.
Use these packs as worked examples, not magic words. Notice how each prompt separates investigation from action, limits scope, asks the agent to state its evidence and ends with verification. Then adapt the structure to your own work.
The anatomy of a stronger prompt
01
02
03
04
05
06
Six questions to answer before you press send
You do not need complicated prompt jargon. You need to remove the ambiguity that forces a model to guess. Start with these six pieces and add only what the task needs.
What context changes the answer?
Share the audience, goal, inputs and relevant background. Leave out unrelated history that hides the signal.
What are the boundaries?
State what the model may change, must preserve, should avoid and needs to ask before doing.
What does good look like?
Give a short example, rubric or acceptance criteria. A model can aim at a standard it can see.
Should it inspect before acting?
For complex work, separate diagnosis from execution. Ask it to show assumptions, risks and missing information first.
How will you verify the result?
Require sources, tests, calculations or a checklist. Fluent output is not the same thing as correct output.
Loading the free packs...
Basic
Go from copying prompts to designing better ones.
The free packs show the method. Basic opens every published pack across 15 topics, with research, search and sources included. Each pair shows you how to inspect first, act second and verify last— a pattern you can reuse far beyond code security.
€9
per month
- 3 days free. Every pack and the whole library, from the first minute.
- Nothing charged until day four. Cancel before then in one click and pay nothing.
- No lock-in. Monthly, cancel any time, access to the end of the month you paid for.
One avoided rework loop can repay the month. The prompting skill stays yours.
Start your 3-day free trial Already have an account? Sign in
Need prompts written for your exact situation, and an email the moment a vulnerability or a new model lands? That is Premium, EUR 24 a month, same trial.
Your situation is specific
Still not getting the result you need?
Bring the prompt, the workflow and the outcome you are aiming for. We will work through the real problem with you and leave you with a structure you can reuse.
Book a prompting consultation