Functions and scripts
Turn a specific input and expected output into a function or a repeatable automation script.
I need a small React app to track equipment loans for our team. Start with a checkout form, a list of current loans and a way to mark returns. Explain the structure and show me a first version. I will test it and bring back the issues we find.
01 / CHOOSE YOUR DIRECTION
Start with the behavior you need, the language or framework, and where the result will run.
Turn a specific input and expected output into a function or a repeatable automation script.
Build a page or component with the states and interactions your users need.
Connect screens, data and user actions into a working first version you can try.
Draft endpoints, validation and data handling around a clear contract.
Add meaningful checks or revise code after you share a failing case or error.
Understand a generated file, a design choice or the change needed for your project.
02 / HOW YOUR AGENT WORKS
Define what the code should do, inspect the first result, then use real test feedback to make the next change.
Give an example input, expected output, users and important edge cases.
Share the language, framework, existing files and constraints that matter.
Identify the files, components or functions needed before writing them.
Create code for the agreed scope and explain how the pieces fit together.
Try the result, check the behavior and record errors or missing cases.
Bring back a failing case or new requirement and revise the relevant code.
03 / THE LONG GAME
The first equipment-loan app works for one item, but your team finds that two people can check out the same item. Bring that case back with the first version and the rule your team needs.
Your agent can trace where availability is checked, revise the flow and add a test for duplicate checkouts. You run the changed code and decide whether it is ready to use.
04 / GOOD TO KNOW
Eazo is a new kind of personal agent: a relational agent built around an ongoing relationship with you. It works with the goals, people and context you choose to share, helping you start a task and keep moving as your needs change. You guide what it works on and decide which suggestions to act on.
A code generator may return a snippet for one prompt. With Eazo, you can start from a feature brief, review a first implementation, then bring back errors and new requirements to change the same project. You still run and review the code before using it.
Start with the scope you need. A function needs inputs, outputs and edge cases; an app also needs screens, data and user flows. Breaking a larger idea into reviewable parts makes it easier to test.
Yes. Include the language, framework, versions and existing project structure in your request so the generated code fits your environment.
Share the relevant files or explain the current structure and the change you want. Include errors or failing examples when you need a fix.
Generated code can contain bugs or miss requirements. Run it in your environment, review security-sensitive behavior and test the cases that matter before deploying.
Yes. Ask for tests tied to the behavior you expect and an explanation of the parts you need to maintain.
Tell your agent what happened when you ran it, include the error or missing behavior, and ask for a focused revision to the existing code.
YOUR NEXT MOVE
Describe the behavior you need, review the code, and improve it with real feedback.
Generate your code