Hello there,
I'm Alex.

Freelancer doing Web-Technolgy-Stuff

Learning about Agentic Coding

Example Prompt

Implement a form to add a new TODO item in the TODO-app.
A TODO item has the following fields:
- `title` (string, required, minimum 3 characters)
- `description` (string, optional)
- `dueDate` (date, required)
On submit, validate all required fields. 
Do not allow submission if validation fails.  
Display inline error messages.
            

Requirements in the Prompt

• Vague defintion
• Contradictory with existing requirements
• Forgotten when chat ends
• No documentation
• No traceability over time

→ Volatile, but we want Stability

Want Stability? Have Tests!

Generating tests ... based on generated implementation

→ Might just validate a flawed implementation 🫠

Test-Driven-Development (TDD)

Write Tests First!

• Might be tedious and time-consuming
• Might not reflect original intent

Behavior-Driven-Development (BDD)

Self validating

System documentation that is automatically checked against the system's behaviour

Human Language

• Better Collaboration with non-devs
• Easy to understand for LLMs

→ Source of truth for human + AI Agent + Test runner

The one only Prompt

Read all `.feature` files and all rules
Repeat until 100% of tests pass:
	- Run the tests
	- Implement the code for failing tests