One Human, One AI, One Verifiable Step: Building MyZubster with Nicola

Description and use

Nicola and an AI built a fully verifiable end‑to‑end observation workflow for the MyZubster MVP, adding automated tests, safe JSON persistence, and strict input validation.

Problem

Ensuring that an independent MVP can create, persist, and retrieve observations reliably while providing automated verification and robust input validation.

Outcome

Pull request #1 passed all tests, demonstrating a verifiable observation workflow without modifying main directly.

Workflow

• Create observation via API
• Receive HTTP 201 Created
• Persist observation to JSON file
• Retrieve observation through API
• Automated tests confirm all steps

Evidence

We completed one user story and made it independently verifiable.

Source

The implementation now verifies that the API accepts a valid observation; the server returns the correct 201 Created response; the observation receives an ID; the data is actually written to JSON; the saved observation can be retrieved through the API; invalid JSON, descriptions and coordinates are rejected; automated tests confirm the complete flow.

Source

GitHub Actions executed the complete test suite. Result: successful.

Source