# Generate test doubles from the API description

Test stand-ins written by hand, replaced with ones generated from the specification so they cannot fall behind.

Category: Mocks, fixtures, and test doubles
Source: https://letsship.ai/use-cases/generate-doubles-from-the-api-description
Published: 2026-08-25

## Goal

Replace hand-written stubs with doubles generated from the API specification.

## Acceptance criteria

- Doubles are generated by a committed command rather than edited by hand
- Regenerating produces no diff when the specification is unchanged
- Every hand-written stub it replaces is deleted
- The generation step runs in the pipeline

## What lands as proof

A regeneration that produces an empty diff, which is what proves the doubles and the specification agree.

## Prompt

Copy this to an agent to open the ticket:

```text
Create a Linear issue in my team and assign it to the SHIP agent.

Title: Generate test doubles from the API specification

Description: Replace hand-written stubs with doubles generated from the API specification.

Acceptance criteria:
- Doubles are generated by a committed command rather than edited by hand
- Regenerating produces no diff when the specification is unchanged
- Every hand-written stub it replaces is deleted
- The generation step runs in the pipeline

Before you create the issue, check whether this repository has a machine-readable API description that doubles could be generated from. If not, that is a prerequisite ticket, so say so.
```

## Questions

### What does the agent actually change?

The ticket is scoped to one outcome: replace hand-written stubs with doubles generated from the API specification. Work that serves that outcome is in scope, and anything outside it is left for a separate ticket, so the pull request stays reviewable.

### How do I know the work is done?

The pull request carries the evidence, not only the diff. Here that means doubles regenerate from the specification, so drift stops being possible, so a reviewer can confirm the result without reproducing the work locally.

### How much oversight does this need?

The run stops once the plan is written. Somebody reads the approach and approves it before any code exists, which is the cheapest moment to redirect the work.