Spec-Driven Implementation FAQs

Question 1

Does it support frontend development and UI testing?

Accepted Answer

Yes, it includes a browser automation mode that uses a browser skill to visually verify UI changes, monitor console errors, and perform accessibility audits as part of the implementation's acceptance criteria.

Question 2

Where does the skill store its temporary artifacts?

Accepted Answer

By default, all implementation artifacts are stored in 'tmp/ship', but this directory is configurable via the CLAUDE_SHIP_DIR environment variable.

Question 3

Can I use this skill if I already have a JSON specification?

Accepted Answer

Yes, the skill is designed with entry points for three distinct phases. If you already have a spec.json, you can skip Phase 1 and start directly at the Prepare or Execute phases.

Question 4

How does the skill ensure that the AI doesn't produce broken code?

Accepted Answer

It enforces 'right-sized' stories that fit within a single context window and requires 'Typecheck passes' as a mandatory acceptance criterion for every story, alongside optional linting and testing commands.

Question 5

What is the primary function of the Implement skill?

Accepted Answer

It automates the process of turning a technical specification (SPEC.md) into a finished feature by breaking it down into small, verifiable user stories and executing them through an iterative subprocess loop.

Spec-Driven Implementation

Spec-Driven Implementation

주요 기능

사용 사례

주요 기능

사용 사례