मुख्य सामग्री पर जाएं

Spec-Driven Development with Claude Code: A Guided Tutorial

Learn how to write a spec, turn it into a plan, and let Claude Code build using spec-driven development. Compare Superpowers, Spec Kit, and BMAD-METHOD to find the right tool for your workflow.
अद्यतन 19 मई 2026  · 15 मि॰ पढ़ना

Vibe-coding with Claude Code works fine on small jobs. You describe a change, the agent writes it, and you check the result. The trouble starts when a feature touches many files at once. By then, the hard part is the design decision, not the implementation.

Spec-driven development handles that design decision in writing, before any code runs. You write a short spec that says what the change should do. You turn the spec into a plan of numbered tasks. Claude Code then writes code against the plan, one task at a time, with a human review between every step.

This tutorial teaches the workflow end-to-end. It walks through three open-source setups that run it inside Claude Code: Superpowers, GitHub Spec Kit, and BMAD-METHOD.

Spec-Driven Development in Claude Code FAQs

What is spec-driven development in Claude Code?

Spec-driven development is a workflow built on three documents in order: one that tells what a change should do, a plan that specifies the steps, and code written against the plan, with a human review between every pair.

How is it different from Claude Code's built-in plan mode?

Plan mode produces a plan inside a single chat turn, in memory, with no persisted spec and no review step. Spec-driven development persists both files on disk, runs each through a human review, and survives across sessions.

Which framework should I start with: Superpowers, GitHub Spec Kit, or BMAD-METHOD?

Start with Superpowers for solo work on a single repo. Reach for the Spec Kit when the spec needs to live in Git and be read by people who never open Claude Code. Reach for BMAD-METHOD when several people work across distinct roles.

How do I install Superpowers in Claude Code?

One command inside Claude Code: /plugin install superpowers@claude-plugins-official. A SessionStart hook auto-loads the workflow, so there is nothing to wire up per project.

What happens when the spec turns out to be wrong mid-implementation?

The loop stops and asks. In Superpowers, you edit the spec and regenerate the affected tasks. In Spec Kit, you run /speckit.analyze to surface the contradiction. In BMAD, a "FAIL" verdict from Phase 3 sends you back to Phase 2 to rewrite the PRD.

विषय
एआई एजेंट्स
कृत्रिम बुद्धिमत्ता