spec-driven-development

Execute spec-driven tasks with TDD and two-stage reviews.

4|1|Updated Apr 7, 2026
One-click install
npx skills add https://github.com/NSObjects/specpowers --skill spec-driven-development-nsobjects
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: spec-driven-development
Source: https://github.com/NSObjects/specpowers/tree/main/skills/spec-driven-development
Command: npx skills add https://github.com/NSObjects/specpowers --skill spec-driven-development-nsobjects

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

It prevents premature coding by enforcing a structured execution cadence that keeps every implementation linked back to approved specs, TDD discipline, and explicit user pacing.

Core Features & Use Cases

  • Task Execution Discipline: Operates on checked tasks.md entries ensuring each task runs through RED→GREEN→REFACTOR within a single uninterrupted session.
  • Two-Stage Reviews: Requires sequential spec compliance and code quality reviews with subagent dispatch when supported before reporting completion.
  • Guardrails & Modes: Prevents spec/design edits during execution, supports Step-by-Step or Fast modes, and triggers milestone verification loops every few tasks.

Quick Start

Use the spec-driven-development skill to run the approved tasks.md plan with two-stage reviews and user-controlled pacing.

Frequently Asked Questions about spec-driven-development

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
What is spec-driven development and when do I need it for code review?

Spec-driven development is a structured execution cadence keeping implementation linked to approved specs and TDD discipline. You need it for code review when a concrete plan defines tasks and spec scenarios, ensuring traceability and preventing premature coding.

Can I run spec-driven tasks with subagent coordination and pacing control?

Yes, you can run tasks with subagent coordination using Step-by-Step or Fast execution modes. The system dispatches subagents for sequential spec compliance and code quality reviews while respecting user-controlled pacing and user-managed git operations.

How do two-stage reviews work for spec compliance and code quality?

Two-stage reviews require sequential checks for spec compliance and code quality before reporting task completion. Subagents are dispatched when supported, ensuring every implementation task meets approved spec scenarios and quality standards without modifying the spec.

Are there limitations on modifying specs during task execution?

Yes, guardrails prevent spec and design edits during task execution to maintain traceability. The system triggers milestone verification loops every few tasks and operates on checked tasks entries, ensuring implementation follows the approved plan without unauthorized modifications.