Spec-Driven Development skill
A skill that interviews you and writes requirements.md, design.md and tasks.md so Claude Code and Cursor work from one plan
- 158 ★ on the repository
- Last commit
- MIT
Extended access
Why this level
- Creates agent config files such as `CLAUDE.md` and `.cursorrules` at the project root that steer the behavior of all your agents
Install
- Works in
In your terminal, with SkillFoxx CLI
npx skillfoxx add skills/spec-driven-developmentDetects the agents on your machine, shows the access level and pins the version.
Other ways to install
Access and checks
How we reviewExtended accessThe tool runs code or makes network calls and only reads external data.
- Creates agent config files such as `CLAUDE.md` and `.cursorrules` at the project root that steer the behavior of all your agents
This is third-party code. Review the repository files before installing.
What it does
The Spec-Driven Development skill builds three specification files in a project before any code exists. `requirements.md` states what the system must do, with REQ-001 style identifiers and acceptance criteria. `design.md` records the structure, and `tasks.md` splits the work into atomic steps such as TASK-003, each linked to a requirement and a check. For a new project the agent asks four short questions one at a time: what the project does and for whom, the stack, the deployment target and which AI tools the team uses. For an existing codebase the skill reconstructs the spec from your description and marks every assumption `[TO VERIFY]`, and the first phase of the plan is spec verification. After the spec the skill creates config files for each tool: `CLAUDE.md`, `.cursorrules`, `.windsurfrules`, `.github/copilot-instructions.md` and `.aider.conf.yml`. Each holds the same rule block: read the three files before acting, never invent requirements, stop and ask when the design must change.
Who it is for. Teams where Claude Code, Cursor and Copilot work side by side and each reads the task its own way.
Good fit when
- A project starts from scratch and needs requirements, a design and a work order before the first line of code
- Agents contradict each other: Cursor redoes what Claude Code did
- There is code but no system description, and it has to be reconstructed from your account
- A team adopts several AI tools at once and wants identical rules in every config file
- An agent invents scope and has to be held to requirements with testable criteria
Not a fit when
- The task is one file: four questions and three documents pay off only on a project that lives longer than a day
- The project already follows another method, such as OpenSpec or Spec Kit, with its own structure
- You want a code generator: the skill writes the specification and leaves the code to the agent
Example request
My AI keeps going off script. Interview me about a time-tracking project and create requirements.md, design.md, tasks.md and config files for Claude Code and CursorLimitations
- The author calls the release a public beta: the test suite is the author's own and testing by ordinary users on natural requests has not happened yet.
- For Aider the config file is only generated, while Claude Code, Cursor, Windsurf and Copilot are marked tested.
- Installing in Claude Code by cloning relies on the `CLAUDE.md` file at the repository root, while in the chat app the skill loads as a `.skill` file.
- The last commit in the repository dates from May 2026.
How to disable. Remove the skill on the Skills page in Claude settings or delete the cloned folder. The generated `requirements.md`, `design.md`, `tasks.md` and the agent config files stay in the project: delete them by hand or keep them as documentation.
FAQ
How is Spec-Driven Development different from OpenSpec and Spec Kit?
This skill runs a four-question conversation, creates three files and immediately emits matching config files for Claude Code, Cursor, Windsurf, Copilot and Aider. It has no CLI of its own: it is one `SKILL.md` with three reference files.
What if the code is already written?
Tell the agent that a codebase exists and descriptions do not. The skill drafts a spec from your account and marks unconfirmed spots, and the first phase of `tasks.md` is verification.
How do agents learn about the spec?
Every config file (`CLAUDE.md`, `.cursorrules` and others) holds a rule block that demands reading `requirements.md`, `design.md` and `tasks.md` before work. Deviating from the design needs your approval.
Which files does the skill create?
`requirements.md`, `design.md`, `tasks.md` and config files for the tools you choose: `CLAUDE.md`, `.cursorrules`, `.windsurfrules`, `.github/copilot-instructions.md` and `.aider.conf.yml`.
What does the skill do with unclear requirements?
It asks a clarifying question and invents no requirements. If the requirements lack a section on what stays out of the first version, the skill flags it and asks what to leave out of scope.
Discussion
No questions or reviews yet.
Related
SkillsTwo skills for spec-driven development: /spec asks questions and saves a document to specs/, /spec-impl implements an approved spec step by step
A plugin with three skills for large development work: project analysis, phased tasks tracked as GitHub Issues, parallel agents and code review
Eleven skills that take an idea through requirements, spec and plan to a verified pull request with an independent review
Orchestrator skill: from an idea the agent writes requirements and a spec, splits work into tasks, builds the project, checks it
14 skills that find the gaps in a task before the code: interviews, plans, assumption tests and proof that a bug fix works
Eleven skills take an agent from an idea to verified code through a PRD and RFCs kept as files in the repository
Comparison with similar tools
Cards that share the main search query, and how they differ.
| Tool | Official | Repository stars | Main difference |
|---|---|---|---|
| Spec-Driven Development (this card) | no | 158 | |
| Spec Kit | no | 148,171 | GitHub's kit with the specify CLI that installs commands and skills for the whole cycle from principles to implementation |
| OpenSpec | no | 74,038 | proposal, requirements and design driven development with a separate change workflow |
README in short
The README starts from the problem of agents diverging and tables of spec files with sample requirements and tasks. Then come a quick start with trigger phrases, install routes for the Claude chat and Claude Code, scenarios for a new project, an existing codebase and a team, the shared rule block, a supported-tools table and a description of the test suite.
README badge
Original SKILL.md
--- name: spec-driven-development version: 1.0.0 description: > Use this skill to scaffold Spec Driven Development: requirements.md, design.md, tasks.md, and cross-compatible AI config files (CLAUDE.md, .cursorrules, .windsurfrules, .github/copilot-instructions.md). Triggers: 'requirements', 'spec', 'design doc', 'tasks', 'CLAUDE.md', 'cursorrules', 'SDD', 'spec driven', 'AI configuration', 'cross-AI', 'multi-agent', 'agent setup'. Also trigger for: 'set up my AI tools', 'keep agents consistent', 'stop AI going off-script', 'plan before coding', 'structure project for AI'. Retrofit triggers: 'document my system', 'document my codebase', 'describe my architecture', 'write up what we built', 'add specs to existing project', 'no specs yet', 'help me document', 'I already have a codebase'. Use whenever a user wants to start a project, configure coding agents, create spec docs, or prevent AI agents from drifting, hallucinating scope, or producing inconsistent implementations. --- # Spec Driven Development Skill This skill guides the user through creating a complete Spec Driven Development environment: - `requirements.md` — what the system must do - `design.md` — how it will be built - `tasks.md` — atomic, ordered implementation steps - AI agent configuration files for Claude Code, Cursor, Windsurf, GitHub Copilot, and Aider **Reference files** (read when needed): - `references/sdd-curriculum.md` — full teaching curriculum, newbie to hero - `references/capabilities-and-cross-ai.md` — Claude Code capabilities, limitations, cross-AI strategy - `references/templates.md` — all spec file and AI config templates --- ## When to Read Reference Files | Situation | Read | |---|---| | User is new to SDD and needs explanation | `sdd-curriculum.md` Levels 1–3 | | User asks about Claude Code + SDD specifics | `capabilities-and-cross-ai.md` Part 1 | | User wants cross-AI config files | `capabilities-and-cross-ai.md` Part 2–3 | | You need template content to generate files | `templates.md` | --- ## Workflow: First Run (New Project) ### Step 1: Interview the User **Conversational rule (non-negotiable):** Ask exactly ONE question, wait for the answer, then ask the next. Never present questions as a numbered list or bullet list — that feels like a form, not a conversation. A newbie who sees three bullets may answer only the first or feel overwhelmed. **The four required answers before any file is generated:** 1. What the project does and who uses it 2a. Tech stack (language, framework, database) 2b. Deployment target (Railway, Fly.io, AWS, Vercel, etc. — offer to suggest if unknown) 3. Which AI coding tools they use (Claude Code, Cursor, Windsurf, Copilot, Aider, other) Stack and deployment are separate required answers because one does not imply the other. "Node.js" tells you nothing about deployment. "Railway" tells you nothing about the language. Both must be known independently before generation proceeds. **If the user's opening message already answers some but not all of these**, acknowledge what you know and ask only for what's missing — still one question at a time. **Optional follow-ups** (ask only when the user's answer raises real ambiguity):