A lightweight markdown-based workflow for collaborating with AI coding assistants using spec-driven development. Enables operations teams to build predictable software with a repeatable AI-guided workflow. Connects to AI coding assistants like Claude and integrates with development tools.
git clone https://github.com/liatrio-labs/spec-driven-workflow.gitA lightweight markdown-based workflow for collaborating with AI coding assistants using spec-driven development. Enables operations teams to build predictable software with a repeatable AI-guided workflow. Connects to AI coding assistants like Claude and integrates with development tools.
1. **Prepare Your Spec**: Create a markdown spec document with clear sections for requirements, endpoints, data models, and validation rules. Store it in your project repo (e.g., `SPEC.md`). 2. **Initialize the Workflow**: Start with the first task from your spec. Use the prompt template with [PROJECT_NAME], [PROJECT_DESCRIPTION], and the first [TASK_DESCRIPTION]. 3. **Iterate with AI**: After receiving the implementation, review the diff and validation. Use the AI's questions/deviations to refine your spec or implementation. Commit changes to a feature branch. 4. **Automate Testing**: Run the provided test cases locally (or in CI) to verify the implementation. Update the spec's test cases if new scenarios emerge. 5. **Progress Through Spec**: Use the AI's next task suggestion to continue the workflow. For complex tasks, break them into subtasks in your spec. Tag completed tasks with `✅` in the spec document. **Pro Tips:** - Use `<!-- TODO: [description] -->` comments in your spec for AI to implement - Maintain a `CHANGELOG.md` to track spec updates from AI questions - For large projects, split specs into logical modules (e.g., `auth/SPEC.md`, `users/SPEC.md`) - Integrate with your IDE's AI assistant (VS Code + GitHub Copilot) for in-editor workflow
No install command available. Check the GitHub repository for manual installation instructions.
git clone https://github.com/liatrio-labs/spec-driven-workflowCopy the install command above and run it in your terminal.
Launch Claude Code, Cursor, or your preferred AI coding agent.
Use the prompt template or examples below to test the skill.
Adapt the skill to your specific use case and workflow.
I'm working on [PROJECT_NAME], a [BRIEF_PROJECT_DESCRIPTION]. Here's the current spec in markdown format: ```markdown [SPEC_CONTENT] ``` Please implement the next task: [TASK_DESCRIPTION]. Follow the spec exactly, including: - [SPEC_SECTION_1] - [SPEC_SECTION_2] - [SPEC_SECTION_3] After implementation, validate against these test cases: ``` [TEST_CASES] ``` Provide: 1. The code changes in a diff format 2. Explanation of how it meets the spec 3. Any deviations or questions about the spec 4. Next task suggestion based on the spec
### Implementation for Task: Implement User Authentication Endpoint
**Diff:**
```diff
+ // src/controllers/authController.ts
+ export const loginUser = async (req: Request, res: Response) => {
+ const { email, password } = req.body;
+
+ // Validate input
+ if (!email || !password) {
+ return res.status(400).json({ error: 'Email and password are required' });
+ }
+
+ // Find user
+ const user = await User.findOne({ email });
+ if (!user) {
+ return res.status(401).json({ error: 'Invalid credentials' });
+ }
+
+ // Verify password
+ const isMatch = await bcrypt.compare(password, user.password);
+ if (!isMatch) {
+ return res.status(401).json({ error: 'Invalid credentials' });
+ }
+
+ // Generate token
+ const token = jwt.sign({ userId: user._id }, process.env.JWT_SECRET!, {
+ expiresIn: '1h',
+ });
+
+ res.json({ token });
+ };
```
**Validation Against Spec:**
✅ Implements POST /api/auth/login endpoint
✅ Requires email and password in request body
✅ Validates input presence
✅ Returns 400 for missing fields
✅ Uses bcrypt for password comparison
✅ Generates JWT token with 1h expiration
✅ Returns 401 for invalid credentials
✅ Returns 200 with token on success
**Test Results:**
```
✅ Valid credentials → 200 with token
✅ Missing email → 400 error
✅ Missing password → 400 error
✅ Invalid email → 401 error
✅ Invalid password → 401 error
```
**Questions/Deviations:**
1. Spec mentions rate limiting but not implemented - should we add this?
2. Password strength requirements not specified - implement basic check?
**Next Task Suggestion:**
Based on the spec's "Authentication" section, next would be implementing the token refresh endpoint.Create and collaborate on interactive animations with powerful, user-friendly tools.
Searchable AI prompt library for all major models
Cloud ETL platform for non-technical data integration
Get more done every day with Microsoft Teams – powered by AI
Customer feedback management made simple
Enterprise workflow automation and service management platform
Take a free 3-minute scan and get personalized AI skill recommendations.
Take free scan