Fullstack coder
Skill batterfried-philosophy172/Agent-Startup-Skills/skills/fullstack-coder
Build SaaS products using an 8-phase AI agent pipeline with mandatory approval gates for Antigravity, Claude Code, and Codex.
npx -y skills add batterfried-philosophy172/Agent-Startup-Skills --skill fullstack-coderAssembled from the repository path, not quoted from the project. Check it against their README if it does not work.
One thing to look at
- 2 stars2 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.
What its author says it does
Copied from the file, not written here
Full-stack implementation agent that writes complete, production-ready code following an approved architecture and schema. Triggers on: write the code, implement features, build the app, code the MVP, generate codebase.
SKILL.md
2.8 KB, as published. Nobody here has run it
Full-Stack Coder — Implementation Agent
Purpose
Write complete, runnable, production-quality code that exactly follows the approved architecture, schema, and MVP scope. No shortcuts, no placeholders.
When to Use
- After architecture AND schema are approved
- When implementing approved MVP features
- Never before Phase 3 and Phase 4 of the SaaS Accelerator pipeline
Hard Rules
- Follow the architecture exactly. Use the approved tech stack, file structure, and API endpoints. Do not substitute technologies.
- No TODOs or placeholders. Every function must be fully implemented.
- No hardcoded secrets. All configuration uses environment variables.
- Every file has a header comment explaining its purpose.
- Error handling is mandatory. Every API endpoint must handle errors gracefully with proper HTTP status codes and user-friendly messages.
- Input validation is mandatory. Validate all user inputs on both client and server side.
Process
Step 1: Project Scaffold
Generate every file and folder from the architecture document.
Start with configuration files: package.json, .env.example, config files.
Step 2: Database Layer
Implement the schema from Phase 4:
- Migration files or ORM schema definitions
- Database connection utility
- Seed script for test data
Step 3: Backend API
Implement each API endpoint from the architecture document, one at a time:
- Route handler
- Input validation
- Business logic
- Error handling
- Response formatting
Step 4: Authentication
Implement the full auth flow:
- Registration (with password hashing)
- Login (with session/token creation)
- Protected route middleware
- Logout
Step 5: Frontend
Implement each page/screen from the MVP scope:
- Layout and navigation
- Forms with client-side validation
- API integration (data fetching and mutations)
- Loading and error states
- Responsive design
Step 6: Environment Configuration
Generate a complete .env.example with descriptions:
# Database
DATABASE_URL=postgresql://user:pass@localhost:5432/dbname
# Authentication
JWT_SECRET=your-secret-key-here
SESSION_EXPIRY=86400
# External Services
SMTP_HOST=smtp.example.com
Output
The complete codebase, presented file-by-file with clear separators.
Exit Criteria
- Every file from the architecture document exists
- Every API endpoint is fully implemented
- Authentication is complete
- Frontend renders and connects to the API
.env.exampleis complete- No TODOs, no placeholders, no stub functions