backend-code-generator

Generates Express, Prisma, BullMQ, and LangChain backend code from architecture specification documents.

Updated Jun 5, 2026
One-click install
npx skills add https://github.com/fen-zx/NovelToScript --skill backend-code-generator-fen-zx
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: backend-code-generator
Source: https://github.com/fen-zx/NovelToScript/tree/main/.agents/skills/backend-code-generator
Command: npx skills add https://github.com/fen-zx/NovelToScript --skill backend-code-generator-fen-zx

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Writing a full backend project from design documents is repetitive and error-prone. This Skill converts architecture and API specification documents into a complete, consistently structured Express + TypeScript backend codebase, eliminating manual scaffolding of controllers, services, repositories, queues, and AI workflows. ## Core Features & Use Cases - Layered Code Generation: Produces Controllers, Services, Repositories, DTOs (Zod), Routes, and Middlewares following strict separation-of-concerns rules. - Infrastructure Scaffolding: Generates Prisma client setup, BullMQ queues and workers, Redis cache config, MinIO storage clients, and Pino logging. - AI Workflow Generation: Builds LangChain chains (PromptTemplate, RunnableSequence, OutputParser) for novel analysis, character extraction, scene generation, and YAML validation. - Use Case: Given ARCHITECTURE.md, API_SPECS.md, and AI_WORKFLOW.md for a novel-to-script application, generate the entire backend/src directory including GenerateScriptWorker, AIService, and all module routes. ## Quick Start Ask the AI to generate the complete backend project code based on your ARCHITECTURE.md and API_SPECS.md documents.

Frequently Asked Questions about backend-code-generator

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

FAQPage Schema
How do I generate an Express backend from architecture documents?▼

Provide specification documents such as ARCHITECTURE.md, API_SPECS.md, and DATABASE_SCHEMA.md as input. The generator produces the full src directory with controllers, services, repositories, routes, DTOs, queues, workers, and config following the documented structure.

What tech stack does the generated backend code use?▼

The generated code uses Node.js with Express and TypeScript, Prisma ORM with SQLite, BullMQ with Redis for queues, MinIO for object storage, LangChain with DeepSeek for AI workflows, Zod for validation, and Pino for logging.

How are LangChain AI workflows structured in the generated code?▼

AI logic is split into discrete chains such as NovelAnalyzerChain, CharacterExtractionChain, SceneGenerationChain, and YamlGenerationChain, each built from PromptTemplate, RunnableSequence, and OutputParser. A single monolithic prompt is explicitly forbidden.

Does the generated code enforce separation of concerns?▼

Yes. Controllers only handle requests and responses, services contain business logic and transactions, and repositories exclusively perform Prisma database operations. Business logic in controllers or HTTP handling in services is explicitly prohibited.

What are the limitations of generated backend scaffolding?▼

Generated code depends on the completeness of input specification documents; unspecified behavior is marked with [TODO] placeholders. Business rules, edge cases, and tests still require manual review and refinement before production deployment.