Where are the generated skills stored and how are they used?

All generated skills are saved to the `.claude/skills/` directory. Claude Code automatically references these skills to provide domain-specific guidance during development.

Does it require manual configuration for each subsystem?

It uses a `subsystem-registry.md` file and automated dependency scanning to identify stacks, though users can customize the agent prompts for specific needs.

Can this skill be used to generate security checklists?

Yes, it generates specific security.md and testing.md files for each language and subsystem to ensure AI-generated code meets safety standards.

Which programming languages and frameworks are supported?

It currently supports Python (FastAPI, Celery, PyTorch), Go (Cobra, Bubbletea), TypeScript, and React, with specific patterns for RAG, Electron, and more.

How does this skill handle inconsistent code patterns in the existing codebase?

The skill is designed to prescribe 'ideal' patterns based on industry standards and specified frameworks, rather than codifying existing technical debt or inconsistencies.

Subsystem Skill Architect

Name: Subsystem Skill Architect
Author: llama-farm

byllama-farm

•

801

•

Deployment & DevOps

Automatically generates specialized AI coding skills and best-practice checklists for every subsystem within a monorepo.

The Subsystem Skill Architect is a meta-utility designed to automate the creation of a comprehensive AI-assisted development environment. By analyzing monorepo structures, it identifies specific technology stacks—such as Python, Go, and TypeScript—to generate tailored Claude Code skills. It produces idiomatic pattern guides, security checklists, and performance benchmarks for varied subsystems like RAG pipelines, FastAPI servers, and React designers. By codifying 'ideal' patterns rather than replicating existing inconsistencies, it ensures that AI-generated code remains high-quality, production-ready, and aligned with modern industry standards.

Key Features

01Prescriptive 'Ideal Pattern' documentation to enforce architectural standards

02Parallel generation of shared language-specific skills for Python, Go, TS, and React

03Automated dependency analysis via pyproject.toml, package.json, and go.mod

04801 GitHub stars

05Structured output to the .claude/skills/ directory for seamless integration

06Framework-aware checklist generation for specialized subsystems like RAG and Electron

Use Cases

01Enforcing high-quality coding standards by providing Claude with domain-specific context

02Automating the creation of AI-driven style and security guides for new subsystems

03Standardizing development patterns across large, multi-language monorepos

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add llama-farm/llamafarm generate-subsystem-skills

For use in Claude.ai and ChatGPT

Download Skill