What are the three tiers of documentation involved?

The system uses Architecture Patterns (Tier 1: Conceptual rules), Reference Implementations (Tier 2: Concrete file paths in the repo), and Code Snippets (Tier 3: Copy-pastable boilerplate).

When should I update my pattern documentation?

Documentation should be updated whenever a core implementation changes. The skill includes a 'last_verified' metadata standard to help track staleness and ensure the AI relies on accurate information.

How does this skill reduce token usage?

By documenting patterns once in small, specialized markdown files (~2k tokens), Claude avoids having to re-scan and process large source files (10k+ tokens) every time it implements a similar feature.

Is this skill only useful for AI agents?

No, while it is optimized for AI token limits, it serves as an excellent standardized onboarding guide and reference manual for human developers joining a project.

Code Pattern Extraction

Name: Code Pattern Extraction
Author: violetio

byvioletio

0•

Learning & Documentation

Standardizes and documents code patterns across three tiers to minimize token usage and improve development consistency.

The Pattern Extraction skill provides a structured framework for identifying and documenting recurring architectural and code patterns within a repository. By organizing knowledge into conceptual architecture, concrete reference implementations, and copy-pastable snippets, it enables AI agents and developers to understand a codebase's requirements without expensive, repetitive exploration. This system significantly reduces token consumption during feature implementation while ensuring all new code adheres to established standards and project-specific best practices.

Key Features

01Workflow for cross-system consistency and architectural alignment

020 GitHub stars

03Three-tier documentation system: Architecture, Reference, and Snippets

04Token usage optimization for AI coding agents and LLM contexts

05Standardized metadata for pattern verification and maintenance

06Built-in templates for documenting new system modules quickly

Use Cases

01Onboarding new developers or AI agents to a complex codebase with established patterns

02Reducing API costs and token overhead when using Claude Code on large repositories

03Enforcing architectural standards and code style across microservices

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add violetio/violet-ai-plugins pattern-extraction

For use in Claude.ai and ChatGPT

Download Skill