What is the 'enrichment' process in Architect Detective?

Enrichment uses an LLM to generate metadata summaries of files and symbols. This allows the skill to search for 'meaning' (e.g., 'where is the business logic?') rather than just specific keywords or variable names.

Why are grep and find forbidden with this skill?

Architect Detective uses claudemem v0.2.0, which relies on semantic indexed memory. Standard grep and find tools are limited to raw text matching, whereas this skill understands the purpose, exports, and context of code components through LLM enrichment.

Can I use this for security audits?

While its primary focus is architecture and design, its ability to map data flow and system boundaries makes it a powerful supporting tool for identifying architectural security flaws and unauthorized data access patterns.

How does it identify design patterns?

It analyzes file exports, dependencies, and purpose descriptions within the enriched index to match code structures against known patterns such as Repository, Factory, Strategy, and Observer.

Architect Detective

Name: Architect Detective
Author: MadAppGang

byMadAppGang

•

177

•

Developer Tools

Analyzes codebase architecture and design patterns using semantic indexed memory to map system boundaries and dependencies.

Architect Detective is a specialized skill for Claude Code designed to perform high-level structural analysis of complex codebases. By leveraging the claudemem v0.2.0 indexed memory system, it moves beyond simple text matching to understand code through its purpose and meaning. It is specifically optimized for mapping system layers, identifying architectural patterns like Clean Architecture or DDD, and uncovering hidden dependency flows. This skill is indispensable for architects onboarding to new projects or auditing existing structures for technical debt and scalability bottlenecks, replacing traditional grep-based searches with LLM-enriched semantic discovery.

Key Features

01Automated layer mapping to visualize presentation, application, and domain boundaries

02Dependency flow analysis to identify circular imports and architectural leaks

03Design pattern detection (Repository, Factory, Strategy, etc.) via file summaries

04Automated generation of comprehensive system architecture reports

05177 GitHub stars

06Semantic codebase searching using vector embeddings and LLM enrichment

Use Cases

01Mapping dependencies and boundaries before a major refactoring or microservices extraction project

02Auditing system design to ensure adherence to architectural principles like SOLID or Hexagonal Architecture

03Onboarding to a large, unfamiliar codebase to quickly understand its high-level structure and entry points

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add madappgang/claude-code architect-detective

For use in Claude.ai and ChatGPT

Download Skill