How does the parallel exploration work?

The skill launches multiple specialized sub-agents in the background. Each agent focuses on a different domain—such as frontend, backend, or data flow—and then synthesizes their findings into a single comprehensive report.

Is it compatible with specific frameworks?

While it works with any language, it features production-ready patterns and specialized sub-agents optimized for FastAPI, React 19, and LangGraph-based AI workflows.

Can this skill assess code quality?

Yes, it includes a dedicated Code Health Assessment phase that rates files on a 0-10 scale across dimensions like readability, maintainability, testability, and complexity.

Does it require manual task management?

No, the skill automates task management using TaskCreate and TaskUpdate tools to provide a live progress dashboard while the agents are working.

What is the 'Agent Teams' mode?

Agent Teams is an advanced mesh orchestration mode where explorers share discoveries in real-time. For example, when the structure explorer finds an entry point, it immediately alerts the data flow explorer to begin tracing it.

Codebase Explorer & Architect

Name: Codebase Explorer & Architect
Author: yonatangross

byyonatangross

•

학습 및 문서화

Automates deep codebase discovery and architectural analysis using parallel AI agents to map data flows, code health, and dependencies.

The explore skill is a high-performance orchestration tool for Claude Code designed for deep repository navigation. It dispatches a specialized team of parallel agents—including a Code Structure Explorer, Data Flow Explorer, Backend Architect, and Frontend Developer—to analyze your codebase from multiple angles simultaneously. Beyond basic file searching, it provides quantitative code health assessments, dependency hotspot mapping, and product-level context, making it the definitive tool for onboarding to new repositories, auditing legacy code, or planning complex refactors.

주요 기능

01Integrated 'Product Perspective' to align technical implementation with business logic

02Native memory integration for context-aware exploration using knowledge graphs

03Parallel agent orchestration for simultaneous code, data, and architecture analysis

04Quantitative code health scoring (0-10) across readability, maintainability, and complexity

0569 GitHub stars

06Automated dependency hotspot mapping with ASCII-based coupling visualizations

사용 사례

01Tracing complex cross-boundary data flows from frontend components to backend storage

02Onboarding to large, unfamiliar repositories or legacy systems to understand patterns quickly

03Performing technical health audits before major migrations or architectural refactors

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add yonatangross/orchestkit explore

For use in Claude.ai and ChatGPT