Does it require external tools to function?

It leverages a suite of tools including WebSearch, GitHub search, and MCP memory connectors to gather, verify, and index the most accurate technical information available.

Can I update existing research files?

Yes, when you research a topic that has already been documented, the skill identifies the existing file and asks if you would like to update, replace, or cancel the operation.

Where are the research results stored?

All findings are synthesized into high-quality Markdown files and saved locally in your ~/.claude/learnings/ directory for persistent access.

What is the difference between Subagent and Team modes?

Subagent mode is optimized for well-defined, independent research dimensions, while Team mode is used for broad or vague topics that require real-time coordination and iterative discovery.

How does deep-learn differ from a standard web search?

Unlike a single search, deep-learn deploys multiple specialized agents to explore different aspects of a topic (such as API docs, edge cases, and community patterns) simultaneously to create a comprehensive synthesis.

Deep Research & Knowledge Base

Name: Deep Research & Knowledge Base
Author: ourines

byourines

•

Learning & Documentation

Conducts multi-agent parallel research to build a persistent, high-fidelity technical knowledge base within Claude Code.

Deep Learn is a sophisticated research orchestration skill that leverages parallel AI agents to explore complex technical topics from multiple dimensions simultaneously. By coordinating either independent subagents or collaborative teams, it synthesizes vast amounts of information—including official documentation, real-world code patterns, and common pitfalls—into structured Markdown files stored locally. It is the ideal tool for developers who need to master new frameworks, investigate intricate system architectures, or maintain a long-term searchable repository of technical expertise for instant recall during coding sessions.

Key Features

01Persistent local storage of structured knowledge in Markdown format

02Parallel agent orchestration for multi-dimensional topic exploration

03Built-in cross-validation and deduplication of research findings

041 GitHub stars

05Dynamic strategy selection between independent subagents and collaborative teams

06Automated synthesis of web search results, GitHub code, and documentation

Use Cases

01Mastering new frameworks like TanStack Router by researching patterns and gotchas

02Building a personalized, persistent technical library for quick reference during development

03Investigating complex infrastructure concepts like Kubernetes networking architectures

What are Skills?·How to Install

Install with 🐟 Skill.Fish

npx skillfish add ourines/claude-code-learn deep-learn

For use in Claude.ai and ChatGPT

Download Skill