Getting Started with Claude Code¶

Claude Code is Anthropic's official CLI tool that transforms Claude into a development environment assistant. Unlike web-based interfaces, Claude Code runs locally on your machine with direct access to your filesystem, terminal commands, and development tools.

Official Documentation: Claude Code Overview

This document provides general background and information about Claude Code to help you understand its core concepts and capabilities. For hands-on lessons applying Claude Code to biocuration workflows, see the other parts of this tutorial: Exploring LinkML Models with Claude Code, Creating Slash Commands for API Lookups, and Using MCPs for Ontology Lookups. Please review this document to familiarize yourself with Claude Code before moving on to those practical lessons.

OBO Academy Training Session Part 1¶

What is Claude Code?¶

Claude Code is an interactive command-line interface that provides:

Direct file operations - Read, write, and edit files conversationally
Command execution - Run terminal commands with natural language
Tool integration - Grep, glob, git operations, and more
Project awareness - Understands your project structure and git status
Permission model - You control what Claude can execute

Key difference from web interfaces: - Web Claude: Upload files manually, copy/paste code, no direct execution - Claude Code: Direct filesystem access, automatic tool selection, command execution

Installation¶

Claude Code requires Node.js 18+ and is available via npm:

npm install -g @anthropics/claude-code

Verify installation:

claude --version

Launching Claude Code¶

Start Claude Code in any directory:

claude

You'll see:

Claude Code vX.X.X
Type /help for commands or just start chatting!

Working directory: /your/current/path
Git status: Clean (on branch main)

Claude Code automatically detects: - Current directory and file structure - Git repository status - Available tools and permissions

Tool Overview¶

Claude Code has access to several built-in tools:

Tool	Purpose	Example Use
Read	Read file contents	Schema files, code, documentation
Write	Create new files	Scripts, config files, data
Edit	Modify existing files	Fix typos, update code, refactor
Bash	Execute commands	Git operations, build commands
Grep	Search file contents	Find patterns, search codebase
Glob	Find files by pattern	List all .yaml files
WebSearch	Search the web	Look up API documentation

Claude automatically selects the appropriate tool based on your request.

Permission Model¶

Claude Code will ask permission before: - ❌ Writing or modifying files - ❌ Executing shell commands - ❌ Deleting anything

Permission dialog example:

Claude wants to run: `git status`
Allow? (yes/no/always):

Options: - yes - Allow this once - no - Deny this action - always - Auto-approve this type of action for this session

Security tip: Start with individual approvals (yes) until you understand what Claude is doing!

Essential Commands¶

Claude Code has special commands starting with /:

Command	Purpose
`/help`	Show available commands
`/clear`	Clear conversation history
`/exit`	Exit Claude Code
`/reset`	Reset the session

Common Workflow Patterns¶

Below are some examples of how Claude Code works in different biocuration scenarios. These patterns demonstrate Claude Code's capabilities working with local files, databases, and APIs. Don't worry if some of the terminology seems unfamiliar—you'll learn about LinkML schemas, database queries, and API interactions in detail in the later parts of this tutorial. For now, focus on understanding the general workflow patterns and how Claude Code can assist at each step.

Pattern 1: Explore → Understand → Modify¶

1. "What LinkML files are in this directory?"
2. "Read the CellLine class definition"
3. "Add a new slot called 'tissue_origin' with type string"

Pattern 2: Search → Analyze → Report¶

1. "Find all files that reference BRCA1"
2. "Summarize how BRCA1 is used across these files"
3. "Create a markdown report of your findings"

Pattern 3: Query → Transform → Export¶

1. "Query the database for all human genes"
2. "Convert the results to TSV format"
3. "Save it as human_genes.tsv"

Best Practices¶

✅ DO:¶

Be specific - "Read lines 10-50 of schema.yaml" vs "Read the file"
Iterate - Build complexity step-by-step
Verify - Ask Claude to show what changed
Use natural language - No special syntax required

❌ DON'T:¶

Blindly approve all permissions without reading
Give vague instructions like "fix it"
Assume Claude remembers files from previous sessions
Skip verifying changes before committing

Troubleshooting¶

"Claude can't find the file"¶

Check your current directory:

What directory am I in? List all files here.

"Permission denied"¶

Grant the necessary permissions when Claude asks. Claude explains what it needs and why.

"Unexpected behavior"¶

Ask Claude to explain:

Can you explain what you just did step-by-step?

Key Advantages for Biocuration¶

Direct Database Access - Query production databases conversationally
LinkML Integration - Read, understand, and modify schema files
Batch Operations - Process multiple files or records at once
Git Integration - Automatic git status, commit assistance
Script Generation - Create data validation or transformation scripts
API Integration - Query external APIs (UniProt, NCBI, etc.)

Next Steps¶

Now that you understand Claude Code basics, explore more advanced features:

Exploring LinkML Models with Claude Code - Learn to navigate and extend AGR LinkML schemas
Creating Slash Commands for API Lookups - Build reusable workflows for ChEBI, GO, and other API lookups
Using MCPs for Ontology Lookups - Connect to EBI's Ontology Lookup Service for integrated ontology queries

Resources¶

Claude Code Documentation: https://docs.claude.com/en/docs/claude-code
GitHub Repository: https://github.com/anthropics/claude-code
Installation Guide: https://docs.claude.com/en/docs/claude-code/installation
MCP Servers: https://github.com/modelcontextprotocol/servers

Tip: Claude Code shines for biocuration tasks that involve repetitive file operations, data validation, schema exploration, and database queries. Start with simple tasks and build up to complex workflows!

Note on Content Generation: This tutorial was developed with the assistance of Claude Code (Anthropic's AI-powered CLI tool) for content generation, organization, and technical accuracy verification. All examples and workflows have been tested and validated for educational purposes.