Claude Configuration

This repository contains configuration files and templates for Claude AI assistant interactions, including specialized agents, custom commands, and workflow automation.

Structure

.claude/
├── README.md                   # This file
├── CLAUDE.md                   # Main configuration and rules
├── .gitignore                  # Excludes personal data from version control
├── agents/                     # Specialized AI agents for different domains
├── commands/                   # Custom slash commands for common tasks
└── hooks/                      # Shell scripts for automation

Components

📋 Configuration (CLAUDE.md)

The main configuration file containing:

• Default mode settings (Architect mode enabled)
• Code style guidelines and consistency rules
• Communication preferences and formatting rules
• Commit message conventions
• Knowledge sharing and persistence guidelines

🤖 Agents (agents/)

Specialized AI agents for different domains:

• api-designer: RESTful API design and review
• aviation-domain-expert: Flight computer algorithms and aviation theory
• cargo-specialist: Rust Cargo and dependency management
• cicd-specialist: CI pipeline design and GitHub Actions workflows
• code-reviewer: Comprehensive code quality review
• cpp-expert: Modern C++ patterns and optimization
• cross-platform-ui-specialist: UI rendering and touch interfaces
• css-specialist: CSS architecture and modern layout systems
• database-optimizer: SQL query optimization and schema design
• dependency-analyst: Dependency updates and security analysis
• devops-engineer: Infrastructure and deployment automation
• documentation-writer: Technical documentation and API docs
• ember-specialist: Ember.js applications and Ember Data modeling
• geospatial-mapping-specialist: Map rendering and coordinate systems
• integration-tester: End-to-end testing strategies
• migration-specialist: Safe database migrations and refactoring
• mobile-android-specialist: Android NDK development and optimization
• nmea-protocol-specialist: NMEA sentence parsing and GPS data handling
• performance-engineer: Performance profiling and optimization
• prompt-engineer: Crafting effective prompts for Claude Code interactions
• rust-expert: Idiomatic Rust patterns and best practices
• security-auditor: Security vulnerability assessment
• shell-script-specialist: Bash/shell scripting and automation
• supply-chain-reviewer: Dependency risk assessment and auditing
• svelte-specialist: Svelte component development and SvelteKit applications
• test-automator: Comprehensive testing strategies
• web-accessibility-specialist: WCAG compliance and inclusive design

⚡ Commands (commands/)

Custom slash commands for common workflows:

• 1-commit: Interactive git commit with message suggestions
• 2-commit-fast: Fast git commit using first suggested message
• 3-pull-request: Generate pull request titles and descriptions

🔧 Hooks (hooks/)

Shell scripts for automation:

• format.sh: Code formatting automation
• notify.sh: Desktop notifications for task completion

Usage

Setting Up

• Clone or copy these configuration files to your ~/.claude/ directory
• Customize CLAUDE.md with your project-specific preferences
• Enable any desired hooks in your Claude settings

Using Agents

Agents are automatically suggested by Claude when working on relevant tasks, or you can explicitly request them:

• "Use the rust-expert agent to review this code"
• "I need the database-optimizer for this SQL query"

Using Commands

Commands are available as slash commands in Claude:

• /1-commit - Interactive commit process
• /2-commit-fast - Quick commit with auto-selected message
• /3-pull-request - Generate PR descriptions

Customization

Adding New Agents

Create new .md files in the agents/ directory following this format:

---
name: agent-name
description: When to use this agent
tools: Read, Write, Bash
model: sonnet
---

Agent instructions and expertise areas...

Modifying Commands

Edit the .md files in commands/ to customize the command behavior. All command files are marked as read_only: true to prevent accidental modification during execution.

Personal vs Shared Files

The .gitignore file excludes personal data while allowing shared configuration:

• ✅ Shared: Agent definitions, commands, hooks, documentation
• ❌ Private: Personal settings, conversation history, project data

Contributing

When contributing to shared configurations:

• Ensure no personal information (paths, API keys, etc.) is included
• Test configurations thoroughly before committing
• Document any new agents or commands clearly
• Follow the existing naming and structure conventions