Getting Started¶

Get productive with Claude Code Tresor in 5 minutes

This guide gets you from zero to productive quickly. Choose your path based on what you need most.

⚡ Quick Start (2 minutes)¶

Installation¶

# Clone repository
git clone https://github.com/YOUR_USERNAME/claude-code-tresor.git
cd claude-code-tresor

# Install everything (recommended for first time)
./scripts/install.sh

# Or selective installation:
./scripts/install.sh --skills-only      # Skills only
./scripts/install.sh --agents-only      # Sub-agents only
./scripts/install.sh --commands-only    # Commands only

Verify Installation¶

# Check installed components
ls ~/.claude/skills/          # 8 skills
ls ~/.claude/agents/          # 8 sub-agents
ls ~/.claude/commands/        # 4 commands

# Test a skill (skills work automatically when coding)
# Test a sub-agent
claude "@code-reviewer Analyze this file"

# Test a command
claude "/scaffold react-component Button"

✅ You're ready! Skills are available and Claude will invoke them during conversations when relevant.

🎯 Choose Your Path¶

Path 1: "I want proactive code quality suggestions"¶

What you get: Skills that Claude invokes during code discussions

# 1. Install skills
./scripts/install.sh --skills

# 2. Start Claude Code and discuss your code
# Skills are invoked by Claude when contextually relevant:
# - code-reviewer: Suggests improvements during code discussions
# - security-auditor: Flags vulnerabilities when reviewing security
# - test-generator: Suggests tests when discussing new functions

# Example: Discuss code with Claude
claude
> "I'm working on src/utils/api.ts - any quality issues?"

# Claude invokes relevant skills:
# ✅ code-reviewer checks patterns
# ✅ security-auditor scans for issues
# ✅ test-generator suggests missing tests

# 3. For deep analysis, invoke sub-agent
claude "@code-reviewer --focus security"

Time investment: 2 minutes setup, proactive assistance during conversations

Path 2: "I need comprehensive code review"¶

What you get: Multi-perspective code analysis

# 1. Install sub-agents and commands
./scripts/install.sh --agents-only --commands

# 2. Stage your changes
git add .

# 3. Run comprehensive review
claude "/review --scope staged --checks all"

# This coordinates:
# - @code-reviewer: Code quality and patterns
# - @security-auditor: Vulnerability scanning
# - @performance-tuner: Performance analysis
# - @architect: Architecture validation

# Result: Detailed report with actionable recommendations

Time investment: 3 minutes setup, 5-10 minutes per review

Path 3: "I want to scaffold projects faster"¶

What you get: Pre-configured templates and scaffolding

# 1. Install commands
./scripts/install.sh --commands

# 2. Scaffold components/projects
claude "/scaffold react-component UserProfile --hooks --tests"
claude "/scaffold express-api user-service --auth --database"
claude "/scaffold nextjs-app my-app --typescript --tailwind"

# 3. Generated structure:
# - Component files with boilerplate
# - Test files with basic structure
# - TypeScript types
# - Documentation stubs

# 4. Customize generated code to your needs

Time investment: 2 minutes setup, saves 15-30 minutes per scaffold

Path 4: "I need automated testing workflows"¶

What you get: Test generation and execution automation

# 1. Install test-focused tools
./scripts/install.sh --skills-only --agents --commands

# 2. Start conversation about testing
# Claude invokes test-generator skill when relevant

# 3. Generate comprehensive test suites
claude "/test-gen --file src/utils/api.ts --framework jest --coverage 90"

# 4. For complex testing needs
claude "@test-engineer Create comprehensive test suite with edge cases"

# Result:
# - Unit tests for all functions
# - Integration tests for workflows
# - Edge cases and error handling
# - 90%+ code coverage

Time investment: 3 minutes setup, saves hours per feature

Path 5: "I want automatic documentation"¶

What you get: Self-updating docs from code

# 1. Install documentation tools
./scripts/install.sh --skills-only --agents

# 2. Discuss documentation with Claude
# Skills are invoked when discussing docs:
# - api-documenter: Suggests OpenAPI specs for APIs
# - readme-updater: Suggests README updates

# Example workflow:
claude
> "I added this API endpoint - help me document it"
function createUser(req, res) { /* code */ }

# Claude invokes api-documenter skill which suggests:
# ✅ OpenAPI spec structure
# ✅ Example requests/responses
# ✅ Documentation updates

# 3. For comprehensive docs
claude "/docs-gen api --format openapi"
claude "@docs-writer Create user guide with examples"

# Result: Well-documented APIs

Time investment: 2 minutes setup, proactive documentation suggestions

🚀 Real-World Scenarios¶

Scenario: New Feature Development¶

Task: Add user authentication feature

# 1. Start Claude Code session
claude

# 2. Discuss your work
> "I'm adding user authentication - can you review as I go?"

# Skills are invoked by Claude during the conversation:
# - code-reviewer checks patterns in real-time
# - security-auditor validates auth implementation
# - test-generator suggests test cases

# 3. Run formal review when ready
git add src/auth/
claude "/review --scope staged"

# 4. Generate comprehensive tests
claude "/test-gen --file src/auth/login.ts --framework jest"

# 5. Update documentation
claude "/docs-gen --update-readme"

# 6. Commit
git commit -m "feat(auth): add login with OAuth2"

# Total time saved: 2-3 hours

Scenario: Bug Fix¶

Task: Fix production bug

# 1. Analyze issue
claude "@debugger Analyze this error: [stack trace]"

# 2. Discuss fix with Claude
claude
> "I'm fixing this issue - can you validate my approach?"
# (code-reviewer skill is invoked during discussion)

# 3. Add regression test
claude "@test-engineer Create test to prevent this regression"

# 4. Security check
claude "@security-auditor Validate fix doesn't introduce vulnerabilities"

# 5. Deploy with confidence
git commit -m "fix(api): resolve memory leak in WebSocket"

# Total time saved: 1-2 hours

Scenario: Code Review for PR¶

Task: Review teammate's pull request

# 1. Fetch PR
gh pr checkout 123

# 2. Comprehensive review
claude "/review --scope pr --checks all"

# Review covers:
# ✅ Code quality (patterns, smells)
# ✅ Security (vulnerabilities, secrets)
# ✅ Performance (bottlenecks, optimizations)
# ✅ Architecture (design patterns, coupling)
# ✅ Tests (coverage, edge cases)

# 3. Provide feedback
# (generates detailed report with line numbers)

# Total time saved: 30-45 minutes per PR

📚 Understanding the 3-Tier Architecture¶

Skills (Tier 1: Contextual Helpers)¶

What: Helpers invoked by Claude during conversations
When: Claude activates them when contextually relevant
Examples: code-reviewer, test-generator, security-auditor

Think of skills like: Smart autocomplete - activates when relevant to your current discussion

Sub-Agents (Tier 2: Manual Experts)¶

What: Invoked specialists for deep analysis
When: You explicitly call them (@agent)
Examples: @code-reviewer, @test-engineer, @docs-writer

Think of sub-agents like: Consulting an expert - you ask when you need deep help

Commands (Tier 3: Workflows)¶

What: Multi-agent orchestrated workflows
When: You run a command (/command)
Examples: /review, /scaffold, /test-gen

Think of commands like: Running a script - automates complex multi-step processes

🔧 Configuration (Optional)¶

Sandboxing (Advanced Users)¶

Default: Sandboxing is OFF (all skills work without it)

Enable sandboxing for additional security isolation:

# During installation
./scripts/install.sh --skills-only --sandboxing

# Or configure manually per skill
vim ~/.claude/skills/security-auditor/config.json

When to enable: - High-security environments - Untrusted code analysis - Compliance requirements

See: SANDBOXING-GUIDE.md for details

Customization¶

Customize existing skills:

# Copy skill and modify
cp -r ~/.claude/skills/code-reviewer \
      ~/.claude/skills/company-code-reviewer

# Edit for your standards
vim ~/.claude/skills/company-code-reviewer/SKILL.md

Create new skills:

See skills/TEMPLATES.md for copy-paste templates

🎓 Learning Path¶

Day 1: Installation & Basics¶

✅ Install skills (./scripts/install.sh --skills)
✅ Start Claude Code session and discuss your code
✅ Observe Claude invoking skills during conversations
✅ Request deep analysis: @code-reviewer Analyze this

Day 2: Commands¶

✅ Try scaffolding: /scaffold react-component TestComponent
✅ Run code review: /review --scope staged
✅ Generate tests: /test-gen --file utils.js

Week 1: Integration¶

✅ Use skills + sub-agents together during conversations
✅ Integrate /review into PR workflow
✅ Discuss documentation updates with Claude (skills suggest improvements)

Week 2: Customization¶

✅ Create custom skill from template
✅ Adjust skill triggers for your workflow
✅ Share custom skills with team

🆘 Troubleshooting¶

Skill Not Activating¶

Check:

# Verify skill installed
ls ~/.claude/skills/ | grep skill-name

# Check Claude Code recognizes it
claude --list-skills

Fix:

# Reinstall skill
./scripts/install.sh --skills

Command Not Found¶

Check:

ls ~/.claude/commands/ | grep command-name

Fix:

./scripts/install.sh --commands

Sub-Agent Error¶

Common issue: Tool access restrictions

Fix:

# Check agent configuration
cat ~/.claude/agents/agent-name.json

# Verify allowed-tools includes required tools

📖 Next Steps¶

Dive Deeper¶

Architecture: ARCHITECTURE.md - Understand the system design
Skills Guide: skills/README.md - Comprehensive skills documentation
Templates: skills/TEMPLATES.md - Create custom skills
Examples: examples/workflows/ - Real-world workflows

For Existing Users¶

Already using Claude Code Tresor? See MIGRATION-GUIDE.md for upgrading to the new skills system.

Get Help¶

GitHub Issues: Bug reports and feature requests
GitHub Discussions: Questions and community support
Documentation: Browse examples/ for detailed workflows
Professional Support: Available for custom development

💡 Pro Tips¶

Start small - Install skills first, get comfortable, then add sub-agents and commands
Let skills run - They're designed to be non-intrusive, leave them on
Combine tools - Skills detect → Sub-agents analyze → Commands orchestrate
Customize gradually - Use defaults first, customize when you know your needs
Share with team - Skills and agents work great for team standards

🎉 You're Ready!¶

You now have: - ✅ Tools installed and verified - ✅ Understanding of 3-tier architecture - ✅ Real-world scenarios to practice - ✅ Next steps for continued learning

Start coding - skills are watching your back!

Questions? See README.md for full documentation or open a GitHub issue.

Created: October 24, 2025 Author: Alireza Rezvani License: MIT