Claude Code Templates

A template repository containing Claude Code command definitions, specialized AI agents, and structured development workflows for enhanced AI-powered software development.

Overview

This repository provides reusable Claude Code templates that can be copied into your projects to enable:

• Specialized AI Agents - Role-based agents for product management, engineering leadership, research, UX design, and administrative tasks
• Development Workflows - Product engineering with Linear integration and PR feedback via GitHub CLI
• Collaborative Development - Multi-agent collaboration for complex feature planning and implementation
• Accessibility Auditing - WCAG 2.1 AA compliance checking with detailed remediation guidance
• Security Auditing - Comprehensive vulnerability assessment and threat analysis
• Analytics Auditing - Pendo tracking implementation verification
• AI Code Review - Automated GitHub Actions workflow for pull request reviews

Quick Start

1. Copy selected templates to your project:

   # Copy desired command templates
   cp dot-claude/commands/feature.md /path/to/your/project/.claude/commands/
   cp dot-claude/commands/fix-pr.md /path/to/your/project/.claude/commands/
   
   # Copy specialized agents
   cp dot-claude/agents/product-feature-planner.md /path/to/your/project/.claude/agents/
   cp dot-claude/agents/engineering-team-lead.md /path/to/your/project/.claude/agents/
   cp dot-claude/agents/research-analyst.md /path/to/your/project/.claude/agents/
   cp dot-claude/agents/intern.md /path/to/your/project/.claude/agents/
   
   # Copy audit commands as needed
   cp dot-claude/commands/security-audit.md /path/to/your/project/.claude/commands/
   cp dot-claude/commands/accessibility-audit.md /path/to/your/project/.claude/commands/
   cp dot-claude/commands/pendo-audit.md /path/to/your/project/.claude/commands/
   
   # Copy GitHub Actions workflow (optional)
   cp -r dot-github/ /path/to/your/project/.github/

2. Customize commands for your project:

   • Update repository-specific references in command files
   • Adjust Linear project references in feature.md
   • Configure analytics patterns in pendo-audit.md

3. Setup AI Code Review (optional):

   • Add ANTHROPIC_API_KEY secret to your GitHub repository
   • The workflow automatically reviews PRs under 2000 lines
   • Comment /claude-review to manually trigger reviews

4. Use commands and agents in Claude Code:

   Commands:

   • /feature [ticket-id] - Product engineering with Linear integration
   • /fix-pr [pr-number] - Review and address PR feedback
   • /security-audit - Run comprehensive security analysis
   • /accessibility-audit - Check WCAG compliance
   • /pendo-audit - Verify analytics tracking

   Agents (via Task tool):

   • Product planning and Linear ticket creation
   • Engineering feasibility and technical design
   • Research and best practices exploration
   • Quick administrative tasks

Available Agents

🧠 Product Feature Planner (product-feature-planner.md)

Expert product management agent for comprehensive feature planning:

• Feature Specification: Creates detailed specs with user stories and acceptance criteria
• Linear Integration: Structures hierarchical tickets (Epics → Stories → Tasks) in Linear
• Prioritization: Applies RICE scoring and identifies trade-offs
• User-Centric: Maps user journeys and defines success metrics
• Stakeholder Communication: Translates technical requirements to business value

👷 Engineering Team Lead (engineering-team-lead.md)

Technical leadership agent for engineering perspective and feasibility:

• Technical Assessment: Evaluates feasibility, scalability, and performance implications
• Architecture Decisions: Recommends design patterns and implementation strategies
• Requirement Refinement: Transforms product specs into technical specifications
• Risk Identification: Flags technical debt, bottlenecks, and integration challenges
• Estimation: Provides realistic effort estimates based on complexity

🔍 Research Analyst (research-analyst.md)

Deep research agent for exploring external resources and best practices:

• Comprehensive Research: Investigates technical solutions beyond the codebase
• Documentation Review: Explores third-party APIs and external documentation
• Competitive Analysis: Researches industry approaches and patterns
• Technology Evaluation: Compares different technical solutions
• Knowledge Synthesis: Presents findings in structured, actionable format

📝 Intern (intern.md)

Administrative assistant agent for quick tasks:

• Note-Taking: Creates notes and reminders without disrupting workflow
• Linear Tickets: Quickly creates tickets for future work
• Documentation: Handles simple documentation tasks
• Context-Independent: Works on tasks that don't require deep project knowledge
• Parallel Execution: Handles administrative work while main agent continues

Available Commands

🚀 Product Engineering (feature.md)

Linear-integrated development workflow for implementing features:

• Ticket Management: Pulls Linear tickets, assigns to user, updates status
• Planning & Documentation: Attaches implementation plans and summaries to Linear
• Devil's Advocate Approach: Considers edge cases and alternative approaches
• Project Context: Reviews parent projects for full context

Requirements: Linear MCP server configuration for ticket management.

🔧 Pull Request Feedback (fix-pr.md)

Structured PR review and feedback resolution workflow:

• GitHub CLI Integration: Direct interaction with pull requests
• Linear Context: Automatically reviews associated tickets (531-XXX format)
• Feedback Triage: Categorizes feedback into immediate fixes vs. future tickets
• Performance & DX Focus: Evaluates edge cases and developer experience

🔐 Security Audit (security-audit.md)

Comprehensive security vulnerability assessment:

• Secret Scanning: Identifies exposed credentials and API keys
• Dependency Analysis: Checks for vulnerable or suspicious packages
• Injection Prevention: Scans for SQL, XSS, and command injection risks
• AI Security: Detects prompt injection vulnerabilities
• Report Generation: Creates prioritized findings in .claude/tmp/

♿ Accessibility Audit (accessibility-audit.md)

WCAG 2.1 AA compliance verification:

• Semantic HTML Review: Validates proper structure and landmarks
• Keyboard Navigation: Checks focus management and tab order
• Screen Reader Support: Verifies ARIA implementation
• Visual Accessibility: Evaluates color contrast and text sizing
• Detailed Remediation: Provides specific code fixes with testing methods

📊 Analytics Audit (pendo-audit.md)

Pendo tracking implementation verification:

• Coverage Analysis: Identifies missing tracking on interactions
• Naming Conventions: Validates feature:component:action patterns
• Privacy Review: Ensures no sensitive data in tracking
• Git Diff Focus: Audits recent changes for proper implementation

GitHub Actions Workflow

🤖 AI Code Review

Automated code review workflow that provides comprehensive feedback on pull requests using Claude AI.

Features:

• Automatic Triggers: Reviews PRs when opened (under 2000 lines)
• Manual Override: Comment /claude-review to force review large PRs
• Smart Analysis: Security, TypeScript compliance, architecture adherence
• Cost Control: Automatic size limits with manual override options

Setup:

1. Add ANTHROPIC_API_KEY as a repository secret
2. Workflow runs automatically on new PRs
3. Review comments appear directly on pull requests

Review Focus:

• Multi-tenant security and data isolation
• TypeScript strict mode compliance
• API authentication and error handling
• Database query security and performance
• Code organization and best practices

Repository Structure

dot-claude/
├── agents/
│   ├── product-feature-planner.md  # Product management expertise
│   ├── engineering-team-lead.md    # Technical leadership
│   ├── research-analyst.md         # Deep research capabilities
│   └── intern.md                    # Administrative assistance
├── commands/
│   ├── feature.md                   # Product engineering workflow
│   ├── fix-pr.md                    # PR feedback resolution
│   ├── security-audit.md            # Security vulnerability scanning
│   ├── accessibility-audit.md       # WCAG compliance checking
│   └── pendo-audit.md               # Analytics tracking verification
└── (additional configs as needed)

dot-github/
└── workflows/
    └── (AI code review workflow)

Using Agents

Agents are invoked through Claude Code's Task tool to provide specialized expertise and enable collaborative AI development.

Benefits of Agent-Based Development

• Domain Expertise: Each agent brings specialized knowledge and best practices
• Parallel Processing: Multiple agents can work simultaneously on different aspects
• Separation of Concerns: Agents focus on their specific domain without context switching
• Collaborative Problem-Solving: Agents work together like a real development team
• Reduced Context Pollution: Main conversation stays focused while agents handle specific tasks

Agent Invocation

Agents work collaboratively to solve complex problems:

User: "I want to add real-time notifications to our chat feature"

Claude: Uses Task tool to invoke product-feature-planner agent
→ Agent creates detailed feature specification and Linear tickets

Claude: Uses Task tool to invoke engineering-team-lead agent
→ Agent reviews technical feasibility and refines requirements

Claude: Implements the feature based on combined insights

Collaboration Patterns

• Sequential: Agents build on each other's work (e.g., product planning → engineering review)
• Parallel: Multiple agents work simultaneously on different aspects
• Research-Driven: Research agent gathers information before implementation
• Review-Based: Engineering lead reviews product specifications

Customization Guide

Adapting Commands

Each command file can be customized for your project:

1. Update Repository References: Replace example repository names with your actual project
2. Configure Tool Integrations: Update Linear project IDs, GitHub repo paths, etc.
3. Adjust Audit Criteria: Modify security rules, accessibility standards, or analytics patterns
4. Add Project-Specific Instructions: Include your team's conventions and requirements

Adapting Agents

Customize agent behavior for your team:

1. Update Tool Access: Modify available tools based on your integrations
2. Adjust Expertise: Tailor agent knowledge to your domain
3. Configure Workflows: Adapt agent workflows to your processes
4. Set Collaboration Rules: Define how agents should work together

Creating New Commands

To create custom commands:

1. Create a new .md file in .claude/commands/
2. Define the role and expertise at the top
3. Provide clear step-by-step instructions
4. Specify output formats and destinations
5. Include any required tool integrations

Creating New Agents

To create custom agents:

1. Create a new .md file in .claude/agents/
2. Add YAML frontmatter with name, description, tools, model, and color
3. Define the agent's expertise and responsibilities
4. Provide examples of when to use the agent
5. Specify collaboration patterns with other agents

Contributing

This is a template repository. Improvements to command templates and new workflow patterns are welcome via pull requests.