Claude Engineering Skills

A collection of Agent Skills for Claude that provide best-practice engineering guidance for code review, generation, and infrastructure management.

What are Agent Skills?

Agent Skills are modular capabilities that extend Claude's functionality with domain-specific expertise. Each skill packages instructions, workflows, and best practices that Claude uses automatically when relevant to your request.

For more information, see Anthropic's Agent Skills documentation.

Available Skills

Skill	Description
core-engineering	Core coding standards, code review patterns, and generation guidelines
agent-workflow	Plan/Implement/Review workflow with audit requirements
python-development	Python 3.14+ standards, t-strings, free-threading, type hints, async patterns, testing
typescript-javascript	TypeScript/JavaScript patterns, strict typing, ESM
go-rust-systems	Go and Rust systems programming patterns
bash-shell-scripting	Bash scripting, CLI design, Makefile patterns
scripting-automation	Advanced Bash automation, retry logic, testing with BATS
containers-orchestration	Docker best practices, multi-stage builds, security scanning
kubernetes-containers	Kubernetes and Helm patterns
infrastructure-iac	Terraform, Ansible, Docker, CloudFormation
cloud-platforms	AWS, Azure, GCP, Cloudflare best practices
cicd-github-actions	GitHub Actions workflows and CI/CD
security-testing	OWASP Top 10, testing strategies, API design
documentation-standards	Markdown, Mermaid diagrams, technical writing
mcp-development	MCP server patterns and AI/ML integration
database-postgresql	PostgreSQL schema design, migrations, performance, security

Installation

Claude Code

Clone this repository and Claude Code will automatically discover the skills:

# Clone to your home directory
git clone https://github.com/d-padmanabhan/claude-engineering-skills.git ~/.claude/skills/claude-engineering-skills

# Or clone to a project directory
git clone https://github.com/d-padmanabhan/claude-engineering-skills.git .claude/skills/claude-engineering-skills

Skills in ~/.claude/skills/ are available globally, while skills in .claude/skills/ are project-specific.

Development Setup (Recommended)

If you want to work on the skills as a development project while keeping them available to Claude Code:

# Clone to your code directory
git clone https://github.com/d-padmanabhan/claude-engineering-skills.git ~/code/claude-engineering-skills

# Create symlink for Claude Code to discover
ln -s ~/code/claude-engineering-skills ~/.claude/skills/claude-engineering-skills

This approach keeps the repository in your organized code structure while making skills globally available. Any changes you make are immediately available to Claude Code.

Claude.ai

1. Go to Settings > Features
2. For each skill you want to use:
   • Zip the skill folder (e.g., python-development/)
   • Upload the zip file

Note: Claude.ai skills are individual to each user.

Claude API

Upload skills using the Skills API:

# Package the skill
zip -r python-development.skill python-development/

# Upload via API (requires skills-2025-10-02 beta header)
curl -X POST https://api.anthropic.com/v1/skills \
  -H "x-api-key: $ANTHROPIC_API_KEY" \
  -H "anthropic-beta: skills-2025-10-02" \
  -F "file=@python-development.skill"

Skill Structure

Each skill follows the Anthropic Agent Skills specification:

skill-name/
├── SKILL.md           # Main instructions with YAML frontmatter
└── references/        # Detailed documentation (loaded on-demand)
    ├── topic-1.md
    └── topic-2.md

Progressive Disclosure

Skills use a three-level loading system:

1. Metadata (~100 tokens) - Always loaded, used for triggering
2. SKILL.md body (<5k tokens) - Loaded when skill triggers
3. References (unlimited) - Loaded only when needed

This ensures efficient context usage while providing comprehensive guidance.

Usage Examples

Once installed, Claude automatically uses relevant skills based on your requests:

• "Review this Python code for security issues" → Triggers python-development and security-testing
• "Show me how to use Python 3.14 template strings" → Triggers python-development
• "Help me set up a Terraform module for AWS" → Triggers infrastructure-iac and cloud-platforms
• "Create a GitHub Actions workflow" → Triggers cicd-github-actions
• "Write a production-grade Bash script with retry logic" → Triggers scripting-automation

Contributing

Contributions are welcome! Please follow these guidelines:

1. Each skill should have a clear, focused purpose
2. Keep SKILL.md under 500 lines
3. Move detailed content to references/
4. Include clear trigger conditions in the description
5. Test with Claude before submitting

License

MIT License - See LICENSE for details.

Acknowledgments

These skills are derived from the cursor-engineering-rules project, adapted for the Anthropic Agent Skills format.