## What it does
Three production-tested Claude Code workflow systems from an AI-native startup. Each workflow ships slash commands for on-demand use and GitHub Actions for automated PR checks.
**Code Review Workflow (code-review/):** Inspired by Anthropic's own Claude Code development process. Implements dual-loop architecture where AI agents handle the "blocking and tackling" of code review — syntax, completeness, style guide adherence, and bug detection. Frees the team to focus on strategic thinking and architectural alignment.
**Security Review Workflow (security-review/):** Automated security review based on Anthropic's security-focused approach and OWASP Top 10 standards. Proactively identifies vulnerabilities, exposed secrets, and potential attack vectors. Returns severity-classified findings (critical/high/medium/low) with clear remediation guidance. Includes on-demand slash commands and automated PR security checks via GitHub Actions.
**Design Review Workflow (design-review/):** Automated UI/UX review using Microsoft's open-source Playwright MCP for browser automation. Specialized Claude Code agents ensure UI/UX consistency, accessibility compliance, and adherence to world-class design standards. Catches visual issues before they reach production.
Each workflow is documented with a video tutorial on Patrick Ellis' YouTube channel.
## Install
```bash
git clone https://github.com/OneRedOak/claude-code-workflows
```
Copy the workflow folder(s) you want into your project's .claude/ directory and configure the GitHub Actions as needed. Each workflow directory is self-contained with its own slash commands and CI configuration.
## When to use
When you want AI-driven code, security, or design review wired into both your local Claude Code workflow (slash commands) and your CI/CD pipeline (GitHub Actions) without building the integration from scratch.