A collection of skills, agents, and workflows for Claude Code.
YouTube | Newsletter | Blog
# Add the marketplace
/plugin marketplace add dgalarza/claude-code-workflows
# Install what you need
/plugin install tdd-workflow@dgalarza-workflowsSee INSTALL.md for full installation options.
- TDD Workflow - Test-driven development, one test at a time
- Conventional Commits - Structured commit messages
- Parallel Code Review - Multi-agent reviews
- Meeting Transcript - Process transcripts into notes
- Cybersecurity Reviewer - Security analysis agent
- Gridfinity Planner - 3D printing baseplate planning
- Rails Toolkit - Complete Rails development workflow
- Tip 1: Use worktrees for parallel work
- Tip 2: Customize your status bar
- Tip 3: Compact context proactively
- Tip 4: Structure your CLAUDE.md files
- Tip 5: Use subagents for focused tasks
- Tip 6: MCP servers worth installing
Install: /plugin install tdd-workflow@dgalarza-workflows
A skill that enforces true test-driven development:
- Write ONE failing test
- Write minimal code to pass
- Refactor
- Repeat
Claude follows the red-green-refactor cycle step by step instead of writing all tests upfront.
When it activates: Automatically when implementing features with TDD.
Install: /plugin install conventional-commits@dgalarza-workflows
Structured commit messages following the Conventional Commits specification.
feat(api): add webhook signature verification
fix: resolve login redirect loop
docs: update README with setup instructions
Includes reference material for commit types, scopes, and examples.
When it activates: When creating Git commits.
Install: /plugin install parallel-code-review@dgalarza-workflows
Run multiple specialized review agents in parallel for comprehensive feedback. Get security, architecture, and style feedback simultaneously.
When it activates: During code review workflows.
Install: /plugin install meeting-transcript@dgalarza-workflows
Process raw meeting transcripts (from Granola, Otter, or manual notes) into structured notes with:
- Summary
- Action items with owners
- Key decisions
- Follow-up questions
When it activates: When processing meeting transcripts.
Install: /plugin install cybersecurity-reviewer@dgalarza-workflows
Security analysis agent for:
- Vulnerability assessment (OWASP Top 10)
- Threat modeling
- Authentication/authorization review
- Data protection analysis
Provides actionable remediation with code examples.
When to use: Security reviews, penetration test prep, architecture review.
Install: /plugin install gridfinity-planner@dgalarza-workflows
Plan and design gridfinity baseplates for 3D printing:
- Calculate optimal grid sizes from measurements (metric or imperial)
- Slice large grids into printable chunks based on printer bed size
- Calculate padding for non-exact fits
- Generate parameters for gridfinity.perplexinglabs.com
When to use: Designing gridfinity storage systems for drawers, toolboxes, or shelves.
Install: /plugin install rails-toolkit@dgalarza-workflows
Complete Rails development workflow bundle. Includes:
Agents:
- Rails Code Reviewer - Conventions, POODR, idiomatic Ruby
- Rails Security Reviewer - Multi-tenant security, ActsAsTenant
- Rails Feature Developer - TDD-driven implementation
- Rails Backend Expert - Architecture guidance
Commands:
/rails-toolkit:full-code-review- Parallel security + Rails review/rails-toolkit:linear-worktree- Set up worktree from Linear issue
Skills:
- Linear implementation workflow
- RSpec testing patterns
Requirements:
- Linear MCP for Linear integration
- Memory MCP for persistent context
Git worktrees let you work on multiple features simultaneously without branch switching or stashing.
# Create a worktree for a new feature
git worktree add ../myproject-feature-x -b feature-x
# Work in isolation
cd ../myproject-feature-x
# When done, remove it
git worktree remove ../myproject-feature-xWhy it matters: Each worktree has its own file state. No more "let me stash this first" when switching contexts.
The rails-toolkit plugin includes /rails-toolkit:linear-worktree which automates this with Linear issue context.
Claude Code's status bar can show useful context while you work.
claude config set --global statusLineTemplate '${cwd.basename} | ${model} | ${tokenUsage}'This shows:
- Current directory (which project you're in)
- Model being used
- Token usage (so you know when to compact)
See configs/status-bar.md for more options.
Don't wait until Claude starts forgetting things. Compact when:
- You finish a logical unit of work
- You're about to start a different task
- Token usage is getting high
- Claude starts repeating itself or forgetting earlier context
/compactThink of it like saving your game - do it at good checkpoints.
A well-structured CLAUDE.md helps Claude understand your project faster.
# Project Name
## Overview
One paragraph on what this project does.
## Tech Stack
- Framework: Rails 7.2
- Database: PostgreSQL
- Background jobs: Sidekiq
## Key Patterns
- Service objects in app/services/
- Result pattern for service returns
- ActsAsTenant for multi-tenancy
## Testing
- RSpec with FactoryBot
- Run tests: `bin/rspec`
## Common Commands
- Setup: `bin/setup`
- Console: `bin/rails c`
- Server: `bin/dev`The goal: Give Claude the context it needs to make good decisions without overwhelming it.
When Claude spawns subagents, each one gets focused context. Use this to your advantage:
- Security review? Let it spawn a security-focused subagent
- Code review? Multiple specialized reviewers in parallel
- Complex implementation? Let it break down and delegate
Subagents complete their task and return results - they don't pollute your main conversation with intermediate steps.
Start with these:
- Linear - If you use Linear for project management
- Memory - Persistent context across sessions
- Sentry - Debug production errors without context-switching
See configs/mcp-servers.md for setup instructions and configuration snippets.
Found a bug? Have a workflow to share? PRs welcome.
MIT
Built by Damian Galarza - Former CTO, 15+ years in software. I make videos about Claude Code and AI development workflows.