Master slash commands, keyboard shortcuts, and Claude Projects. Everything you need to know about Claude.ai's power features—from basic commands to advanced workflows.
Claude's command system is one of the least understood but most powerful features in Claude.ai. After months of using it daily, I've learned that most people only use a fraction of what's available—and they're missing out on massive productivity gains.
This guide is the reference I wish I'd had when I started. It covers everything: slash commands, keyboard shortcuts, how Projects work as your command center, and how to chain commands into powerful workflows.
Slash commands are shortcuts that trigger specific Claude behaviors. They start with / and handle everything from code review to task organization. They're context-aware—Claude understands what you're asking for and adjusts its response style accordingly.
Think of them as preset instructions that save you from repeating yourself. Instead of: "Please review this code for security vulnerabilities, performance issues, and style problems," you type /review and Claude knows exactly what you need.
| Command | What It Does | Best For |
|---|---|---|
/help |
Opens help documentation | Learning about Claude features |
/clear |
Clears conversation history | Starting fresh without context |
/code |
Specializes in code generation | Writing functions, scripts, algorithms |
/research |
Gathers information and summarizes | Understanding complex topics |
/summarize |
Condenses long text | Extracting key points from articles |
/explain |
Breaks down difficult concepts | Learning new technologies |
/review |
Critiques code or content | Quality assurance and feedback |
Projects are like folders for related conversations. Instead of drowning in a flat list of discussions, you group related work together—and Claude maintains context across all conversations within that project.
I use Projects for major initiatives. Right now I have three active: a design system overhaul, an analytics implementation, and a content strategy. Each project keeps related conversations organized, and I can reference earlier decisions instantly.
Project workflow: Create a project → add descriptions and context → start conversations within it. Claude remembers everything from previous conversations in that project, so it never loses context about your goals or decisions.
These shortcuts compound throughout your day. Cmd+Enter alone saves you hundreds of mouse clicks per week.
Your custom instructions are system-level prompts. They set the baseline behavior for every conversation. I use mine to define:
This single setup saves me from repeating preferences thousands of times. Every response arrives exactly how I want it.
The real power emerges when you combine commands. Here's how I use them sequentially:
/research to gather information/summarize to extract key points/explain to understand implications/code to implement solutions/review to catch issuesEach step builds on the previous one. Claude maintains full context, so later steps reference earlier decisions. This sequential approach has cut my iteration cycles in half.
This is muscle memory now. One keystroke instead of reaching for the mouse. Over a full workday, it saves 2-3 minutes. Sounds small—but it eliminates friction from the workflow that matters most.
I use this before writing any function or implementing logic. It switches Claude's mode to code-first thinking. Responses are more structured, examples are immediate, and explanations assume technical context. It's my most-used command.
I use this when starting a fundamentally new task. It resets the conversation so Claude doesn't carry assumptions from previous work. Essential for keeping contexts clean and preventing confusion.
Opens everything. Need to access settings? Search for a command? Cmd+K. It's faster than remembering where things are in the UI. This single shortcut has probably saved me more time than everything else combined.
I pin my three active projects at the top. Switching between them takes one click. Without Projects, I'd spend 20+ minutes per day hunting through conversation history. With them, I'm instantly back in context.
Combine commands with custom instructions. Your system prompt defines defaults, slash commands handle specific tasks. Use them together: system prompt says "always markdown," /code specializes that for programming work.
Create conversation templates. For repetitive workflows, I create a template conversation with the exact sequence of steps I always follow. I reference it when starting similar work, saving setup time.
Build project hierarchies. Instead of flat Projects, I organize by domain: Design (with sub-projects for components, patterns), Development (with sub-projects for frontend, backend), Content (with sub-projects for blog, docs).
Want to go deeper? These resources have shaped my understanding of Claude:
docs.anthropic.com for the authoritative reference on features, API usage, and best practices. I reference it constantly for feature details I might have missed.Start small. Pick one command from this list and use it intentionally for a week. Let it become habit. Then add another. The compound effect of small optimizations builds into massive productivity gains.
Your future self—the one working faster and never losing context—will thank you.
Comments
Learn AI productivity strategies and Claude workflows delivered to your inbox. No spam, ever.