10 Secret Claude Code Updates You Can’t Afford to Miss!

Anthropic has been quietly rolling out a series of game-changing updates for Claude Code, its AI coding assistant for VS Code. These aren’t just minor tweaks; they are powerful new features that can fundamentally transform your workflow, making you a more efficient and capable developer. Many of these updates were only revealed on specific Twitter accounts, meaning you’ve likely missed them. This guide will walk you through 10 of the most impactful secret updates that will make using Claude Code a significantly better experience, whether you’re a seasoned pro or just starting.

“If you use Claude Code, this is going to be the most important video you watch this month. I promise you whether you’re a beginner or a pro, you will be using Claude Code completely differently after these 10 updates.”

1. Resume Previous Sessions

[00:31.848]

One of the most significant quality-of-life improvements is the ability to resume past sessions. If you’re like most developers, you’ve probably lost valuable context after restarting your computer or closing VS Code. Previously, this meant painstakingly re-feeding information to the AI. Now, that’s a thing of the past. By simply running a command in your terminal, you can pull up a list of your recent conversations and jump right back in where you left off, with all the previous context intact.

To access your previous sessions, open your terminal in VS Code and type the following command:

claude --resume

This will display a list of your past sessions for the current project, allowing you to select and restore a conversation instantly, saving you from the tedious task of rebuilding context from scratch.

2. Visualize Your Context Window

[01:47.888]

Have you ever noticed Claude’s performance degrading or “hallucinating” during a long conversation? This often happens when the context window becomes cluttered. To combat this, Anthropic introduced the /context slash command. This incredible tool gives you a detailed visual breakdown of what’s currently occupying your context window, showing the token usage for the system prompt, system tools, past messages, and more.

“People on X complain right now, ‘Oh, Claude’s getting worse.’ No, that’s a skill issue. They’re just not managing their context well.”

By running /context, you can quickly identify what’s taking up the most space. In many cases, it’s the history of past messages. Armed with this knowledge, you can then use another command like /clear to free up tokens and restore Claude’s performance, ensuring you’re always getting the most accurate and efficient responses.

3. Track Your Usage with /stats

[03:15.068]

A common concern for users of any AI tool is tracking usage and knowing when you might be approaching plan limits. The new /stats command provides a comprehensive dashboard right within your terminal. It shows a complete overview of your activity, including which models you’re using most frequently, your total tokens used, your longest session, your daily streak, and even your peak usage hours. This feature not only helps you manage your subscription but also offers fun insights into your coding habits.

4. Keep Organized by Naming Your Sessions

[04:15.688]

As you start working on multiple features or tasks simultaneously, juggling different Claude Code sessions can become confusing. The /rename slash command solves this by allowing you to give each conversation a custom, descriptive name. This is especially powerful when combined with the claude --resume feature, as your named sessions will appear in the list, making it effortless to find and return to the exact conversation you need. To name your current session, simply type:

/rename <your-session-name>

5. Instantly Undo Mistakes with Rewind

[05:37.528]

Experimentation is key to development, but it sometimes leads to mistakes. With the new Rewind feature, you can easily undo any changes you regret. By simply double-tapping the Escape key, you can bring up an interactive timeline of your session. This allows you to select a previous point in the conversation and restore both your code and the conversation state to that exact moment. It’s like having a time machine for your coding session, giving you the freedom to experiment without fear of permanently messing things up.

6. Unleash Deeper Reasoning with Ultrathink

[06:57.170]

For particularly complex problems, bugs, or creative tasks, you can now activate the Ultrathink mode. This isn’t a slash command, but a keyword you include directly in your prompt. By adding the word “ultrathink” to your request, you instruct Claude to dedicate extra computational resources and time to your problem. The result is a more thorough, well-reasoned, and often higher-quality output. While it takes longer and uses more tokens, it’s an invaluable tool for tackling your most challenging coding tasks.

Add a beautiful calendar section to our app using the design skill. ultrathink about this.

7. Create Persistent Custom Memories

[08:34.900]

While Claude Rules are great for setting overarching guidelines, they can become cluttered. The new Custom Memories feature offers a more flexible way to make Claude remember specific instructions. By starting your prompt with a hashtag (#), you can add a piece of information to either your personal “User Memory” (for all projects) or the current “Project Memory.” This ensures Claude consistently follows specific rules, like always using a particular skill or adhering to a certain coding style, without you having to repeat the instruction.

# Always use the design skill when changing the UI

8. Enable “YOLO Mode” for Maximum Speed

[10:13.166]

For advanced users who have built a strong rapport with Claude Code, the constant permission prompts for file changes can slow down the workflow. “YOLO Mode” is the solution. By launching Claude with a special flag, you grant it permission to make changes automatically without asking for confirmation each time. This allows for a truly seamless, autonomous coding experience where you can assign long-running tasks and let Claude work independently. To start a session in this mode, use:

claude --dangerously-skip-permissions

9. Expand Capabilities with the Plugin Store

[11:31.540]

Claude Code now has its own version of an app store, accessible via the /plugins command. This opens up a world of integrations and new skills. You can connect Claude to tools like GitHub, Figma, Linear, and Notion, or install powerful new skills to enhance its capabilities. One of the most highly recommended plugins is frontend-design, which dramatically improves the quality of the UI that Claude generates. Exploring the plugin store is a must to unlock the full potential of the platform.

10. Never Lose a Prompt with Stashing

[12:36.567]

It’s a common scenario: you’re halfway through writing a complex prompt when you realize you need to do something else first. Instead of deleting your work, you can now stash your prompt by pressing Ctrl+S (or Cmd+S on Mac). This temporarily saves your current input, allowing you to run another command. Once that’s done, your stashed prompt will automatically reappear, ready for you to finish and send. This simple but effective feature eliminates the frustration of losing a carefully crafted prompt.