NemoClaw Knowledge Wiki

New Claude Code features

5 min read

New Claude Code features

https://www.youtube.com/watch?v=o_qZa4ifDdQ This video provides a detailed overview of several new features and improvements in Claude Code, covering updates from versions 1.0.71 down to 1.0.57. Here’s a breakdown of the features:

  1. Customizable Status Lines (1.0.71): What it is: Allows users to personalize the status line at the bottom of the Claude Code terminal. How it works: Users can issue commands like /statusline add the weather in London with emojis or /statusline add how long I've been in the session, the model being used, the git branch, the folder name alongside the weather, weather should be last. Claude will then generate and modify a Bash script (statusline-command.sh in ~/.claude/settings.json) that runs in the background to fetch and display this information. Usefulness: Provides real-time context and personalized metrics (like session duration, current model, Git branch, local folder, or even data fetched from an API endpoint, e.g., unread emails, Stripe earnings) without interrupting the workflow.

  2. Background Commands (1.0.71): What it is: The ability to run Bash commands in the background, allowing Claude to continue processing other requests while monitoring the background task’s output. How it works: Prepend Ctrl-B to any Bash command to run it in the background. Users can then use /bashes to list and manage these running background shells, view their output, or kill them. Usefulness: Ideal for tasks like running development servers locally (e.g., npm run dev), continuously monitoring Docker container logs (docker logs -f), or tailing log files. Claude can keep track of error logs or server status, providing proactive assistance.

  3. Added Support for @-mentions in Slash Command Arguments (1.0.70): What it is: Enables specifying specific files or folders as arguments to custom slash commands using the @ symbol. How it works: When using a custom slash command (e.g., /optimize), typing @ will bring up suggestions for files and folders. Selecting one will pass its path as an argument to the command. Usefulness: Allows for more precise targeting of operations. Instead of analyzing an entire project, users can direct Claude to focus on a specific component or directory for tasks like optimization or security review.

  4. Upgraded Opus to Version 4.1 (1.0.69): What it is: An upgrade to the underlying Opus language model. Usefulness: Implies general improvements in reasoning, code generation, and understanding for tasks where Opus is used.

  5. Agents: Added Model Customization Support (1.0.64): What it is: Users can now specify which Claude model (Sonnet, Opus, or Haiku) a custom agent should use. How it works: When creating a new agent, a prompt appears to select the desired model. This setting is then saved in the agent’s definition file (e.g., i18n-sync-manager.md will include model: haiku). Usefulness: Allows for optimizing the agent’s performance and cost. Simpler, faster tasks can use Haiku, while more complex reasoning can leverage Opus, avoiding overkill and saving tokens.

  6. Added Hidden Files to File Search and @-mention Suggestions (1.0.64): What it is: Hidden files and folders (e.g., .DS_Store, .claude/, .env, .git/) are now included in the auto-completion suggestions when using @-mentions. Usefulness: Improves discoverability and allows Claude to interact with important configuration or version control files that might otherwise be overlooked.

  7. Transcript Mode (Ctrl+R): Changed Esc to Exit Transcript Mode Rather Than Interrupt (1.0.61): What it is: A workflow improvement for transcript mode. How it works: Pressing Ctrl+R reveals Claude’s internal thought process and the exact context it’s operating on. Previously, Esc would interrupt the current agent’s task. Now, Esc simply exits the transcript view, allowing the agent to continue its work uninterrupted. Usefulness: Facilitates debugging and understanding Claude’s decision-making without disrupting ongoing operations.

  8. IDE: Added Support for Pasting Images in VSCode macOS using Cmd+V (1.0.61): What it is: Allows users to paste images directly into the Claude Code terminal on macOS. How it works: After taking a screenshot (e.g., using Cmd+Shift+4 to copy to clipboard), simply press Cmd+V in the Claude Code input to paste the image. Usefulness: Streamlines the process of providing visual context to Claude, such as UI mockups, error screenshots, or diagrams, making interaction more fluid.

  9. Added Support for Reading PDFs (1.0.58): What it is: Claude Code can now read and process PDF files. How it works: Users can drag and drop a PDF file into the Claude Code terminal, then provide a prompt (e.g., summarize this). Claude will read the PDF content and respond. Usefulness: Highly beneficial for developers who receive feature specifications, technical documentation, or research papers in PDF format, enabling quick summaries and information extraction.

  10. Added Support for Specifying a Model in Slash Commands (1.0.57): What it is: Allows developers to hardcode a specific Claude model to be used for a given custom slash command. How it works: By adding a model: field in the frontmatter of a slash command’s markdown definition (e.g., model: opus), that command will always use the specified model, regardless of the default conversation model. Usefulness: Provides granular control over resource usage and performance for specific automated tasks, ensuring the right model is used for the right job without wasting tokens on simpler commands or underperforming on complex ones.

Graph View