new·The score now tells you which way it movedA brain's exam only ever grows: its own material writes questions, and so does every question a real caller asked and did not get answered. The score is a percentage over that growing set, so a brain that learned more could post a smaller number — and this week three did. One of them answered two MORE questions than the week before and showed eighteen points less. Printed as a single percentage, that reads as decline to a reader and as punishment to anyone who contributes material.all news →
mozg.beta
Sign in

AI SDK · Cookbook · all subjects

computer-use/overview

8 notes, read out of this brain and free to use. Each one was extracted from a source and is re-checked against its exam.

Computer Use tool actions available

The Computer Tool enables basic computer control actions including: moving the cursor, clicking buttons, typing text, taking screenshots, and reading screen content.

Three main Computer Use tools in API

Computer Use provides three main tools: the Computer Tool (enables basic computer control like mouse movement, clicking, and keyboard input), the Text Editor Tool (provides functionality for viewing and editing text files), and the Bash Tool (allows execution of bash commands).

Computer Use best practices

Best practices for Computer Use include: (1) Specify simple, well-defined tasks with explicit instructions for each step, (2) Prompt Claude to verify outcomes through screenshots, (3) Use keyboard shortcuts when UI elements are difficult to manipulate, (4) Include example screenshots for repeatable tasks, (5) Provide explicit tips in system prompts for known tasks.

Computer Use security measures required

Computer Use poses unique risks and requires security precautions: (1) Use a dedicated virtual machine or container with minimal privileges to prevent system attacks or accidents, (2) Avoid giving the model access to sensitive data like account login information to prevent information theft, (3) Limit internet access to an allowlist of domains to reduce exposure to malicious content, (4) Ask a human to confirm decisions with meaningful real-world consequences such as accepting cookies, executing financial transactions, or agreeing to terms of service.

Computer Use is currently beta with limitations

Computer Use is currently in beta and may be error-prone at times. Anthropic recommends starting with low-risk tasks and implementing appropriate safety measures. The feature has documented limitations that should be reviewed before use.

Computer Use how it works process flow

Computer Use works through four steps: (1) Start with a prompt and tools - add Anthropic-defined Computer Use tools to your request with a task (e.g. 'save an image to your downloads folder'), (2) Select the right tool - the model evaluates which tools can help and sends a formatted tool_call, (3) Execute the action and return results - the AI SDK processes Claude's request by running the selected tool and sends results back to Claude via tool_result message, (4) Complete the task through iterations - Claude analyzes each result to determine if more actions are needed and continues requesting tool use until task completion or additional input is required.

Model version for Computer Use

Computer Use is available with Claude 3.5 Sonnet model. In code examples, use the model identifier 'anthropic/claude-sonnet-4-20250514' or check Anthropic's current model recommendations for Computer Use support.

generateText experimental_sandbox parameter

The experimental_sandbox parameter is optional and accepts Experimental_SandboxSession. It is passed through to prepareStep, tool description functions, and tool execution. Tools can access it from their description function options and execution options.

Give your agent this brain