AWS bedrock-agentcore documentation change
Summary
Expanded documentation with detailed workflow steps, security features explanation, session isolation details, recording capabilities, and added use cases section
Security assessment
The changes document security features like session isolation, automatic timeout termination, and secure session recording, but there's no evidence of addressing a specific security vulnerability. This enhances documentation of existing security controls rather than fixing issues.
Diff
diff --git a/bedrock-agentcore/latest/devguide/browser-tool.md b/bedrock-agentcore/latest/devguide/browser-tool.md index 07e0ec2be..40d5a83eb 100644 --- a//bedrock-agentcore/latest/devguide/browser-tool.md +++ b//bedrock-agentcore/latest/devguide/browser-tool.md @@ -5 +5 @@ -OverviewWhy use remote browsers for agent development?Security FeaturesHow it works +OverviewWhy use remote browsers for agent development?Security FeaturesHow it worksUse cases @@ -58,0 +59,2 @@ The Browser Tool includes several security features to help protect your environ +The Amazon Bedrock AgentCore Browser provides a session-based model for secure web browsing with comprehensive observability features. Here's how the complete workflow operates: + @@ -61 +63,38 @@ The Browser Tool includes several security features to help protect your environ -Create a Browser Tool to enable web browsing capabilities. The Browser Tool allows you to augment your agent runtime to securely interact with web applications, fill forms, navigate websites, and extract information. These interactions can be performed in a fully managed environment with low latency. +Create a Browser Tool to enable web browsing capabilities. You can choose between the AWS managed Browser (aws.browser.v1) for quick setup, or create a custom browser with advanced features like session recording, custom network settings, and specific IAM execution roles. The Browser Tool allows you to augment your agent runtime to securely interact with web applications, fill forms, navigate websites, and extract information in a fully managed environment. + + 2. ###### Start a browser session + +The Browser Tool uses a session-based model where each session runs in an isolated environment. After creating a Browser Tool, you start a session with a configurable timeout period (default is 15 minutes, extendable up to 8 hours). Sessions automatically terminate after the timeout period, and multiple sessions can be active simultaneously for a single Browser Tool. + + 3. ###### Interact with the browser + +Once a session is started, you can interact with the browser using WebSocket-based streaming APIs. The Automation endpoint enables your agent to perform browser actions such as navigating to websites, clicking elements, filling out forms, taking screenshots, and more. Libraries like Strands, Nova Act, or Playwright can be used to simplify these interactions. Meanwhile, the Live View endpoint allows an end user to watch the browser session in real time and interact with it directly through the live stream. + + 4. ###### Monitor and record sessions + +All browser sessions provide built-in observability through live viewing capabilities and optional session recording. Live view allows real-time monitoring of browser activity, while session recording (available for custom browsers) captures comprehensive interaction data including DOM changes, user actions, console logs, and network events. Recorded sessions are stored in your Amazon S3 bucket and can be replayed through the AWS Console or accessed programmatically for debugging and analysis. + + 5. ###### Assess performance using observability + +Monitor key metrics for each tool in CloudWatch to get real-time performance insights. Session recordings provide detailed analysis capabilities including video playback, timeline navigation, user action tracking, and comprehensive logs for troubleshooting and optimization. + + + + +## Use cases + +The AgentCore Browser can be used for a wide range of use cases, enabling AI agents to interact with web applications just as humans do. This section describes common use cases. + +With the AgentCore Browser, you can: + + * Test web applications in a secure environment + + * Access online resources and services + + * Perform web-based tasks and workflows + + * Interact with web interfaces + + * Capture screenshots and record browser sessions + + * Build AI agents that can navigate the web @@ -63 +102 @@ Create a Browser Tool to enable web browsing capabilities. The Browser Tool allo - 2. ###### Integrate it within an agent to invoke + * Automate form submissions and data entry @@ -65 +104 @@ Create a Browser Tool to enable web browsing capabilities. The Browser Tool allo -Copy the built-in tool resource ARN into your runtime agent code to invoke it as part of your session. For browser use tools, you can navigate websites and interact with web elements in real-time. + * Extract information from websites @@ -67 +106 @@ Copy the built-in tool resource ARN into your runtime agent code to invoke it as - 3. ###### Assess performance using observability + * Perform e-commerce transactions @@ -69 +108 @@ Copy the built-in tool resource ARN into your runtime agent code to invoke it as -Monitor key metrics for each tool in CloudWatch to get real-time performance insights. + * Monitor website changes and updates