Cisco AI Defense - Command Line Interface (CLI)

Welcome to this video presentation where we will explore the automation options available through the command line interface or CLI as part of Cisco AI defense. This session will provide an in depth understanding of how to leverage the CLI for efficient automation and cyber security defenses.00:04Using state of the Art AI technologies, let's begin this journey into enhancing our security posture through the power of automation.00:20In this slide, we are discussing the Cisco AI defense CLI. This command line interface is designed to simplify and accelerate the process of scanning AI models and associated files for security risks. Its primary advantage lies in its developer friendly approach, which integrates seamlessly into local development environments continuous integration pipelines.00:28And controlled evaluation settings. The CLI supports the scanning of various AI model formats without the need for initial downloads, streamlining the workflow integration process. Importantly, it outputs scan results in JSON format, facilitating easy parsing and integration into automated CICD pipelines.00:48For use an AI defense API token and base URL are required. Let's discuss the supported platforms for the AI defense CLI. It is distributed as a Linux binary compatible with systems through methods like the Windows subsystem for Linux.01:06WSL for Windows users. Mac users on Apple silicon may not be able to run the CLI and may need alternative solutions or workarounds such as AWS EC two instances. This versatility ensures that a broad range of users can leverage the CLI according to their platforms.01:23In this segment, we'll cover how to install the Cisco AI defense CLI. Begin by downloading the binary from the specified Cisco software download page. Once downloaded, change the permissions of the binary file to make it executable. This is done using the command provided.01:41After setting up, you can use the command line help feature to explore its functionalities and capabilities. This slide covers the critical aspect of authentication and configuration for the CLI. Successful operation requires an AI defense API token and a base URL pertinent to your environment.01:56From a security standpoint, the protocol ensures that model files are transmitted securely to Cisco AI defense for analysis. Note that NO credentials are embedded within the binary itself, and authentication is managed through tenant scoped API tokens. This setup underpins the integrity and confidentiality of data handled during the scanning process.02:15The authentication process with a Cisco AI defense CLI involves generating and configuring an API key. You have three methods to provide credentials using environment variables, which is recommended for CICD workflows, creating a configuration file such as.config.TOML or directly via command line arguments. Each method offers.02:35Flexibility depending on your operational environment and security practices. Remember to use your specific regions based URL during these configurations.02:55Here, we detail what the Cisco AI defense CLI accomplishes. This tool is designed to perform local validation of model files prior to secure uploads for cloud based security analysis. Once the analysis is complete, the results are presented in JSON format. This format is optimal for integration into various automated processing and reporting systems.03:05Allowing teams to easily manage and respond to potential security risks highlighted during the scans. Now let's explore the model scan command. This command manages AI or ML model security scans offering several capabilities, running a local scan, triggering a cloud based scan, listing previous scans, and retrieving detailed information.03:25Information about specific scans. The examples demonstrate how you can execute various types of scans to suit your security assessment needs, utilizing the flexibility of the CLI for robust and effective scanning.03:45The MCP scan command focuses on managing model context protocol server security scans. This section introduces various commands available for MCP servers such as registration, deletion, running scans, and pulling detailed scan reports using the command.03:58You can register and monitor MCP servers, trigger on demand scans, and manage server data securely and efficiently. These capabilities support continuous oversight of server security infrastructure. With the AI bomb command, you can manage the AI bill of materials for AI and ML applications.04:13This functionality allows you to generate, submit, and manage AI bomb reports, which are crucial for maintaining an overview of all components within an AI ML application. This command suite supports component listing and provides a summary of bomb statistics enabling comprehensive management of AI ML projects.04:32Thank you for taking the time to engage with our presentation. We hope it provided valuable insights into using the Cisco AI defense CLI to enhance AI model security. Your attention and participation are greatly appreciated and we look forward to addressing any further questions you might have. Please don't hesitate to reach out for more information or support. Thank you once again for being part of this session.04:52