msrd-rest-samples/bash

MSRD Linux Automation Example

Overview

The msrd.sh shell script is an example of how to use the MSRD HTTP API to perform automated Linux job submission. It should only be considered an example of automation, and uses a mix of SDK API calls and regular shell scripting. In the future, the job submission portion of this script will be replaced by API calls that use the MSRD SDK.

This sample can be used as a starting point for custom Linux job submission scripts. In a production scenario, it would be enhanced with additional logging, exhaustive error handling, &c.

Requirements

The following dependencies are likely already installed on your system, but worth noting:

• bash shell
• curl
• scp
• ssh

Less likely to be installed by default are:

• jq (JSON parsing tool)
• sshpass (for easily using ssh/scp with password authentication)

These are all available on most popular Linux distros.

For example, on Debian/Ubuntu, just run:

sudo apt install curl jq sshpass

Script Steps

To run the script against the sample fuzz target, set the $ACCOUNT_ID and $API_TOKEN script variables, then invoke ./msrd.sh or bash msrd.sh. Note: invoking as sh msrd.sh will typically fail, as this script uses features specific to the bash shell.

The script then attempts to:

1. Initializes a new Linux job preparation VM
2. Fetches the address and SSH credentials for the prep VM
3. Upload a fuzz target, seeds, and fuzzing parameters
4. Submits the job for validation and fuzzing

Details

Job identifiers

The script attempts to create a new job using the given target data, and keeps job-specific data in a subfolder alongside msrd.sh. This includes API-fetched login credentials and a job-specific known_hosts file which is used for SSH commands. The script does not attempt to modify the user known_hosts.

Script globals

There are two global variables that must be set when using the script, even with the default fuzz targets:

• ACCOUNT_ID: The GUID of a user account.
• API_TOKEN: An API token for the above user account, generated from the Account Settings > Api Tokens page.

These must be set to properly direct and authenticate API requests.

Job parameters

The script expects a file named JobParams.json to be located alongside it. The object in this file contains fuzzing parameters that would otherwise be obtained from the user interacting with the sfwizard command line tool, after logging into the prep VM. The paths in the file refer to the filesystem on the preparation VM, not the environment where the msrd.sh script is being run.

Fuzz target and seeds

This example script is designed to create a new fuzz job for the SampleFuzzingJob/DemofuzzLinux target. This is reflected in two places:

1. The script globals $TARGET_PATH, $TARGET_NAME, and $TARGET_SEED_PATH
2. The JobParams.json config file

To use this script with your own job, you must update these values both in the script and in the configuration file. Note that the configuration file paths are on the remote preparation VM, and the script paths are on the local system being used to create and submit the job.