Host Requirements

The local host must have Ansible installed and support SSH.

Server-setup

This section will cover the server setup which must be done before proceeding to other parts.

Server Requirements

The OS of the server must be Debian 12. The server must support SSH. Assumption for the following sections is that you have a working SSH-connection setup to the server.

Populate Playbook

Navigate to ansible/inventory/hosts and fill the IP-address with that of the server.

Create Forks

You need to create forks of the [[][targets]] and schemas. This is to support mutation testing. Despite whether you use mutation testing or not, this is mandatory.

Schemas

There is only one repository for the schemas.

• https://github.com/suoja/API-schemas

Targets

There are four targets with two repositories. You need to fork both of them repositories.

• https://github.com/WebFuzzing/Dataset
• https://github.com/SeUniVr/api_case_studies

Populate .env-file

Navigate to .env and fill API_REPO_URL and FORK_OWNER.

Run Ansible

You can now run Ansible to setup the environment to the server with the command ansible-playbook -i ansible/inventory/hosts ansible/playbook.yaml. Note that due to the path, the working directory must be at the root of the project or modify the path. This will take a moment.

Using the environment

Now that environment has been moved to the server, you can SSH to the server: ssh root@IP-ADDRESS

Initialize

First of all you should initialize the testing environment. That is simply done by running init. There is a minor bug that might state that a env_cache_file.env doesn’t exist. This can be ignored.

The environment will ask you for the git-branch of the target. Without mutation testing this can be left empty which will default to master. Note that with mutation testing you need to create the same branch for all the target-repositories.

Following is a question of API branch. This means the git-branch of the schema-files. This defaults to main and only has to be changed in case of API-schema mutation testing.

Now the fuzzer will ensure a clean testing environment which will take a couple of minutes. You might bump into a Docker-bug about coroutines. In this case simply rerun the init command as long as there is need.

When the initialization is ready, the environment shows the running targets, fuzzers and API schemas.

Running Fuzzers

There are two options; either you run one fuzzer or you run all fuzzers. Both of the options will run testing against all fuzzers.

• test_single will present you the available fuzzers from which you choose the desired fuzzer
• test_all will start fuzzing with all fuzzers one by one

Collecting Outputs

Once the testing have finished, you can find the results from $HOME/outputs. You probably want to create an archive of it and then move it to localhost for further inspection via scp or similar tool.