WrongSecrets CTF Party

Powered by MultiJuicer

Want to play OWASP WrongSecrets in a large group in CTF mode, but not go over all the hassle of setting up local copies of OWASP WrongSecrets? Here is OWASP WrongSecrets CTF Party! This is a fork of OWASP MultiJuicer, which is adapted to become a dynamic multi-tenant setup for doing a CTF together!

Note that we:

• have a Webtop integrated for each player
• have a WrongSecrets instance integrated for each player
• A working admin interface which can restart both or delete both (by deleting the full namespace)
• Do not support any progress watchdog as you will have access to it, we therefore disabled it.
• It can cleanup old & unused namespaces automatically.

Special thanks

Special thanks to @commjoen, @madhuakula, @bendehaan, and @mikewoudenberg, and @osamamagdy for making this port a reality!

Sponsorships

We would like to thank the following parties for helping us out:

GitGuardian for their sponsorship which allows us to pay the bills for our cloud-accounts.

Jetbrains for licensing an instance of Intellij IDEA Ultimate edition to the project leads. We could not have been this fast with the development without it!

Docker for granting us their Docker Open Source Sponsored program.

1Password for granting us an open source license to 1Password for the secret detection testbed.

What you need to know

This environment uses a webtop and an instance of wrongsecrets per user. This means that you need per user:

• 1.5 CPU (min = 0.5 , limit = 2.5)
• 2 GB RAM (min 1 GB, limit = 3.5GB)
• 4GB HD (min 3 GB, limit = 8GB)

Running this on minikube

A 4-10 contestant game can be played on a local minikube with updated cpu & memory settings (e.g. 6 virtual CPUs, 9 GB ram).

Running this on AWS EKS with larger groups

Small Game

We recently played a small CTF with 40 relatively active players using version 1.5.10 of wrongSecrets and the T6 version of the virtualdesktop-k8s. This could have easily ran on 5 T3A-X2large nodes for a day.

Large Numbers

A 100 contestant game can be played on the AWS, GCP, and Azure setup, which will require around 150 (100-250) CPUs, 200 (150-350) GB Ram, and 400 GB of storage available in the cluster. Note that we have configured everything based on autoscaling in all cloud providers. This means that you can often start with a cluster about 20% of the size of the "limit" numbers and then see how things evolve. You will hardly hit those limits, unless all players are very actively fuzzing the WrongSecrets app, while runnign heavy appss on their Webtops. Instead, you will see that you are using just 25% of what is provided in numbers here. So, by using our terraform (including an autoscaling managed nodegroup), you can reduce the cost of your CTF by a lot!

Status - Experimental release

This is an experimental release. It showed to work at 6 CTFs already, we just did not complete the documentation and the cleaning up of the Helm chart yet. However: it is working in its basis, and can support a good crowd. Currently, we support using Minikube, AWS EKS, GCP GKE, and Azure AKS (Please follow the readme in the folder for each cloud provider if you want to use it, as the guides section is not updated yet).

How to use it

The different setups are explained in OWASP WrongSecrets CTF-instructions. With the 3-domain approach you generate flags for CTFD automatically, while with the 2-domain setup you need to set it up manually.

Approach 1: 3-domain setup

You need 3 things:

• This infrastructure
• The actual place where correct answers are exchanged for CTFD-flags. This can be your fly.dev/heroku/etc. or local container of WrongSecrets running in CTF mode with the additional key setup for challenge 8.
• A CTFD/Facebook-CTF host which is populated with the challenges based on your secondary hosted WrongSecrets application.

Approach 2: 2-domain setup

You need 2 things:

• This infrastructure
• A CTFD/Facebook-CTF host which is populated with the challenges based on your secondary hosted WrongSecrets application (this can be the helm chart included in the EKS installation script)

To use the 2 domain setup with CTFD:

1. Set up the CTFD and WrongSecrets instances using your preferred method and docs e.g. AWS here, GCP here, or Azure here.
2. Set up a team with spoilers available (On Cloud providers AWS, GCP, or Azure this can be done by changing the deployment of a team you have created and setting ctf-mode=false).
3. Use these spoilers to manually copy the answers from WrongSecrets to CTFD.
4. Delete the team used to get these spoilers ( You can delete the entire namespace of the team)

General Helm usage

This setup works best if you have Calico installed as your CNI, if you want to use the helm directly, without the Cloud Challenges, do:

helm repo add wrongsecrets https://owasp.org/wrongsecrets-ctf-party

helm upgrade --install my-wrongsecrets-ctf-party wrongsecrets/wrongsecrets-ctf-party

Play with Minikube:

** NOTE: The below steps require at least minikube version v1.30.1 and yq (https://github.com/mikefarah/yq/) version v4.34.1. **

For minikube, run:

minikube start  --cpus=6 --memory=10000MB --network-plugin=cni --cni=calico --driver=docker --kubernetes-version=1.30.0
eval $(minikube docker-env)
./build-and-deploy-container.sh
kubectl port-forward service/wrongsecrets-balancer 3000:3000

or use build-and-deploy-container-minikube.sh to do all of the above in one script.

Want to know whether your system is holding up? use

minikube addons enable metrics-server
kubectl top nodes
kubectl top pods

Develop with Minikube

minikube start  --cpus=6 --memory=10000MB --network-plugin=cni --cni=calico --driver=docker --kubernetes-version=1.30.0
eval $(minikube docker-env)
./build-and-deploy.sh
kubectl port-forward service/wrongsecrets-balancer 3000:3000

or use build-and-deploy-minikube.sh to do all of the above in one script.

Play with AWS EKS

NOTE: SEE SECTIONS ABOVE ABOUT WHAT YOU NEED AND THE COST OF THINGS: This project is not responsible, and will not pay for any part of your AWS bill.

For AWS EKS follow the instructions in the /aws folder. This setup also includes a helm installation of CTFd.

Then open a browser and go to localhost:3000 and have fun :D .

Play with GCP GKE

NOTE: SEE SECTIONS ABOVE ABOUT WHAT YOU NEED AND THE COST OF THINGS: This project is not responsible, and will not pay for any part of your GCP bill.

For GCP GKE follow the instructions in the /gcp folder. This setup also includes a helm installation of CTFd.

Then open a browser and go to localhost:3000 and have fun :D .

Play with Azure AKS

NOTE: SEE SECTIONS ABOVE ABOUT WHAT YOU NEED AND THE COST OF THINGS: This project is not responsible, and will not pay for any part of your Azure bill.

For Azure AKS follow the instructions in the /azure folder. This setup also includes a helm installation of CTFd.

Then open a browser and go to localhost:3000 and have fun :D .

Some production notes

See production notes for a checklist of values you'll likely need to configure before using Wrongsecrets-ctf-party in proper events.

Customizing the Setup

You got some options on how to setup the stack, with some option to customize the WrongSecrets and Virtual desktop instances to your own liking. You can find the default config values under: helm/wrongsecrets-ctf-party/values.yaml

The default ctfd config values are here: aws/k8s/ctfd-values.yaml. Note that these values are not used, and instead only se in the file aws/build-and-deploy-aws.sh.

Download & Save the file and tell helm to use your config file over the default by running:

helm repo add wrongsecrets https://owasp.org/wrongsecrets-ctf-party

helm install -f values.yaml my-wrongsecrets-ctf-party wrongsecrets/wrongsecrets-ctf-party

Deinstallation

helm delete my-wrongsecrets-ctf-party

And if you are running AWS, GCP, or Azure (including CTFd):

helm delete ctfd -n ctfd

FAQ

Why a custom LoadBalancer?

There are some special requirements which we didn't find to be easily solved with any pre build load balancer:

• Restricting the number of users for a deployment to only the members of a certain team.
• The load balancers cookie must be save and not easy to spoof to access another instance.
• Handling starting of new instances.

If you have awesome ideas on how to overcome these issues without a custom load balancer, please write us, we'd love to hear from you!

Why a separate kubernetes deployment for every team?

There are some pretty good reasons for this:

• The ability delete the instances of a team separately. Scaling down safely, without removing instances of active teams, is really tricky with a scaled deployment. You can only choose the desired scale not which pods to keep and which to throw away.
• To ensure that pods are still properly associated with teams after a pod gets recreated. This is a non problem with separate deployment and really hard with scaled deployments.
• The ability to embed the team name in the deployment name. This seems like a stupid reason but make debugging SOOO much easier, with just using kubectl.

How to manage WrongSecrets easily using kubectl?

You can list all WrongSecrets with relevant information using the custom-columns feature of kubectl. You'll need to down load the wrongsecrets.txt from the repository first:

kubectl get -l app=wrongsecrets -o custom-columns-file=wrongsecrets.txt deployments

There are a few more ways how you can check whether all is going well: have a look in the /scripts folder for various tools that can help you to see if there are too many namespaces created for instance. This does require you to export the teams and players from ctfd.

Did somebody actually ask any of these questions?

No 😉

The webtop cannot reach the Kubernetes API ! Can you help?

Make sure you make the users connect to the right IP and port number. This you can see for the API server by running the following command on your host where you connect to the cluster with:

kubectl -n kube-system get pod -l component=kube-apiserver -o=jsonpath="{.items[0].metadata.annotations.kubeadm\.kubernetes\.io/kube-apiserver\.advertise-address\.endpoint}"

Still having trouble to connect to that host at that port? run ./scripts/patch-nsp-for-kubectl.sh to make sure the NSPs are updated.

Talk with Us

You can reach us in the #project-wrongsecrets channel of the OWASP Slack Workspace. We'd love to hear any feedback or usage reports you got. If you are not already in the OWASP Slack Workspace, you can join via this link