SIGPwny's CTF Challenge Repo Template

Flag format: sigpwny{...}

For Challenge Devs: Adding a Challenge

• Do you need a container?

  • YES:
    • cd into /challenges/<category>
    • kctf chal create --template <templatename> <chalname> --challenge-dir ./<chalname>
    • Available templates: pwn, web, xss-bot
    • Note: the kCTF config is challenge.yaml and the CTFd config is challenge.yml. Confusing? Yes.
  • NO:
    • mkdir /challenges/<category>/<chalname>

• Your challenge folder MUST have a challenge.yml file for CTFd, following the specification here

• Your challenge must have a healthcheck script if it is deployable - attempt to make it solve the challenge

• Your challenge should have a SOLUTION.md writeup (it's ok if it's simple/concise or a TL;DR version)

For Challenge Devs: Local Development for Containerized Challenges

Initial setup

• Follow kCTF setup instructions here
  • umask a+x
  • Install dependencies (CLI tools, Docker)
  • Enable user namespaces:
    • `echo 'kernel.unprivileged_userns_clone=1' | sudo tee -a /etc/sysctl.d/00-local-userns.conf
    • sudo service procps restart
• Helpful: export DOCKER_SCAN_SUGGEST=false - disables annoying Snyk messages from newer Docker versions which break kCTF parsing

After initial setup

Every time you open a new shell, you will need to do the following:

• cd to root of this repository
• source kctf/activate

Testing locally

• Switch to and start local cluster:
  • kctf cluster load local-cluster
  • kctf cluster start
• Start challenge and port forward to access it:
  • kctf chal start
  • kctf chal debug port-forward
• When done testing:
  • kctf cluster stop to shutdown local k8s cluster
    • Do NOT run this command on remote-cluster or you will delete the Google Cloud cluster
  • deactivate to exit ctfcli

Testing deployed challenge on remote cluster

• Push to repo, and run the kCTF GitHub action
• Switch to remote cluster:
  • kctf cluster load remote-cluster
• Port forward to access it:
  • kctf chal debug port-forward
• When done testing:
  • deactivate to exit ctfcli

For Infrastructure Admins: Setting Up Google Cloud

These instructions only need to be done once before the CTF.

Prerequisites

• Install gcloud: https://cloud.google.com/sdk/docs/install-sdk
• Authenticate with Google Cloud: gcloud auth login
• Follow kCTF setup instructions here

Set up Kubernetes

Create cluster:

kctf cluster create --project dotted-forest-314903 --domain-name chal.uiuc.tf --start --email-address sigpwny@gmail.com --zone us-central1-a --registry us.gcr.io remote-cluster --disable-src-ranges

Note: --disable-src-ranges disables Cloud Armor. To remove, you need the SECURITY_POLICIES quota.

Resize cluster (to reduce costs before CTF starts):

kctf cluster resize --min-nodes 1 --max-nodes 1 --num-nodes 1 --machine-type e2-standard-4 --pool-name default-pool --spot

Test challenge deployment

cd to a challenge folder with a deployment challenge.yaml file and run the following:

kctf chal start

Set up CTFd

Enable services

You may need to enable SQL and Redis services. Run the following commands. If you see a prompt like API [sqladmin.googleapis.com] not enabled on project [648434879266]. Would you like to enable and retry (this will take a few minutes)? (y/N)?, press y.

gcloud sql instances list
gcloud redis instances list --region us-central1

Setup script

Run from the root directory:

./ctfd/chal setup

Setting up CI/CD

GitHub Actions needs some secrets to automatically sync with the CTFd instance. After creating a CTFd admin account, go to http:///settings#tokens to obtain a token.

From the root of the repository, create the .ctf/config file with the new IP and token. Note that you need git-crypt to unlock and edit the file. These credentials will be automatically used by the GitHub Actions workflow to connect to CTFd and sync/install challenges.