🍪 CookieFarm - Exploiter Manager

Python decorator for automating exploit execution in CTF Attack & Defense competitions

📦 What is it?

This package provides a @exploit_manager decorator designed to automate the parallel execution of exploits in CTF (Attack & Defense) settings, specifically for use with the CookieFarm project.

It handles:

• Authentication with the central server
• Retrieving team configuration
• Automatic flag parsing from stdout

⚠️ Note: This package is not standalone. It must be used together with the CookieFarm client. The client provides the required APIs and team configurations.

📦 Installation

To install the package:

pip install --upgrade cookiefarm requests

⚙️ How it works

The @exploit_manager decorator takes care of:

• Calling your exploit(ip, port, name_service, flag_ids) function
• Retrieving the CookieFarm server configuration
• Fetching the full flag IDs JSON at every tick
• Normalizing competition-specific flag IDs structures into a per-team list
• Passing only the current team/service flag IDs to each exploit thread
• Capturing your exploit's stdout
• Parsing flags via regex
• Logging the result in JSON format, including: team ID, port, service name, and the flag found

🚀 Example usage

from cookiefarm import exploit_manager
import requests

@exploit_manager
def exploit(ip, port, name_service, flag_ids):
    # flag_ids contains only the IDs for the current team and service
    for flag_id in flag_ids:
        response = requests.get(
            f"http://{ip}:{port}/",
            params={"id": flag_id},
        )

        # Just print the flag to stdout
        print(response.text)

# Run from the command line with arguments from CookieFarm
# python3 myexploit.py -s <server_address> -t <tick_time> -W <worker_number> -p <port> -n <name_service>

For execution, you have to pass the required arguments from the command line, which are provided by the CookieFarm client. The decorator will handle the rest.

python3 myexploit.py -s <server_address> -t <tick_time> -W <worker_number> -p <port> -n <name_service> -x [test mode]

🧩 Flag IDs normalization

Different Attack/Defense infrastructures may expose flag IDs with different JSON layouts. CookieFarm fetches the full flag IDs JSON at every tick and normalizes it locally according to the shared.flagids_format value configured on the server.

The exploit function always receives:

flag_ids: list

This list contains only the flag IDs for the current team and the current service.

Format syntax

The flagids_format path supports dynamic tokens:

• [service] — the service name passed with -n / --name_service
• [team] — the team key to iterate over
• [id] — the terminal node containing the flag IDs

Literal keys can be written directly in the path.

CyberChallenge-style layout

Raw flag IDs:

{
  "Service1": {
    "team_1": {
      "1": ["id-a"],
      "2": ["id-b"]
    }
  }
}

Server configuration:

shared:
  url_flag_ids: "http://10.10.0.1/flagids"
  flagids_format: "[service].[team].[id]"

The exploit running against Service1 and team_1 receives:

["id-a", "id-b"]

Team-first layout

Raw flag IDs:

{
  "team_1": {
    "Service1": ["id-a", "id-b"]
  }
}

Server configuration:

shared:
  url_flag_ids: "http://172.168.1.0/flagids"
  flagids_format: "[team].[service].[id]"

The exploit running against Service1 and team_1 receives:

["id-a", "id-b"]

Nested layout with a literal key

Raw flag IDs:

{
  "teams": [123, 456, 789],
  "flag_ids": {
    "service1": {
      "123": ["abc123", "def456"],
      "789": ["xxx", "yyy"]
    }
  }
}

Server configuration:

shared:
  url_flag_ids: "http://example.local/flagids"
  flagids_format: "flag_ids.[service].[team].[id]"

The exploit running against service1 and team 123 receives:

["abc123", "def456"]

If flagids_format is omitted, CookieFarm uses:

flagids_format: "[service].[team].[id]"

🛠️ Requirements

• Python ≥ 3.13
• Working CookieFarm client installed

📝 License

Distributed under the MIT License. Feel free to use, modify, and contribute.

For any questions, suggestions, or issues, feel free to open a GitHub issue!

Created with ❤️ by ByteTheCookies (feat. @0xMatte)