cloud_enum

Multi-cloud OSINT tool. Enumerate public resources in AWS, Azure, and Google Cloud.

Currently enumerates the following:

Amazon Web Services:

  • Open S3 Buckets
  • Protected S3 Buckets

Microsoft Azure:

  • Storage Accounts
  • Open Blob Storage Containers
  • Hosted Databases
  • Virtual Machines
  • Web Apps

Google Cloud Platform

  • Open GCP Buckets
  • Protected GCP Buckets

By "open" buckets/containers, I mean those that allow anonymous users to list contents. if you discover a protected bucket/container, it is still worth trying to brute force the contents with another tool.

IMPORTANT: Azure Virtual Machine DNS records can span a lot of geo regions. To save time scanning, there is a "REGIONS" variable defined in cloudenum/azure_regions.py. You'll want to look at this file and edit it to be relevant to your own work.

Usage

Setup

You'll need the requests-futures python package, as this tool uses it for multi-threading HTTP requests. It's a very cool package if you're already using requests, I highly recommend it.

pip3 install -r ./requirements.txt

Running

The only required argument is at least one keyword. This is what will be used to find the names of S3 Buckets, Azure Storage Accounts, and Google Cloud Buckets. This keyword will be mutated by prepending and then appending the strings in cloudenum/mutations.txt.

You can provide multiple keywords by specifying the -k argument multiple times.

Azure Containers required two levels of brute-forcing, both handled automatically by this tool. First, by finding valid accounts. Then, by brute-forcing the name of containers inside that account. That second layer of discovery will use the strings in the cloud_enum/brute.txt file.

Alternatively, you can specify your own mutation and brute files.

Let's say you were researching "somecompany" whose website is "somecompany.io" that makes a product called "blockchaindoohickey". You could run the tool like this:

cloudenum.py -k somecompany -k somecompany.io -k blockchaindoohickey

By default, the tool uses 5 HTTP threads for brute forcing and enumerating. You can try increasing this, but eventually the cloud providers will rate limit you. Here is an example to increase to 8.

DNS brute-forcing uses a hard-coded 5 threads.

cloudenum.py -k keyword -t 8

Complete Usage Details

usage: cloud_enum.py [-h] -k KEYWORD [-m MUTATIONS] [-b BRUTE]

Multi-cloud enumeration utility. All hail OSINT!

optional arguments:
  -h, --help            show this help message and exit
  -k KEYWORD, --keyword KEYWORD
                        Keyword. Can use argument multiple times.
  -m MUTATIONS, --mutations MUTATIONS
                        Mutations. Default: cloud_enum/mutations.txt.
  -b BRUTE, --brute BRUTE
                        List to brute-force Azure container names. Default:
                        cloud_enum/brute.txt.
  -t THREADS, --threads THREADS
                        Threads for HTTP brute-force. Default = 5
  -ns NAMESERVER, --nameserver NAMESERVER
                        DNS server to use in brute-force.
  --disable-aws         Disable Amazon checks.
  --disable-azure       Disable Azure checks.
  --disable-gcp         Disable Google checks.

Roadmap

I plan to implement some more things, like:

  • Adding content to mutations.txt and brute.txt - they are most POC-length right now.
  • Adding additional public resources, where it makes sense

GitHub