Awspx - A Graph-Based Tool For Visualizing Effective Access And Resource Relationships In AWS Environments

auspex [ˈau̯s.pɛks] noun: An augur of ancient Rome, especially one who interpreted omens derived from the observation of birds.

awspx is a graph-based tool for visualizing effective access and resource relationships within AWS. It resolves policy information to determine what actions affect which resources, while taking into account how these actions may be combined to produce attack paths. Unlike tools like Bloodhound, awspx requires permissions to function. It is not expected to be useful in cases where these privileges have not been granted.

Quick start
Install (see installation), load the sample database, and search for attacks:

awspx db --load-zip sample.zip
awspx attacks

OR run it against an environment of your own (attack information is included by default in this case):

awspx ingest

Browse to localhost and see what you can find!


Installation
awspx requires Docker.

git clone git@github.com:FSecureLABS/awspx.git
cd awspx && ./INSTALL 

If it doesn't work out of the box, here are some things to check:

• The docker container runs a Neo4j database that will forward TCP ports 7687, 7373 and 7474 to these same ports on localhost. If an existing Neo4j installation is present (e.g. BloodHound) awspx will fail. You will need to disable this service before continuing. Alternatively, you can modify network mappings yourself by editing INSTALL.
• The docker container also forwards to TCP port 80, resulting in similar issues.
• SELinux may prevent the docker container from doing everything it needs to. If you are running SELinux (props) and encounter issues, check SELinux.
• Docker makes changes to iptables. You may need to adjust your iptables configuration to get awspx to work.

AWS permissions
The following AWS-managed policies can be used.

• SecurityAudit will allow you to ingest everything except S3 objects.
• Add ReadOnlyAccess to also ingest S3 objects (warning: this can be very slow).

Data collection
Once awspx has been installed, you can create a profile by running awspx profile --create my-account, or invoke the ingestor by running awspx ingest on the command line. By default the ingestor will utilise a profile called default unless you specify something else using --profile:

awspx ingest --profile my-other-account

If the profile my-other-account does not exist, you will prompted to enter a AWS access key ID and secret for it. You will also be prompted an output format, which you can ignore, and a region which is not important for IAM but required for other services. You can also create a profile this without ingesting any data by using awspx profile:

awspx profile --create work

Further commands and arguments are provided for tweaking ingestion and attack path computation, and for managing AWS profiles and Neo4j databases. Run awspx -h and awspx {profile|ingest|attacks|db} -h to learn more.
Supported services: IAM, EC2, S3, Lambda

Examples

awspx ingest --profile my-account --services S3

The ingestor will pull only S3 data using the my-account profile and store it in a database named my-account.db. Resource based policies (and Bucket ACLs in this case) will be processed automatically. Identify based policies will be ignored since IAM has been omitted from this list of services.

awspx ingest --profile my-account --services IAM EC2 --database db-for-ec2

The ingestor will pull only IAM and EC2 data, using the my-account profile, and store it in a database named db-for-ec2.db. Since IAM includes Identity based policies and assume role policy documents, this infromation will be included in db-for-ec2.db

awspx ingest --profile my-account \
    --except-types AWS::S3::Object \
    --except-arns arn:aws:s3:::broken-bucket arn:aws:ec2:eu-west-1:123456789012:instance/i-1234

awspx will pull data for all supported services using the my-account profile but will not attempt to load S3 objects. It will also skip the bucket named broken-bucket and the EC2 instance named i-1234. A full list of recognised resource types can be found in lib/aws/resources.py.

awspx ingest --profile my-account --skip-attacks

awspx will pull data for all supported services using the my-account profile but will not compute attacks. This can be useful for large environments. Attacks can be computed separately later on by running awspx attacks.

awspx attacks --only-attacks AssumeRole CreateGroup

Using the current database, awspx will only compute only the Assume Role and Create Group attacks.

awspx db --load-zip sample.zip

awspx will create a new database named sample from sample ZIP file. Files must be placed in /opt/awspx/data so that they can be accessed by the docker container. Note that attack information is not included with zip data. To include this information awspx attacks must be run after a zip has been loaded.

awspx db --use my-other-account

awspx will switch the database to my-other-account. You will need to refresh your browser to see the changes.

Using the frontend
Once you've loaded a database (hint: load the sample data by running awspx db --load-zip sample.zip) you can explore it by visiting localhost in your browser.
To get started, find a Resource (or Action) you're interested in and see where the path takes you (right click on Resources to bring up the context menu, left click to see its properties).

Action colors
Action Effect color palette:

• Allow: Green edges
• Deny: Red edges
• Conditional: Dashed edges

Action Access Type color palette:

• List: Yellow
• Read: Pink
• Write: Indigo
• Tagging: Teal
• Permissions Management: Purple

Actions are represented visually using a linear gradient comprised of the Effect and Access colors (in that order). Conditional attacks are presented using a dotted line.

Shortcut keys

Key	Action
Alt + Enter	Rerun Layout
Tab	Switch between Actions and Resources search view
Ctrl + Drag	Box select
Ctrl + Left Click	toggle selection
Delete	Remove selected nodes
Escape	Close properties
Ctrl + C	Copy selection properties (JSON)
Ctrl + A	Select all
Ctrl + S	Open search bar

About
awspx was developed by Craig Koorn and David Yates using Python (ingestor); Neo4j (DB); and Vue (front-end); and Cytoscape (front-end graph visualization).