Cloudsplaining - An AWS IAM Security Assessment Tool That Identifies Violations Of Least Privilege And Generates A Risk-Prioritized Report

Cloudsplaining is an AWS IAM Security Assessment tool that identifies violations of least privilege and generates a risk-prioritized HTML report.

• Example report

Documentation
For full documentation, please visit the project on ReadTheDocs.

• Installation
• Cheat sheet
• Example report

Overview
Cloudsplaining identifies violations of least privilege in AWS IAM policies and generates a pretty HTML report with a triage worksheet. It can scan all the policies in your AWS account or it can scan a single policy file.
It helps to identify IAM actions that do not leverage resource constraints. It also helps prioritize the remediation process by flagging IAM policies that present the following risks to the AWS account in question without restriction:

• Data Exfiltration (s3:GetObject, ssm:GetParameter, secretsmanager:GetSecretValue)
• Infrastructure Modification
• Resource Exposure (the ability to modify resource-based policies)
• Privilege Escalation (based on Rhino Security Labs research)

Cloudsplaining also identifies IAM Roles that can be assumed by AWS Compute Services (such as EC2, ECS, EKS, or Lambda), as they can present greater risk than user-defined roles - especially if the AWS Compute service is on an instance that is directly or indirectly exposed to the internet. Flagging these roles is particularly useful to penetration testers (or attackers) under certain scenarios. For example, if an attacker obtains privileges to execute ssm:SendCommand and there are privileged EC2 instances with the SSM agent installed, they can effectively have the privileges of those EC2 instances. Remote Code Execution via AWS Systems Manager Agent was already a known escalation/exploitation path, but Clou dsplaining can make the process of identifying theses cases easier. See the sample report for some examples.
You can also specify a custom exclusions file to filter out results that are False Positives for various reasons. For example, User Policies are permissive by design, whereas System roles are generally more restrictive. You might also have exclusions that are specific to your organization's multi-account strategy or AWS application architecture.

Motivation
Policy Sentry revealed to us that it is possible to finally write IAM policies according to least privilege in a scalable manner. Before Policy Sentry was released, it was too easy to find IAM policy documents that lacked resource constraints. Consider the policy below, which allows the IAM principal (a role or user) to run s3:PutObject on any S3 bucket in the AWS account:
This is bad. Ideally, access should be restricted according to resource ARNs, like so:
Policy Sentry makes it really easy to do this. Once Infrastructure as Code developers or AWS Administrators gain familiarity with the tool (which is quite easy to use), we've found that adoption starts very quickly. However, if you've been using AWS, there is probably a very large backlog of IAM policies that could use an uplift. If you have hundreds of AWS accounts with dozens of policies in each, how can we lock down those AWS accounts by programmatically identifying the policies that should be fixed?
That's why we wrote Cloudsplaining.
Cloudsplaining identifies violations of least privilege in AWS IAM policies and generates a pretty HTML report with a triage worksheet. It can scan all the policies in your AWS account or it can scan a single policy file.

Installation

• Homebrew

• Pip3

• Now you should be able to execute cloudsplaining from command line by running cloudsplaining --help.

Scanning an entire AWS Account

Downloading Account Authorization Details
We can scan an entire AWS account and generate reports. To do this, we leverage the AWS IAM get-account-authorization-details API call, which downloads a large JSON file (around 100KB per account) that contains all of the IAM details for the account. This includes data on users, groups, roles, customer-managed policies, and AWS-managed policies.

• You must have AWS credentials configured that can be used by the CLI.
  
• You must have the privileges to run iam:GetAccountAuthorizationDetails. The arn:aws:iam::aws:policy/SecurityAudit policy includes this, as do many others that allow Read access to the IAM Service.
  
• To download the account authorization details, ensure you are authenticated to AWS, then run cloudsplaining's download command:
  

• If you prefer to use your ~/.aws/credentials file instead of environment variables, you can specify the profile name:

It will download a JSON file in your current directory that contains your account authorization detail information.

Create Exclusions file
Cloudsplaining tool does not attempt to understand the context behind everything in your AWS account. It's possible to understand the context behind some of these things programmatically - whether the policy is applied to an instance profile, whether the policy is attached, whether inline IAM policies are in use, and whether or not AWS Managed Policies are in use. Only you know the context behind the design of your AWS infrastructure and the IAM strategy.
As such, it's important to eliminate False Positives that are context-dependent. You can do this with an exclusions file. We've included a command that will generate an exclusions file for you so you don't have to remember the required format.
You can create an exclusions template via the following command:
This will generate a file in your current directory titled exclusions.yml.
Now when you run the scan command, you can use the exclusions file like this:
For more information on the structure of the exclusions file, see Filtering False Positives

Scanning the Authorization Details file
Now that we've downloaded the account authorization file, we can scan all of the AWS IAM policies with cloudsplaining.
Run the following command:
It will create an HTML report like this:


It will also create a raw JSON data file:

• default-iam-results.json: This contains the raw JSON output of the report. You can use this data file for operating on the scan results for various purposes. For example, you could write a Python script that parses this data and opens up automated JIRA issues or Salesforce Work Items. An example entry is shown below. The full example can be viewed at examples/output/example-authz-details-results.json

See the examples/files folder for sample output.

Filtering False Positives
Resource constraints are best practice - especially for system roles/instance profiles - but sometimes, these are by design. For example, consider a situation where a custom IAM policy is used on an instance profile for an EC2 instance that provisions Terraform. In this case, broad permissions are design requirements - so we don't want to include these in the results.
You can create an exclusions template via the following command:
This will generate a file in your current directory titled exclusions.yml.
The default exclusions file looks like this:

• Make any additions or modifications that you want.
  • Under policies, list the path of policy names that you want to exclude.
  • If you want to exclude a role titled MyRole, list MyRole or MyR* in the roles list.
  • You can follow the same approach for users and groups list.

Now when you run the scan command, you can use the exclusions file like this:

Scanning a single policy
You can also scan a single policy file to identify risks instead of an entire account.
The output will include a finding description and a list of the IAM actions that do not leverage resource constraints.
The output will resemble the following:

Cheatsheet

FAQ
Will it scan all policies by default?
No, it will only scan policies that are attached to IAM principals.
Will the download command download all policy versions?
Not by default. If you want to do this, specify the --include-non-default-policy-versions flag. Note that the scan tool does not currently operate on non-default versions.
I followed the installation instructions but can't execute the program via command line at all. What do I do?
This is likely an issue with your PATH. Your PATH environment variable is not considering the binary packages installed by pip3. On a Mac, you can likely fix this by entering the command below, depending on the versions you have installed. YMMV.
I followed the installation instructions but I am receiving a ModuleNotFoundError that says No module named policy_sentry.analysis.expand. What should I do?
Try upgrading to the latest version of Cloudsplaining. This error was fixed in version 0.0.10.

References

• Policy Sentry by Kinnaird McQuade at Salesforce
• Parliament by Scott Piper at Summit Route and Duo Labs.
• AWS Privilege Escalation Methods by Spencer Gietzen at Rhino Security Labs
• Understanding Access Level Summaries within Policy Summaries
• Leveraging next-generation blockchain-based AI across multiple service meshes to transparently automate multi-cloud IAM wizardry 慄‍♂️