mitraed/s3-inspector

Tool to check AWS S3 bucket permissions

Tool to check AWS S3 bucket permissions.

Compatible with Linux, MacOS and Windows, python 2.7 and 3. May be used as AWS Lambda function.

What it does

• Checks all your buckets for public access
• For every bucket gives you the report with:
  • Indicator if your bucket is public or not
  • Permissions for your bucket if it is public
  • List of URLs to access your bucket (non-public buckets will return Access Denied) if it is public

Prerequisites

Create a new IAM User

• Create IAM user with AmazonS3ReadOnly policy attached
  • Go to IAM (https://console.aws.amazon.com/iam/home)
  • Click "Users" on the left hand side menu
  • Click "Add user"
  • Fill in user name and check Programmatic access
  • Click "Next: Permissions"
  • Click "Attach existing policies directly"
  • Check AmazonS3ReadOnly policy
  • Click "Next: Review"
  • Click "Create user"
  • Copy the credentials
    • Access key ID
    • Secret access key
• Create ~/.aws/credentials file or paste the credentials in when you run the script
  • Put the credentials you copied in the previous step here in this format:

[default]
aws_access_key_id = <your access key ID goes here>
aws_secret_access_key = <your secret_access_key goes here>

Use existing configured IAM User

• use your existing credentials or profile if you have a file ~/.aws/credentials like this:

[default]
aws_access_key_id = <your access key ID goes here>
aws_secret_access_key = <your secret_access_key goes here>
[my_profile_name]
aws_access_key_id = <your access key ID goes here>
aws_secret_access_key = <your secret_access_key goes here>

• and pass the profile name or leave blank for default when requested:

python s3inspector.py
Enter your AWS profile name [default]:

Usage

python s3inspector.py

Report example

Usage as Lambda function

Lambda function to perform the same check as above.

Lambda Setup & Prerequisites

Rather than a IAM user, we need a role that permits lambda execution as well as read-only access to S3 buckets and the ability to publish to SNS. First we need to create an SNS endpoint.

• Go to the SNS console (https://console.aws.amazon.com/sns/v2/home)
• Select along the sidebar 'Topics'
• In the topics screen, click 'Create New Topic'
• In the popup, add the name and description
• Click 'Create Topic'
• When the topic finishes creation, enter the topic by clicking on the ARN
• Click 'Create Subscription'
• In the popup, change the protocol to 'EMail'
• Enter the email address of whoever will be sent the reports in the 'Endpoint'
• Click 'Create subscription'
• Select the subscription and click 'Request confirmations'
• In the receivers email client, confirm the subscription via the link provided.
• Copy arn of created topic(can be viewed under 'Topic details') and set this value to SNS_RESOURCE_ARN variable in s3inspector.py.

Once done we can now create the lambda function

• Go to the lambda console (https://console.aws.amazon.com/lambda/home)
• Click on 'Create Function'
• Click on 'Author from Scratch'
• Give the function the name 's3inspector' (or the name of the file containing the function)
• Apply the role created above
• Click 'Create Function'
• On the configuration page
  • Change the Runtime to 'Python 2.7'
  • Change the Handler to 's3inspector.lambda_handler'
• Copy & Paste the contents of the lambda function file into the onscreen editor & click 'Save'
• Increase the timeout of the function to something suitable for the number of S3 buckets in the account (we tested with 1 minute and 128Mb)

You can now run the function with an empty test event, or configure a trigger for the function.