Cloud for Penetration Testers
Cloud pentesting examines the configuration and identity setup of platforms such as AWS, Azure and Google Cloud. Most findings are misconfigurations rather than software bugs: public storage buckets, over-permissive IAM roles, exposed metadata services and leaked access keys. Testing needs explicit permission from both the customer and, where required, the provider's policy.
Stage: Application Testing · Part of the Penetration Tester roadmap · Last reviewed 19 November 2026
What to learn in Cloud
- Review IAM policies for wildcard actions and excessive permissions
- Find publicly exposed storage buckets and check their contents safely
- Explain how a server-side request to a metadata endpoint can steal credentials
- Use a cloud auditing tool to enumerate resources in a test account
- Describe the shared responsibility model and what testers may touch
Cloud interview questions
Why is the cloud metadata service a pentest target?
Instances can fetch temporary role credentials from a link-local metadata address. A server-side request forgery bug can make the app fetch that address and leak the credentials. Enforcing IMDSv2 on AWS makes this harder.
How does the shared responsibility model limit what a cloud penetration test may cover?
The cloud provider secures the underlying infrastructure, while the customer secures what they configure on it, such as identities, data, network rules and operating systems on virtual machines. Pentesters focus on the customer's side and must follow provider testing policies. Testing stays inside the customer's scope and the provider's rules.
Where Cloud fits in the Penetration Tester roadmap
Cloud belongs to the Application Testing stage (advanced level), which takes about 38 hours. Web and API.
Other units in this stage
Back to the full Penetration Tester roadmap