Skip to content
Cloud Security Risk Assessment

Cloud security risk assessment: score your cloud access layer in 20 minutes

The cloud security risk assessment asks 40 questions about the cloud access layer of AWS, Azure and Google Cloud, in six domains, and takes about 20 minutes. It returns a score per domain, the five fixes that matter most, and a PDF and XLSX export made in your browser. Answers never leave the page.

40 questions in 6 domains. Answers stay in this browser; nothing is sent, nothing is stored.

Identities and entitlements (domain weight 25)

1. Does the root user (or Global Administrator, or Organization Admin) of every account have phishing-resistant MFA and no access keys?

From CIS AWS Foundations Benchmark v5.0 1.4 and 1.5: No root access keys and MFA on root. Why this question

2. Do all people reach the cloud through central single sign-on with MFA, with no local IAM users or standalone cloud passwords?

From Microsoft cloud security benchmark IM-6: Use strong authentication controls. Why this question

3. Are long-lived access keys and service account keys absent, or rotated at least every 90 days and unused ones deleted?

From Google Cloud security best practices Service account keys: Best practices for managing service account keys. Why this question

4. Do workloads get cloud access through attached roles, managed identities or workload identity federation instead of stored secrets?

From Microsoft cloud security benchmark IM-3: Manage application identities securely and automatically. Why this question

5. Is the number of people with administrator-level roles small, named and separate from their daily accounts?

From Microsoft cloud security benchmark PA-1: Separate and limit highly privileged/administrative users. Why this question

6. Are privileged roles granted only for a limited time, on request, instead of held permanently?

From Microsoft cloud security benchmark PA-2: Avoid standing access for user accounts and permissions. Why this question

7. Are permissions reviewed at least every 90 days and unused ones removed?

From Microsoft cloud security benchmark PA-4: Review and reconcile user access regularly. Why this question

8. Are emergency (break-glass) accounts in place, protected, monitored and tested at least twice a year?

From Microsoft cloud security benchmark PA-5: Set up emergency access. Why this question

Network reachability (domain weight 20)

9. Is every SSH, RDP and database port closed to the whole internet (0.0.0.0/0 and ::/0)?

From Google Cloud security best practices VPC firewall rules: VPC firewall rules overview. Why this question

10. Do your data stores and platform services (storage, databases, secrets, AI) accept traffic only through private endpoints?

From Microsoft cloud security benchmark NS-2: Secure cloud native services with network controls. Why this question

11. Do new compute instances and databases get no public IP address by default?

From Microsoft cloud security benchmark NS-1: Establish network segmentation boundaries. Why this question

12. Do administrators reach servers through a brokered, logged service (Session Manager, Bastion, Identity-Aware Proxy) with no inbound admin ports?

From Microsoft cloud security benchmark PA-6: Use privileged access solution. Why this question

13. Is the instance metadata service hardened (IMDSv2 required on AWS, managed identity endpoint protected on Azure, metadata concealment on GKE)?

From AWS documentation Instance metadata: Configure the instance metadata service. Why this question

14. Is outbound traffic from production workloads filtered to known destinations?

From Microsoft cloud security benchmark NS-3: Deploy firewall at the edge of enterprise network. Why this question

15. Do you analyze which resources are reachable from the internet and from other networks (peering, transit, cross-account) at least monthly?

From Microsoft cloud security benchmark NS-7: Manage network security centrally and effectively. Why this question

Data perimeter (domain weight 20)

16. Is public access to object storage blocked at the organization level, not just per bucket?

From AWS documentation S3 Block Public Access: Blocking public access to your Amazon S3 storage. Why this question

17. Do your resource policies and sharing rules limit access to identities in your own organization?

From AWS documentation Data perimeter: Building a data perimeter on AWS. Why this question

18. Can your most sensitive data be read only from trusted networks or endpoints (VPC endpoints, service perimeters, private links)?

From Google Cloud security best practices VPC Service Controls: VPC Service Controls overview. Why this question

19. Do you review findings about resources shared outside your organization at least weekly?

From AWS documentation IAM Access Analyzer: Getting started with Access Analyzer. Why this question

20. Are sensitive data encrypted with customer-managed keys whose use is limited to named identities and services?

From Microsoft cloud security benchmark DP-5: Use customer-managed key option in data at rest encryption when required. Why this question

21. Do you know which storage locations hold regulated or sensitive data, and is that list up to date?

From Microsoft cloud security benchmark DP-1: Discover, classify, and label sensitive data. Why this question

22. Is every third-party and cross-account role protected by an external ID or condition, with an owner and an end date?

From AWS documentation IAM best practices: Security best practices in IAM. Why this question

AI workloads (domain weight 15)

23. Does each AI agent, pipeline and application have its own cloud identity, not a shared one or a person's credentials?

From Microsoft cloud security benchmark AI-4: Apply least privilege for agent functions. Why this question

24. Is the list of actions each AI agent can perform limited to what its task needs, and written down?

From Microsoft cloud security benchmark AI-4: Apply least privilege for agent functions. Why this question

25. Can only approved models and AI services be used in your accounts, enforced by policy?

From Microsoft cloud security benchmark AI-1: Ensure use of approved models. Why this question

26. Is the data an AI system can read (retrieval stores, training data, tool outputs) limited to what its users may see?

From Google Cloud security best practices AI and ML security: Google Cloud AI and ML perspective: security. Why this question

27. Do destructive or high-impact agent actions (delete, grant access, pay, send externally) need a human approval?

From Microsoft cloud security benchmark AI-5: Ensure human-in-the-loop. Why this question

28. Are model calls, tool calls and the identity behind each logged and kept?

From Microsoft cloud security benchmark AI-6: Establish monitoring and detection. Why this question

Guardrails and drift (domain weight 10)

29. Do organization-level policies block the riskiest actions before they happen (SCPs, Azure Policy deny effects, organization policy constraints)?

From AWS documentation Service control policies: Service control policies (SCPs). Why this question

30. Do infrastructure changes go through code review and pipelines, with console changes in production flagged?

From Microsoft cloud security benchmark PV-1: Define and establish secure configurations. Why this question

31. Does a continuous check compare live configuration against your rules and report differences within a day?

From AWS documentation AWS Config: What is AWS Config. Why this question

32. Do you test a new guardrail in audit or report-only mode and see what it would block before you enforce it?

From Microsoft documentation Azure Policy: Overview of Azure Policy. Why this question

33. Does every guardrail exception have an owner, a reason and an expiry date?

From Microsoft cloud security benchmark PV-2: Audit and enforce secure configurations. Why this question

34. Can administrators inside a workload account not turn off logging, guardrails or security services?

From Microsoft cloud security benchmark PV-2: Audit and enforce secure configurations. Why this question

Evidence (domain weight 10)

35. Are control-plane audit logs on in every account, subscription and project, in every region, including data access logs for sensitive stores?

From Microsoft cloud security benchmark LT-3: Enable logging for security investigation. Why this question

36. Are audit logs stored in a separate, restricted account where workload administrators cannot change or delete them?

From Microsoft cloud security benchmark LT-5: Centralize security log management and analysis. Why this question

37. Do you keep audit logs for at least 365 days, with the most recent 90 days searchable?

From Microsoft cloud security benchmark LT-6: Configure log storage retention. Why this question

38. Do alerts fire, and reach a person, for root or global administrator sign-ins, policy changes, new keys and disabled logging?

From Microsoft cloud security benchmark LT-2: Enable threat detection for identity and access management. Why this question

39. Within one working day, can you produce a list of who can access a given sensitive data store and how they got that access?

From Microsoft cloud security benchmark PA-4: Review and reconcile user access regularly. Why this question

40. Have you practiced cutting off all access of one compromised identity within one hour, and written down the steps?

From NIST SP 800-53 Rev. 5 IR-4: Incident Handling. Why this question

Answer the questions to see a score per domain and the five fixes that matter most. Unanswered questions count as "not known".

How the score is made ยท The checklist version

The six domains

DomainWeightQuestionsWhy it is weighted
Identities and entitlements258People and workloads that hold standing power over the cloud account are the most common route from one stolen credential to a full breach.
Network reachability207Reachability decides which of your services an attacker can touch before any credential is needed.
Data perimeter207A data perimeter limits who can reach your data and from where, even when an identity or a policy is wrong.
AI workloads156AI services and agents act with cloud credentials, so what they can call and read is an access question.
Guardrails and drift106Preventive controls stop a bad change before it lands, and drift checks catch the ones that slip past.
Evidence106Logs and records are what let you prove who had access, detect misuse and respond.

Workload vulnerabilities and posture scanning are outside this assessment: it scores who and what can reach your cloud resources, not the software running on them.

Other ways to assess

The directory of assessment options lists 15 provider tools, open source scanners and vendor-run assessments by the same fields, in alphabetical order.

Sources

  1. CIS AWS Foundations Benchmark v5.0 1.4 and 1.5, read 16 Nov 2026
  2. Microsoft cloud security benchmark IM-6, read 16 Nov 2026
  3. Google Cloud security best practices Service account keys, read 16 Nov 2026
  4. Microsoft cloud security benchmark IM-3, read 16 Nov 2026
  5. Microsoft cloud security benchmark PA-1, read 16 Nov 2026
  6. Microsoft cloud security benchmark PA-2, read 16 Nov 2026

Last reviewed 16 Nov 2026.