Practice exams / CGE-AUD CGE-AUD practice exam 20 questions across the 6 exam domains, weighted like the real exam. Untimed. Submit to see your score by domain and an explanation for every answer. The real exam is 50 questions in 75 minutes and you pass at 70%. These questions are for practice and do not appear on the graded exam.
Question 1 of 20 · Cloud & Code Literacy for Auditors A client runs its application on EC2 virtual machines that its own engineers launch and manage. Under the shared responsibility model, which control belongs on the customer side of the audit scope? Physical access restrictions at the data center facilities Maintenance and patching of the underlying hypervisor layer Patching the guest operating system on those instances Secure disposal of failed storage hardware at the provider
Question 2 of 20 · Cloud & Code Literacy for Auditors An auditor is handed the policy attached to a reporting application's role: "Effect": "Allow" | "Action": "s3:GetObject" | "Resource": "arn:aws:s3:::acme-payroll-exports/*". What does this statement permit? Reading objects in the acme-payroll-exports bucket only Reading and writing objects in every bucket in the account Deleting the acme-payroll-exports bucket and its contents Listing every bucket in the account, but not opening them
Question 3 of 20 · Cloud & Code Literacy for Auditors An auditor testing database controls opens a Terraform file and finds: resource "aws_db_instance" "orders" { engine = "postgres" | storage_encrypted = false | publicly_accessible = true }. What is the correct reading of this block? The database is still encrypted, because the provider overrides a false value at deployment It shows only the engine type, since security settings live in the state file, not the code It cannot be interpreted until an engineer applies it and exports the resulting configuration It defines a Postgres database with encryption at rest off and public access on
Question 4 of 20 · Cloud & Code Literacy for Auditors A repository's branch protection requires one approving review before merging to main. In a sample, the auditor reads this pull request record: PR #482 "Update database connection settings" | Author: mlopez | Approving reviews: 0 | Status checks: 3 passed | Merged to main by: mlopez. What is the most appropriate response? Accept the change, because three passing status checks stand in for the approving review Treat it as a merge without independent review and trace how the rule was bypassed Conclude branch protection was never enabled at any point in the period and stop relying on it Replace the item with another pull request, because self-merged changes fall outside the population
Question 5 of 20 · Cloud & Code Literacy for Auditors During scoping, a client states that all customer data is stored in eu-central-1. What does that identifier refer to? An AWS region, a geographic area with several availability zones A single data center building operated by the cloud provider An availability zone located inside a larger United States region The account alias the client chose for its European subsidiary
Question 6 of 20 · Auditing Infrastructure as Code A client says every merged change has been applied, so code and production match. The auditor asks for terraform plan to be run against production and receives: ~ aws_s3_bucket_public_access_block.logs | block_public_policy = false -> true | Plan: 0 to add, 1 to change, 0 to destroy. What does this output tell the auditor? The code was edited to turn the setting off, and the plan is waiting to push that edit to production Nothing is being added or destroyed, so deployed infrastructure matches the code and no drift exists Running the plan already corrected the setting, so the deployed bucket is back in line with the code The deployed setting is false while the code says true, which points to a change made outside the code
Question 7 of 20 · Auditing Infrastructure as Code An auditor reviewing a storage module needs to know when the line that sets the encryption algorithm was last changed and who reviewed that change. Which evidence source answers both questions? The Terraform state file, which records the reviewer of each attribute The version control history for the file, tied to its pull request The module's outputs.tf, which lists each change with its author The module README, which carries a dated approval for each edit
Question 8 of 20 · Auditing Infrastructure as Code A client's Terraform defines 60 databases through one shared module. Four more databases are written as standalone resource blocks in the root configuration, each with its own settings. How should the auditor scope encryption testing? Test the shared module once and treat that result as covering all 64 databases Pull a random sample of 25 from the 64 databases and test each one the same way Test the shared module once for the 60, then test each of the 4 standalone blocks Test the shared module once and exclude the 4, since they sit outside the module
Question 9 of 20 · Auditing CI/CD Pipelines A client says a failed vulnerability scan blocks production deploys. Its workflow file shows: jobs: test | scan (needs: test) | deploy (needs: test). What should the auditor conclude? Deploy waits only on test, so a failed scan would not stop a production deploy Deploy waits on scan through test, so the client's description of the gate holds Deploy runs first, because jobs execute in reverse order of how they are listed Deploy needs a human approver, because needs: names the team that must sign off
Question 10 of 20 · Auditing CI/CD Pipelines A client wants every change under .github/workflows/ to be reviewed by the security team before it merges. Which mechanism enforces that on the platform? A notice in the repository README asking engineers to tag security Repository admin rights granted to each security team member A longer retention period for workflow run logs and artifacts A CODEOWNERS entry for that path, with code owner review required
Question 11 of 20 · Auditing CI/CD Pipelines An auditor traces one production release and reads the run record: commit 7f3e9d2 by rchen | unit tests: PASS | dependency scan: PASS, 0 high | production approval: granted by rchen | deploy: SUCCESS. What should the auditor conclude? The release is clean, because every automated gate passed before the deploy step ran The automated gates passed, but the author approved their own release, a duties gap The scan result is unreliable, because zero high findings is not a realistic outcome The record is incomplete evidence until management signs a copy of the run summary
Question 12 of 20 · Auditing Cloud-Native Monitoring & Continuous Controls While reviewing a year of AWS audit logs, an auditor finds this event: eventName: StopLogging | userIdentity: dkim | eventTime: 2026-08-09T03:15Z | requestParameters: name = org-trail. Why does this event matter? It is routine trail maintenance and has no effect on the evidence for the period It shows that log files already delivered to the S3 bucket were permanently deleted The trail stopped recording, so find when it resumed and who approved the pause It proves dkim made unauthorized changes, so the logging control fails for the full year
Question 13 of 20 · Auditing Cloud-Native Monitoring & Continuous Controls An automated check flags any access key older than 90 days. Its execution log shows hourly runs with no gaps for the full year. The check's configuration history shows the threshold was changed to 365 days in July. What should the auditor conclude? The control operated effectively all year, because the execution log has no gaps The control failed for the full year, because any configuration change voids it The threshold change is out of scope, because the auditor tests executions only The check ran all year, but from July it tested a weaker rule that needs evaluating
Question 14 of 20 · Auditing Cloud-Native Monitoring & Continuous Controls Policy requires compliance drift to be fixed within 5 days. An auditor reads this record: Config rule: restricted-ssh | Resource: sg-0a12 | NON_COMPLIANT: 2026-05-02 | Alert sent: 2026-05-02 | Ticket OPS-311 opened: 2026-05-02 | COMPLIANT: 2026-05-03 | Ticket closed: 2026-05-03. How should the auditor read it? Drift occurred, and the alert, ticket, and fix chain worked within policy No drift occurred, because the resource ended the period in a compliant state The response failed, because any NON_COMPLIANT status breaches the five day policy The record cannot be used until a console screenshot of the fixed rule is provided
Question 15 of 20 · Claude Code & AI Tooling for Auditors An auditor wants Claude Code to pull branch protection settings directly from a client's GitHub organization. Which part of the extension system provides that connection to the external system? A SKILL.md playbook A persona definition An MCP server A slash command
Question 16 of 20 · Claude Code & AI Tooling for Auditors Claude drafts a finding: "14 IAM users lack MFA. Evidence: iam-credential-report.csv." The auditor opens the cited file and counts 9 users with mfa_active set to false. What should the auditor do? Change 14 to 9 and file the rest of the draft, since the other findings were not disputed Reject the draft as failing a spot check, then verify its claims against the raw evidence Keep 14, since the tool read the whole file and a manual count is more likely to be wrong Ask the client which number is right and record their answer as support for the finding
Question 17 of 20 · Claude Code & AI Tooling for Auditors An auditor needs a Club toolkit plugin that collects configuration evidence from a client's AWS account. Which category of plugin does that job? A framework plugin, such as soc2 A persona plugin, such as grc-auditor An OSCAL plugin for machine-readable output A connector plugin, such as aws-inspector
Question 18 of 20 · Evidence Evaluation in a GRC-Engineered Organization A client emails the auditor a file named mfa_status.csv and says an engineer produced it with a script. Before relying on it, what should the auditor establish? The script used, the source system, when it ran, and that the population is complete That the engineer's manager reviewed the file and signed a statement that it is accurate That the file was converted to PDF before it was sent, so that nobody could edit the rows Nothing more, because any export that comes out of a system counts as system-generated
Question 19 of 20 · Evidence Evaluation in a GRC-Engineered Organization An auditor chooses temporal sampling to test a daily log review control over a calendar year audit period. Which selection best supports a conclusion for the full period? The final four weeks of December, when the activity is most recent The four weeks with the lowest event volume, to limit review time One week from each of the four quarters, with reasons written down Any single month, provided that every day in that month is reviewed
Question 20 of 20 · Evidence Evaluation in a GRC-Engineered Organization An auditor tested encryption on databases built from a Terraform module. Which evidence reference in the workpaper would best hold up to a reviewer? Module code for the database build, as shown on screen by the platform lead during a walkthrough modules/rds/main.tf line 42, storage_encrypted = true, at commit 9c1e7fa, with a clean dated plan Console screenshots of 25 database instances, provided by management, each showing encryption on Terraform repository reviewed in full by the audit team, with no encryption gaps noted in any file
Submit answers 0 of 20 answered