SecuSparkSecuSpark
Practice TestsFlashcardsFeaturesPricingBlogChangelogAbout
Start Free

© 2026 SecuSpark. CompTIA, Security+, A+, Network+, CySA+, PenTest+, and SecAI+ are registered trademarks of CompTIA, Inc. SecuSpark is not affiliated with, endorsed by, or sponsored by CompTIA, Inc.

PrivacyTermsCookies
SecuSparkSecuSpark
Practice TestsFlashcardsFeaturesPricingBlogChangelogAbout
Start Free
  1. Hub
  2. Blog
  3. Free PenTest+ PT0-003 Sample Questions With Answers (2026)
Exam Info

Free PenTest+ PT0-003 Sample Questions With Answers (2026)

20 free CompTIA PenTest+ PT0-003 sample questions with answers and explanations, four per exam domain, weighted like the real exam. No PDF, no signup.

SecuSpark TeamAugust 25, 202612 min read
Pawel's in-game slime character, KingSpark
Pawel SlobodaFounder

Builder of SecuSpark. 24 shipped projects across healthcare, defense, and education. Built this platform because textbooks never worked for my ADHD brain — so I turned exam prep into an RPG. @PawelBuilds

333 Practice Exams7,500+ QuestionsFact-Checked Content

Table of Contents

  • 1.0 Engagement Management (13%)
  • 2.0 Reconnaissance and Enumeration (21%)
  • 3.0 Vulnerability Discovery and Analysis (17%)
  • 4.0 Attacks and Exploits (35%)
  • 5.0 Post-exploitation and Lateral Movement (14%)
  • What the Real PT0-003 Exam Looks Like
  • References

Here are 20 CompTIA PenTest+ (PT0-003) sample questions with the answers and the reasoning, four from each of the five exam domains. They are written to the current PT0-003 objectives — the version that launched December 17, 2024 and replaced PT0-002 [1] — not recycled PT0-002 material. No PDF download, no email gate: read them here, then take a full timed exam free.

Two notes before you start. First, PT0-003 weights Attacks and Exploits at 35%, so that domain gets the same four questions here as everyone else but deserves double the study time. Second, the real exam mixes these multiple-choice items with performance-based questions (PBQs) that hand you tool output and ask what to do next — the questions below mirror that scenario style on purpose. When you want the full 90-question experience, the PenTest+ practice test gives you 3 complete exams free from a bank of 552 questions.

How to use these

Cover the answer box, commit to a letter, then read the explanation even when you were right. On PenTest+ the distractors are usually real techniques applied at the wrong phase — knowing why B beats A is the skill the PBQs test.

1.0 Engagement Management (13%)

Scoping, rules of engagement, legal boundaries, and the remediation advice that goes in the report. Small domain, easy points if you know the process.

1. Halfway through an internal test you recover domain credentials that would also unlock a sister company’s network reachable over a trust relationship. That network is not named in the Statement of Work. What do you do?

  1. Pivot into it while the credentials are valid — the trust makes it in scope
  2. Note the trust in the report but do not test the sister network without a scope change
  3. Test it quietly and only mention it if you find something serious
  4. Delete the credentials so you are not tempted to cross the boundary

Answer: B. The SoW defines what you are authorized to touch, and a network trust does not extend it. Document that the trust exists and that the credentials would reach the other environment, then get a written scope change before testing there. Deleting evidence (D) hides a real finding the client needs.

2. At the report review the client asks you to drop a medium-severity finding because they have decided not to fix it. What is the correct way to handle the request?

  1. Remove it — the client owns the report and the risk decision
  2. Keep the finding but downgrade it to informational to satisfy them
  3. Keep the finding and record that the client has formally accepted the risk
  4. Replace it with a more general recommendation that does not name the system

Answer: C. A penetration-test report is a record of what you found; deleting a real finding misrepresents the system’s security. You document that the client accepted the risk, which keeps the report honest and puts the decision where it belongs. Silently downgrading or genericizing it (B, D) is the same problem in a smaller font.

3. You are about to run an on-site physical and social-engineering test. Which document should you carry to prove the engagement is authorized if a guard or employee challenges you?

  1. The Master Service Agreement
  2. A signed authorization letter naming a client point of contact to call
  3. The non-disclosure agreement
  4. A copy of your PenTest+ certification

Answer: B. The authorization letter (sometimes called a “get out of jail” letter) states that you are permitted to be there and gives a client contact who can confirm it on the spot. The MSA and NDA govern the business relationship, not your presence in a hallway, and a certification proves nothing about authorization.

4. For a single engagement, which document defines the exact systems in scope, the testing dates, and the types of testing allowed?

  1. The Master Service Agreement
  2. The Statement of Work
  3. The Rules of Engagement appendix in the final report
  4. The mutual non-disclosure agreement

Answer: B. The Statement of Work scopes one engagement: which assets, which dates, which test types. The MSA sets ongoing legal terms across engagements, and the NDA covers confidentiality. Knowing which document answers a scoping question is a recurring Engagement Management item.

2.0 Reconnaissance and Enumeration (21%)

Passive OSINT, active discovery, and service enumeration. Expect questions that name a constraint (no active traffic yet, one open port, a default community string) and ask for the technique that still works.

5. The client wants the lightest possible footprint during early recon and has asked you not to send traffic to their servers yet. Which technique still lets you enumerate subdomains?

  1. Brute-forcing hostnames against the target’s DNS server
  2. Running a full port scan of the public IP range
  3. Searching public certificate-transparency logs for issued hostnames
  4. Sending crafted HTTP requests with different Host headers

Answer: C. Certificate-transparency logs record every TLS certificate a CA issues, so they leak subdomains without your ever touching the client’s infrastructure. Brute-forcing DNS, port scanning, and Host-header probing all send traffic to the target, which the client asked you to hold off on.

6. A network device has UDP 161 open and is using a default read community string. What is the most direct way to pull its interface list, routing table, and running-configuration details?

  1. Run an SNMP walk against it with the community string
  2. Open an SSH session and read the config manually
  3. Send ICMP timestamp requests and infer the config
  4. Perform a reverse DNS lookup on every interface

Answer: A. UDP 161 is SNMP; with a valid community string an SNMP walk enumerates the device’s MIB — interfaces, routes, ARP entries, and often configuration data. SSH assumes credentials you may not have, and the other two reveal almost nothing about the device’s internals.

7. A hardened host exposes only TCP 22 and drops everything else. You need to identify the operating system and the SSH implementation. What is the most reliable approach?

  1. Assume it is Linux because port 22 is open
  2. Grab the SSH service banner and run OS and stack fingerprinting against the open port
  3. Launch a UDP scan of the top 1,000 ports
  4. Send a malformed HTTP request and read the error page

Answer: B. The service banner on 22 usually names the SSH implementation and often the distribution, and stack fingerprinting (for example Nmap’s -O) infers the OS from TCP/IP behavior on the one open port. Assuming Linux from a port number (A) is a guess, and there is no web service to error out (D).

8. You are mapping a web application’s attack surface and want to find administrative paths the developers kept out of the navigation. What is the fastest first check before any brute-forcing?

  1. Read robots.txt and the sitemap for disallowed or listed paths
  2. Fuzz the login form with a wordlist
  3. Scan the whole IP range for other web servers
  4. Decompile the site’s JavaScript for API keys

Answer: A. robots.txt frequently lists the exact directories a team wanted hidden from crawlers — admin panels, staging paths, backups — and the sitemap enumerates known URLs. It costs one request and often hands you the endpoints before you spend time on content discovery. The other options are heavier and answer different questions.

3.0 Vulnerability Discovery and Analysis (17%)

Choosing the right scan type, reading results, and deciding what is real. PT0-003 added more application-security and container content here than PT0-002 had.

9. A scan of a client environment returns 240 findings and you have two days. Two of them share the same CVSS base score of 9.8. How should you decide which to pursue first?

  1. Alphabetically, so the report is easy to follow
  2. By which one is exposed to the network and has a working public exploit
  3. By whichever host has more open ports
  4. By the order the scanner reported them

Answer: B. CVSS base score is a starting point, not a ranking. Between two equal scores, the one that is actually reachable and has a proven exploit is the real risk you can demonstrate. Prioritizing by exploitability and exposure is exactly the analysis PT0-003 expects over sorting by the scanner’s output.

10. Before a container image is deployed, the client wants to know whether its base layers ship known-vulnerable packages. Which scan answers that most directly?

  1. A dynamic scan of the running container’s web endpoints
  2. A container image scan that inventories each layer’s packages against vulnerability databases
  3. A port scan of the host running the container
  4. A password-strength audit of the container’s accounts

Answer: B. Image scanning (Trivy, Grype, and similar) reads every layer’s installed packages and matches their versions against known CVEs — the direct answer to “are the base layers vulnerable.” A dynamic scan tests runtime behavior, and a host port scan looks at the wrong boundary.

11. You are assessing a proprietary service that speaks an undocumented binary protocol on a custom port. You want to surface memory-corruption bugs in how it parses input. Which technique fits best?

  1. Static analysis of the public documentation
  2. Fuzzing the service with malformed and mutated inputs while watching for crashes
  3. A credentialed vulnerability scan of the host
  4. Reviewing the TLS certificate chain

Answer: B. Fuzzing feeds the service malformed and mutated inputs and watches for crashes or hangs — the standard way to find parsing and memory-corruption bugs in a closed protocol with no source. A signature-based scan will not understand a custom protocol, and there is no documentation to analyze.

12. You have full source code to a Python API and need to catch hardcoded secrets and dangerous function calls before the service is ever run. Which analysis type fits?

  1. Dynamic application security testing against a deployed instance
  2. Static application security testing of the source code
  3. Software composition analysis of the container registry
  4. An unauthenticated network scan

Answer: B. Static analysis reads the source without executing it, which is exactly what surfaces hardcoded secrets, unsafe calls, and injection sinks before deployment. DAST needs a running target, and SCA looks at third-party dependencies rather than the code the team wrote.

4.0 Attacks and Exploits (35%)

The biggest domain by a wide margin: network, wireless, application, cloud, mobile, and social-engineering attacks, plus the tools that run them. If your practice scores are weak anywhere, make it not here.

13. You have a foothold on a Windows domain and want service-account passwords without triggering account lockouts. You request Kerberos service tickets for accounts that have a Service Principal Name and crack them offline. What is this technique called?

  1. Kerberoasting
  2. Pass-the-hash
  3. LLMNR poisoning
  4. Golden ticket forgery

Answer: A. Requesting TGS tickets for SPN-bearing accounts and cracking them offline is Kerberoasting; because the cracking happens off the wire, it never touches the account’s lockout counter. Pass-the-hash reuses a hash rather than recovering a password, and a golden ticket requires the KRBTGT hash you do not have yet.

14. Testing a billing portal, you change the invoice number in the URL from 4471 to 4470 and the application returns another customer’s invoice. Which vulnerability class is this?

  1. Cross-site scripting
  2. Insecure direct object reference
  3. Server-side request forgery
  4. SQL injection

Answer: B. The app trusts a client-supplied identifier to fetch a record without checking that the record belongs to you — an insecure direct object reference (a broken-access-control flaw). Nothing here injects script or SQL, and no server-side request to another host is involved.

15. You are assessing a corporate wireless network protected by WPA2-PSK. Which sequence lets you recover the passphrase offline?

  1. Set up a rogue DHCP server and wait for clients
  2. Deauthenticate a connected client, capture the 4-way handshake, then crack it offline against a wordlist
  3. ARP-spoof the access point to intercept plaintext
  4. Send oversized frames until the AP reboots into an open state

Answer: B. Forcing a client to reconnect lets you capture the WPA2 4-way handshake, which contains the material needed to test passphrase guesses offline. WPA2 does not fall back to open on a reboot, and ARP spoofing does not defeat the encryption itself.

16. You have a low-privilege shell on a Linux host. Running sudo -l shows you may run /usr/bin/vim as root without a password. What is the most direct path to a root shell?

  1. Edit /etc/passwd by hand in a normal editor session
  2. Use vim’s shell-escape to spawn a root shell from within the editor
  3. Recompile the kernel with a backdoor
  4. Brute-force the root password locally

Answer: B. An editor you can run as root is a classic privilege-escalation primitive: vim can spawn a shell (for example :!/bin/sh) that inherits root, no password needed. This is the GTFOBins pattern the exam expects; the other options are slower, noisier, or need privileges you do not have.

5.0 Post-exploitation and Lateral Movement (14%)

What you do after the first shell: enumerate, persist, move, exfiltrate, and clean up. New as a standalone domain in PT0-003, so older study guides cover it thinly.

17. You have a shell on a Windows workstation and need access to survive a reboot with as little noise as possible. Which technique fits best?

  1. Leave the interactive shell open and hope no one reboots
  2. Register a scheduled task that re-establishes your access at logon
  3. Disable the host firewall entirely
  4. Change the local administrator password to lock others out

Answer: B. A scheduled task (or an equivalent registry Run key) quietly re-runs your payload after a reboot, which is what persistence means. Holding a session open does not survive a restart, and disabling the firewall or changing the admin password is loud and likely to trip alerts.

18. From a compromised host you hold a local administrator NTLM hash that is reused on other workstations. You want to run commands on a second machine without cracking the password. Which technique applies?

  1. Pass-the-hash over SMB to authenticate as that admin on the second host
  2. A dictionary attack against the domain controller
  3. DNS cache poisoning of the second host
  4. Downgrading the second host to LM hashing

Answer: A. Pass-the-hash presents the captured NTLM hash directly to SMB on the target, so a reused local-admin hash gives you code execution on the second workstation with no cracking step. The other options either attack the wrong target or require conditions the scenario does not give you.

19. The compromised network allows only DNS traffic outbound; HTTP, HTTPS, and everything else are blocked at the egress firewall. How can you exfiltrate a small amount of data?

  1. Email the data from the host’s mail client
  2. Encode the data into DNS queries to a name server you control (DNS tunneling)
  3. Upload it to a cloud storage bucket over HTTPS
  4. Open a reverse shell on port 4444

Answer: B. When DNS is the only protocol permitted out, you encode the data into lookups for a domain whose authoritative server you control and reassemble it on the other side — DNS tunneling. Every other option needs a port or protocol the firewall is dropping.

20. You have compromised a dual-homed host that can reach an internal segment your testing laptop cannot route to. What is the standard way to reach that segment through the host?

  1. Reconfigure the client’s core router to add a static route to your laptop
  2. Set up a pivot such as a SOCKS proxy or port forwarding through the compromised host
  3. Ask the client to place your laptop on the internal VLAN
  4. Broadcast your laptop’s ARP entry across both segments

Answer: B. Pivoting routes your tools through the foothold: a SOCKS proxy or port forward over your session lets you reach the otherwise-unroutable segment using the compromised host’s connectivity. Changing the client’s routing or asking for VLAN access defeats the point of testing the segmentation.

What the Real PT0-003 Exam Looks Like

Here is the format these 20 questions are modeled on. The domain weights are CompTIA’s published figures for PT0-003 [1]; the question estimates assume the full 90.

Domain Weight ~Questions (of 90)
1.0 Engagement Management13%~12
2.0 Reconnaissance and Enumeration21%~19
3.0 Vulnerability Discovery and Analysis17%~15
4.0 Attacks and Exploits35%~31
5.0 Post-exploitation and Lateral Movement14%~13
Detail PT0-003
QuestionsUp to 90 (multiple choice + PBQs)
Time165 minutes
Passing score750 on a 100–900 scale (how PenTest+ scoring works)
Voucher$439 (2026)
LaunchedDecember 17, 2024 (PT0-002 retired June 17, 2025)

If you missed more than three of the twenty above, work the domain you missed them in against the PT0-003 exam objectives before buying a voucher. If you missed one or two, you are ready for full-length timed practice: the PenTest+ practice test runs 25 exams built the same way, with per-answer explanations, and the first 3 are free.

Start a Free PenTest+ Exam

References

  1. CompTIA. "CompTIA PenTest+ (PT0-003)." comptia.org/certifications/pentest. Domain weights, 90-question maximum, 165-minute length, 750 passing score, December 17, 2024 launch and June 17, 2025 PT0-002 retirement. Accessed August 25, 2026.

These questions are original SecuSpark items written to the PT0-003 objectives. They are not, and are not meant to resemble, actual CompTIA exam content.

Stop Reading, Start Practicing

7,500+ practice questions across 7 certification paths: Security+, A+, Network+, CySA+, PenTest+, SecAI+, and Claude Architect. 3 free exams per cert to start. Campaign Pass unlocks all exams, premium AI deep dives, and the RPG battle campaign.

Start Free Practice TestEnter RPG Campaign

Free Study Tools

Port Number LookupAcronym FinderSubnet Calculator
pentest+pt0-003sample questionspractice questionsfreecomptiapenetration testing

Related Articles

Exam Info

PenTest+ Passing Score: 750 out of 900 (PT0-003) Explained

The CompTIA PenTest+ PT0-003 passing score is 750/900 — about 83%. See the 5 domain weights, how many questions you can miss, and how to prepare.

6 min readRead more
Exam Info

CompTIA Exam Cost 2026: Every Cert's Price + How to Pay Less

Security+ $439. A+ $548 (two exams). Network+ $399. Post-June-2026 prices for every CompTIA exam, plus the voucher discounts that actually cut the bill.

8 min readRead more