CSPT overview

  • 200 practice questions
  • 32 flashcards
  • 8 study guide topics
  • 40 key terms

What you'll study

  • Engagement Management, Scoping and Rules of Engagement10%
  • Reconnaissance and Information Gathering12%
  • Vulnerability Analysis and Validation12%
  • Network Exploitation and Pivoting15%
  • Web Application Exploitation18%
  • Identity and Directory Service Attack Paths13%
  • Privilege Escalation (Windows and Linux)10%
  • Reporting, Evidence Handling and Remediation Guidance10%

CSPT is an acronym used by more than one credential programme, and in the profile this page follows it refers to a practical, hands-on penetration-testing certification. The exam is aimed at security practitioners who already work with offensive tooling and want a formal benchmark of their ability to run an engagement end to end: agreeing scope and rules of engagement, performing reconnaissance, validating vulnerabilities rather than trusting scanner output, exploiting networks and web applications, moving through identity and directory-service attack paths, escalating privileges on Windows and Linux, and writing findings that a client can actually act on. Because the exam is scenario-driven, preparation is less about memorising trivia and more about repetition under realistic constraints: build a lab with a Windows domain and a small Linux estate, practise pivoting through dual-homed hosts, rehearse web exploitation against deliberately vulnerable applications, and keep notes and screenshots as you go so that reporting becomes routine rather than a last-minute scramble. Candidates should also read the awarding body's own published objectives, since the domain names and weightings shown here are provisional working assumptions and are not an official blueprint. The current logistics - question count, duration, fee, passing score and retake policy - are published by the certification body on its own website; candidates should verify them there before booking. This page is independent study material and is not affiliated with or endorsed by any awarding body.

Common mistakes

Start scanning a hostname because it is in scope by name, without resolving it first and confirming the resulting IP is inside the in-scope range - and letting the proxy or browser follow redirects/CNAMEs to cloud or third-party hosts.

Scope documents usually list DNS names and URL ranges, while tools operate on IPs and follow referrals automatically; the mismatch is invisible until someone reviews the traffic log.

What to do instead

Before active testing, resolve every in-scope name and record name-to-IP mappings with timestamps. Build explicit allowlists in the intercepting proxy and in each tool's target file, disable automatic redirect-following to out-of-scope hosts, and re-resolve periodically since CDN and cloud IPs change. When a name resolves outside the range, pause and get written confirmation before touching it.

Testing begins before the signed authorisation and RoE are returned, with the tester relying on an email thread saying 'go ahead'.

Scheduling pressure and the assumption that paperwork is a formality; the email often lacks named signatory, dates, or defined targets.

What to do instead

Hold all active traffic until a countersigned authorisation is on file that names the signing authority, the exact in-scope targets, the test window, permitted techniques, and emergency stop contacts. Keep a pre-engagement checklist and treat missing items as a hard stop rather than a note for later.

Hitting the first internal service with a full-port aggressive scan and a dozen exploit attempts on discovery, without agreement on testing intensity or an escalation threshold for critical findings.

Momentum after the first foothold; the tester optimises for findings rather than for the client's tolerance for downtime and disclosure timing.

What to do instead

Agree in advance the escalation contacts, the maximum notification delay for critical findings (for example domain admin compromise or access to regulated data), and the permitted intensity per environment. When a critical finding appears, notify within the agreed window before continuing exploitation, and log the notification time.

Skipping passive reconnaissance and going straight to a port scan of the known IP range, missing leaked credentials in public repositories, certificate-transparency subdomains, and the organisation's cloud and ASN footprint.

Active scanning produces immediate visible output, while passive collection feels like it produces nothing - the missing subdomains and credentials are invisible, so the omission never self-corrects.

What to do instead

Run a fixed passive checklist first: all DNS record types, zone transfer attempts, certificate transparency and subdomain enumeration, ASN and whois, reverse-IP lookups, public code repositories, and document metadata. Record the results as a target inventory before any active scan, and use it to expand and re-verify scope.

Copying scanner output directly into the report as confirmed vulnerabilities, including items that were never manually reproduced or that have no demonstrable security impact.

Output volume and time pressure; scanners flag anything that resembles a signature, and a tester who has not reproduced a finding cannot tell a true positive from a misconfigured page.

What to do instead

Reproduce every candidate manually with a minimal request/response pair or command transcript, record preconditions, and state what an attacker gains. Move unreproduced items to an appendix as false positives or unverified observations rather than to the findings list, and note the scanner rule and why it did not hold.

Setting up a pivot without verifying routing first - starting a scan through a single SOCKS proxy with default timeouts, forgetting routes to the second subnet, and leaving the foothold with only one callback method.

Pivoting is improvised under time pressure; a proxy that works for one host is assumed to work for a whole range, and the fragility of the foothold only becomes apparent after it dies.

What to do instead

Document the pivot chain as you build it. Verify egress through each hop against a known reachable host before scanning, tune proxy timeouts and DNS resolution explicitly, add and confirm routes for every additional subnet, and establish a redundant callback (a second listener or a stable agent) before launching broad scans through the pivot.

Testing the application only as an administrator, or only unauthenticated, and never repeating object requests as a second low-privileged user - so broken object-level authorisation and privilege-boundary issues are never exercised.

One working account is convenient, and authenticated pages return data, which feels like progress; access control failures only appear when the same request is made by someone who should be denied.

What to do instead

Enumerate all roles and capture a session for at least two privilege levels. For every object identifier observed, replay the request from each session, swap identifiers, and compare status codes and response bodies. Record which role could reach which object as a table, and treat every successful cross-role read or write as a finding.

Treating Active Directory as a list of individual attacks - running Kerberoasting, AS-REP roasting and ACL scanners blindly and reporting each as a separate low-severity finding, with no attempt to connect them into a path.

Tooling makes each technique a one-line command, so the engagement drifts into checkbox collection rather than graph analysis of who can reach what.

What to do instead

Collect AD data early and map reachability from the starting principal to tier-zero assets, then validate only the edges on a candidate path. Present the result as a chain showing each hop, the credential or permission obtained, and where a control would have broken it, with individual techniques as supporting detail rather than headline findings.

Reporting a local privilege escalation as confirmed after reading that a binary is listed as exploitable, without executing it end-to-end from a clean low-privileged session and without checking that the current user can actually invoke it.

Reference lists such as GTFOBins are consulted for the binary name only; version, build, group membership, and whether the service runs as the target account are skipped because they slow the flow.

What to do instead

Prove escalation from a fresh low-privileged shell: capture whoami/id and the applicable version or build, show the exact command or misconfiguration, capture the resulting privileged context, and note any required group membership or interactive logon. If any precondition is unverified, downgrade the finding and state the missing condition.

Leaving report writing to the final hours, producing a chronological narrative with findings that lack business impact, reproduction steps, or remediation detail, and with the critical attack chain buried among informational items.

Technical work consumes the whole window and writing is treated as an administrative afterthought; notes taken during testing are too sparse to reconstruct the context afterwards.

What to do instead

Write each finding as it is validated using a fixed template (title, affected assets, preconditions, evidence reference, impact in business terms, likelihood, remediation with a specific configuration change, references). Order by exploitability and impact, lead with an executive summary containing the attack narrative and the top three priorities, and cross-check the findings list against the scope before delivery.

Capturing screenshots and terminal output with no host, user, date or address visible, storing captured credentials and files on the compromised host, and keeping no hash or index of collected evidence.

Capture happens mid-task and the surrounding context is assumed to be memorable; the same session that produced the evidence is also the most convenient place to store it.

What to do instead

Configure shell prompts and terminal logging to include timestamp, user, hostname and working directory, and include the target address in every capture. Name files to a fixed convention, hash any collected file or credential dump at the moment of capture, keep evidence only on the tester's system in a single case folder with an index listing each item and where it came from, and note the handling and destruction of any client data at close-out.

Exam requirements can change. Confirm current details with the exam provider.

CSPT practice

200 questions. Each one lists the choices, the correct answer, and an explanation. With JavaScript, this list opens one question at a time. The checker at the top of this page saves progress in this browser when it is available.

  1. Question 1Engagement Management, Scoping and Rules of Engagement

    Which statement best describes the function of a Rules of Engagement (RoE) document?

    1. It holds the technical findings that will be delivered to the client.
    2. It sets permitted targets, methods, timing and constraints. Correct answer.
    3. It lists the commercial milestones and invoicing schedule for the project.
    4. It replaces the master services agreement for the engagement.

    Explanation

    The RoE records how testing may be carried out: which assets may be touched, which techniques are allowed, when testing may occur, and how incidents are escalated. It is agreed and signed by someone able to authorize the work before activity starts. Commercial terms live in the contract, and results live in the report.

  2. Question 2Engagement Management, Scoping and Rules of Engagement

    Before testing an external client's internet-facing systems, what does the tester most need from the client?

    1. A list of previous penetration test reports from the client
    2. Verbal confirmation from the technical point of contact
    3. Signed authorization from an authorized officer Correct answer.
    4. A copy of the client's internal information security policy

    Explanation

    Written authorization signed by an individual with authority to bind the organization is what separates authorized testing from unauthorized access. A technical contact's verbal approval may not survive legal or contractual scrutiny. The signed document should identify the assets and the permitted activity.

  3. Question 3Engagement Management, Scoping and Rules of Engagement

    Which item is normally recorded in the scope section of a penetration test engagement?

    1. The hourly rate applied to retesting work
    2. The CVSS scores of the findings discovered so far
    3. The hosts, domains and applications to be tested Correct answer.
    4. The names of the testers assigned to the project

    Explanation

    Scope enumerates the assets that may be tested, such as IP ranges, domains, applications, APIs and accounts, and normally states explicit exclusions as well. Rates, staffing and findings belong in other documents or later deliverables. Anything not listed is out of bounds.

  4. Question 4Engagement Management, Scoping and Rules of Engagement

    A client insists that all test data, credentials and findings stay confidential. Which agreement primarily governs this?

    1. The service level agreement
    2. The master services agreement
    3. The statement of work addendum
    4. The non-disclosure agreement Correct answer.

    Explanation

    A non-disclosure agreement binds the parties to protect confidential information exchanged or discovered during the work. The other documents address commercial terms, deliverables or service levels rather than confidentiality of information. NDAs are typically signed before any sensitive material changes hands.

  5. Question 5Engagement Management, Scoping and Rules of Engagement

    An RoE states that testing may occur only between 22:00 and 04:00 local time. A tester's automated scanner is still running at 05:00 and producing results. What is the correct action?

    1. Let the scan finish so that the coverage is not lost.
    2. Continue the scan and disclose the overrun in the report.
    3. Restart the scan during the next permitted window.
    4. Stop the scan and log the overrun for the client. Correct answer.

    Explanation

    The testing window is a hard constraint in the RoE, so activity must cease when the window closes and any overrun must be recorded and disclosed promptly. Continuing past the window risks disrupting systems during business hours that the client deliberately excluded. Stopping and reporting preserves both the agreement and the evidence chain.

  6. Question 6Engagement Management, Scoping and Rules of Engagement

    A client asks you to test a web application hosted on a public cloud provider's infrastructure. What must be established first?

    1. That the cloud provider offers a bug bounty program
    2. That the provider's policy permits the activity Correct answer.
    3. That the client's DNS records are managed internally
    4. That the application uses a content delivery network

    Explanation

    Where assets are hosted by a third party, the provider's acceptable-use and testing policies still apply and may prohibit certain tests or require advance notification. The client's authorization does not override the provider's terms. This should be confirmed before any traffic is sent.

  7. Question 7Engagement Management, Scoping and Rules of Engagement

    During testing, a production database becomes unresponsive and the client's service degrades. What should the tester do first?

    1. Finish the current request and then check the logs
    2. Wait for the client to notice and contact the team
    3. Suspend testing and notify the emergency contact Correct answer.
    4. Roll back the database to the previous night's backup

    Explanation

    The RoE defines stop conditions that require the tester to halt activity immediately and use the agreed escalation channel so the client can respond. Testers should not attempt to remediate, restore or reconfigure client systems themselves. Prompt notification limits the impact and preserves the client's own recovery procedures.

  8. Question 8Engagement Management, Scoping and Rules of Engagement

    The client's SOC is receiving alerts from your scanning. What should have been agreed in advance to prevent wasted incident response effort?

    1. A signed waiver of the client's incident response duties
    2. A daily schedule of findings to be reported
    3. A copy of the tester's liability insurance certificate
    4. A list of tester source IPs and a deconfliction contact Correct answer.

    Explanation

    Deconfliction arrangements, typically known source addresses and named contacts on both sides, let the SOC distinguish tester traffic from a genuine intrusion. Without them the blue team may block the engagement or spend effort chasing authorized activity. These details belong in the RoE before testing begins.

  9. Question 9Engagement Management, Scoping and Rules of Engagement

    While testing, you extract a database table containing real customer records. What is the appropriate handling?

    1. Move it to your personal cloud drive for safekeeping
    2. Share it with the client's marketing team for analysis
    3. Keep a full copy in the report appendix as proof of impact
    4. Store it encrypted and delete it per the agreement Correct answer.

    Explanation

    Data encountered during testing remains the client's, and the RoE or contract should specify how it is protected, minimized and destroyed. Real personal data must not be redistributed or retained beyond the agreed period. Sufficient redacted evidence should be captured instead.

  10. Question 10Engagement Management, Scoping and Rules of Engagement

    Mid-engagement the client asks you to also test a newly acquired subsidiary's systems. What is the correct response?

    1. Obtain written authorization and updated scope first Correct answer.
    2. Begin testing immediately because the client asked for it
    3. Refuse, because scope cannot be altered once signed
    4. Test only the subsidiary's public website and note it

    Explanation

    Scope changes need the same authorization as the original engagement, because the subsidiary is a separate legal entity with its own assets and sign-off. The RoE and statement of work should be amended before any testing starts. Informal requests do not extend authorization.

  11. Question 11Engagement Management, Scoping and Rules of Engagement

    The client supplies privileged test accounts for the engagement. Which practice is appropriate?

    1. Store the credentials in the shared team chat for reference
    2. Reuse the same password across all test accounts for speed
    3. Use them only within scope and rotate afterwards Correct answer.
    4. Email the credentials to the whole team for convenience

    Explanation

    Credentials provided by the client are part of the authorized access path and must be protected and confined to the agreed targets. Rotation and revocation at engagement close prevent residual access. Handling rules should be written into the RoE.

  12. Question 12Engagement Management, Scoping and Rules of Engagement

    The RoE prohibits denial-of-service testing. A tester identifies a potentially crashing endpoint. What is the correct approach?

    1. Ask a colleague to test it without telling the client
    2. Send a single crafted request and observe quietly
    3. Run a small load test to confirm the crash remotely
    4. Report the suspicion without triggering the condition Correct answer.

    Explanation

    Techniques excluded by the RoE stay out of bounds even when a vulnerability looks promising, because availability impact was never authorized. The issue should be reported as a suspected finding with a recommendation for controlled verification. The client can then decide whether to widen scope.

  13. Question 13Engagement Management, Scoping and Rules of Engagement

    A client says the goal is 'to know if we are secure'. Which refinement best supports a measurable engagement?

    1. Extend the testing window by two additional weeks
    2. Increase the number of findings required in the final report
    3. Agree specific and measurable objectives with the client Correct answer.
    4. Add a second tester to the delivery team for coverage

    Explanation

    Objectives should be specific and verifiable so that coverage and success can be judged at the end of the engagement. Adding time, people or a findings quota does not make the underlying goal measurable. Clear objectives also make it easier to scope the assets and techniques required.

  14. Question 14Engagement Management, Scoping and Rules of Engagement

    You discover that an in-scope server stores credentials for an out-of-scope system. What should you do?

    1. Use the credentials to prove the out-of-scope risk
    2. Record it and raise it through the agreed channel Correct answer.
    3. Test the out-of-scope system with read-only actions
    4. Quietly ignore it, as it is outside the scope

    Explanation

    Findings that touch out-of-scope assets are reported, not exploited, because authorization does not extend to those systems. The RoE escalation path lets the client decide whether to expand scope. Documenting the observation preserves value without exceeding authority.

  15. Question 15Engagement Management, Scoping and Rules of Engagement

    How does a statement of work (SOW) differ from rules of engagement (RoE)?

    1. The SOW lists techniques; the RoE lists pricing
    2. The SOW is signed by testers and the RoE by the client
    3. The SOW covers deliverables; the RoE covers conduct Correct answer.
    4. The SOW is optional and the RoE is legally binding

    Explanation

    The statement of work is the commercial document describing services, deliverables and timelines, while the RoE describes how testing will be conducted. Both should be agreed and signed before testing starts. They complement each other rather than duplicating content.

  16. Question 16Engagement Management, Scoping and Rules of Engagement

    What is the main purpose of a kickoff meeting before testing begins?

    1. To demonstrate exploits to the client's developers
    2. To negotiate the final price of the engagement
    3. To confirm contacts, scope, timing and escalation Correct answer.
    4. To deliver the draft report to the client's executives

    Explanation

    A kickoff aligns both sides on scope, schedule, key contacts and escalation routes before any traffic is sent. It reduces the chance of scope disputes and missed incident notifications. Commercial negotiation and reporting happen at other stages.

  17. Question 17Engagement Management, Scoping and Rules of Engagement

    Why does an engagement contract typically include indemnification and liability clauses?

    1. To allocate responsibility for unintended damage Correct answer.
    2. To fix the number of testers deployed on the project
    3. To guarantee that no vulnerabilities will be missed
    4. To ensure findings remain confidential between the parties

    Explanation

    Penetration testing carries a risk of disruption, so contracts allocate financial responsibility and set liability caps for unintended harm. These clauses are negotiated legal terms, not technical promises about outcomes. Confidentiality and staffing are handled by separate provisions.

  18. Question 18Engagement Management, Scoping and Rules of Engagement

    You plan to bring a subcontractor onto the engagement. What must happen before they touch client systems?

    1. They should sign a personal promise to work carefully
    2. They should receive the scope spreadsheet from the lead
    3. They should be added to the client's internal chat channels
    4. They must be bound by the same agreements and authorized Correct answer.

    Explanation

    Subcontractors need their own confidentiality and authorization coverage, and the client must agree to their involvement, since authorization extends only to named parties. Handing over scope documents does not confer authorization. This should be resolved contractually before any testing activity.

  19. Question 19Engagement Management, Scoping and Rules of Engagement

    Your team is in one country and the client's servers are in another. Why does this matter for authorization?

    1. Foreign systems can be tested under the tester's home law
    2. The client's written authorization applies worldwide
    3. Only the country of the tester's company has jurisdiction
    4. Other jurisdictions may impose additional legal limits Correct answer.

    Explanation

    Computer-misuse and data-protection laws differ between countries, so the location of the tester, the client and the assets can each impose obligations. Legal review should confirm that the authorization satisfies every relevant jurisdiction. A single signed letter may not cover all of them.

  20. Question 20Engagement Management, Scoping and Rules of Engagement

    The client asks for a retest six weeks after the report is delivered. What should the retest scope define?

    1. A full repeat of the original engagement at no cost
    2. Which findings, hosts and techniques the retest covers Correct answer.
    3. Only the critical findings, unless the client objects
    4. The new vulnerabilities found by the client's own team

    Explanation

    Retest scope should state exactly which findings and assets will be re-verified, along with any techniques and constraints, so effort and evidence are bounded. It is not automatically a repetition of the whole original engagement. Agreeing this in advance prevents disputes over coverage.

  21. Question 21Reconnaissance and Information Gathering

    An assessor wants to discover hostnames on a target domain by looking at certificates that were issued for it. Which data source is most directly suited to this?

    1. Historical WHOIS records for the domain
    2. The target's published DNS zone file
    3. Certificate Transparency logs Correct answer.
    4. The target's robots.txt file

    Explanation

    Certificate Transparency logs are public, append-only records of every certificate a CA issues, including the subjectAltName entries. Querying them (for example through crt.sh) yields hostnames that were never linked publicly. WHOIS, zone files, and robots.txt do not enumerate certificates.

  22. Question 22Reconnaissance and Information Gathering

    What does a WHOIS lookup on a domain primarily return?

    1. The complete set of DNS records published for the domain
    2. The open ports detected on the registrant's mail server
    3. The TLS certificate chain presented by the web server
    4. Registrar, creation date, and registrant contacts Correct answer.

    Explanation

    A WHOIS query reaches the registry or registrar directory and returns registration metadata: registrar, key dates, nameservers, and any published contacts or organization details. DNS and TLS data come from separate services.

  23. Question 23Reconnaissance and Information Gathering

    A web application's robots.txt file lists several Disallow entries. What is the security-testing relevance of this file?

    1. It encrypts the listed paths so they cannot be enumerated
    2. It logs every crawler that has visited the site
    3. It enforces access control, so listed paths cannot be requested by a tester
    4. It can disclose sensitive paths the owner would rather keep unindexed Correct answer.

    Explanation

    robots.txt tells well-behaved crawlers which paths to skip; it does not enforce access control, so listed paths remain reachable. Because owners often list admin panels, staging areas, or backup directories there, it is a useful enumeration lead.

  24. Question 24Reconnaissance and Information Gathering

    Which search engine operator limits results to pages on a single specified host or domain?

    1. inurl:
    2. intitle:
    3. site: Correct answer.
    4. filetype:

    Explanation

    The site: operator restricts results to URLs on the specified host or domain, which makes it useful for enumerating indexed content belonging to a target. filetype: filters by document extension, intitle: matches page titles, and inurl: matches text inside the URL.

  25. Question 25Reconnaissance and Information Gathering

    Which DNS record type should an assessor review first when the goal is to map a target's email infrastructure?

    1. CNAME
    2. AAAA
    3. SOA
    4. MX Correct answer.

    Explanation

    MX records name the mail exchangers that accept mail for a domain, which immediately identifies the mail provider and the hosts worth examining. From there an assessor can check SPF alignment, gateway versions, and related subdomains.

  26. Question 26Reconnaissance and Information Gathering

    An assessor queries a third-party passive DNS database for hostnames belonging to a target. How is this activity classified?

    1. Passive, because no traffic is sent to the target's own systems Correct answer.
    2. Active, because the DNS queries reach the target's nameservers
    3. Active, because the database resolves names live at query time
    4. Active, because DNS is inherently an active protocol

    Explanation

    Reconnaissance is passive when the assessor never sends traffic to the target's own systems; querying a third-party database keeps the target unaware. Active reconnaissance involves direct interaction, such as resolving against the target's nameservers or scanning its hosts.

  27. Question 27Reconnaissance and Information Gathering

    A nameserver answers a dig AXFR request for the zone from an unauthorized client with the full record set. What should the assessor record as the finding?

    1. DNSSEC validation is disabled for the zone
    2. The nameserver is vulnerable to DNS cache poisoning attacks
    3. The server is running a version of BIND with known flaws
    4. The zone is transferable by unauthorized clients Correct answer.

    Explanation

    A successful AXFR from an unauthorized source means the nameserver hands the entire zone, including internal-only hostnames, to any client that asks. The finding should record the exposed record set and the nameserver responsible.

  28. Question 28Reconnaissance and Information Gathering

    During subdomain brute forcing, every random label resolves to the same IP address. What does this indicate?

    1. A wildcard DNS record is answering, so hits must be filtered Correct answer.
    2. The target publishes wildcard certificates for all names
    3. The wordlist contains only valid hostnames
    4. The resolver is returning cached answers from a previous query

    Explanation

    A wildcard record such as *.example.com matches any label, so every lookup in the brute force returns the same address. Results must be compared against that wildcard baseline before any discovered host is reported.

  29. Question 29Reconnaissance and Information Gathering

    What does an SPF TXT record for a domain specify?

    1. Which ports the mail server accepts connections on
    2. Which certificate authorities may issue certificates for the domain
    3. Which hosts are authorized to send email for the domain Correct answer.
    4. Which nameservers are authoritative for the domain

    Explanation

    SPF is published in a TXT record and lists the hosts and mechanisms authorized to send mail for the domain, helping receivers reject forgeries. It also discloses third-party senders such as marketing platforms.

  30. Question 30Reconnaissance and Information Gathering

    Which capability most distinguishes Shodan from a conventional web search engine during reconnaissance?

    1. Indexing only the pages linked from a target's home page
    2. Resolving domain names faster than public resolvers
    3. Searching internet-wide service banners by product or port Correct answer.
    4. Storing historical WHOIS records for each domain

    Explanation

    Shodan continuously scans the internet and indexes the banners services return, so it can be searched by product, port, or header without any traffic reaching the target. That makes it a passive reconnaissance resource rather than a page index.

  31. Question 31Reconnaissance and Information Gathering

    An assessor computes a hash of a target's favicon and searches for that hash in Shodan. What is the purpose of this step?

    1. To brute force the directory containing the favicon
    2. To verify the favicon's TLS certificate is valid
    3. To find other hosts serving the same application Correct answer.
    4. To detect whether the favicon file contains embedded malware

    Explanation

    A favicon hash acts as a fingerprint for a web application, so searching the hash returns other hosts running the same product, including staging or forgotten deployments. It is a passive way to expand the asset inventory.

  32. Question 32Reconnaissance and Information Gathering

    What is the primary purpose of running WhatWeb against a live web endpoint?

    1. To fingerprint servers, frameworks, and libraries in use Correct answer.
    2. To enumerate valid usernames through response timing
    3. To capture and replay authenticated sessions
    4. To fuzz request parameters for injection vulnerabilities

    Explanation

    WhatWeb identifies the server, frameworks, CMS, and JavaScript libraries by matching responses and headers against signatures. The result guides later testing but does not itself test for vulnerabilities.

  33. Question 33Reconnaissance and Information Gathering

    Why would an assessor query the Wayback Machine during reconnaissance?

    1. To view the source code of the server's backend
    2. To check whether the target's TLS certificate has expired
    3. To recover retired pages and endpoints no longer linked Correct answer.
    4. To obtain the target's current DNS zone data

    Explanation

    Web archives store snapshots of pages over time, so retired endpoints, old parameters, and removed comments can still be retrieved. Those artefacts often point at functionality that remains live on the current site.

  34. Question 34Reconnaissance and Information Gathering

    An assessor searches a public code-hosting site for the target's domain name. What class of exposure is being looked for?

    1. Missing DMARC records on subdomains
    2. Secrets and internal hostnames in committed code Correct answer.
    3. Open recursive DNS resolvers
    4. Unpatched TLS ciphers on the target's web servers

    Explanation

    Public repositories frequently contain secrets, internal hostnames, and configuration committed by developers, and a domain-based search surfaces them. Any discovered credential must be handled according to the engagement's rules.

  35. Question 35Reconnaissance and Information Gathering

    A subdomain's CNAME points to a cloud service hostname that now returns NXDOMAIN. What is the most likely security implication?

    1. An attacker may claim the resource and host content on that name Correct answer.
    2. The record will be automatically removed after the TTL expires
    3. The subdomain is protected by DNSSEC and cannot be hijacked
    4. The parent domain's registrar has expired the registration

    Explanation

    When a CNAME points at a resource that no longer exists, whoever registers that resource controls the content served under the trusted subdomain. This is the classic dangling-record subdomain takeover pattern.

  36. Question 36Reconnaissance and Information Gathering

    Which query best identifies the IP ranges an organization owns when the goal is to scope a perimeter assessment?

    1. A lookup of the organization's ASN and its announced prefixes Correct answer.
    2. A PTR lookup against each address of the mail server
    3. An NS lookup to identify the authoritative nameservers
    4. A TXT query against the organization's primary domain

    Explanation

    Every announced prefix belongs to the ASN that originates it, so looking up the organization's ASN and its announced prefixes gives the address space to scope. Individual PTR, TXT, or NS lookups reveal only isolated names.

  37. Question 37Reconnaissance and Information Gathering

    What does the Nmap -sn option perform during host discovery?

    1. SYN scan of the thousand most common TCP ports
    2. Ping scan that identifies live hosts without port scanning Correct answer.
    3. OS fingerprinting based on TCP/IP stack behaviour
    4. UDP scan of well-known service ports

    Explanation

    The -sn option performs host discovery only, using probes such as ICMP echo, TCP SYN to 443, and ARP on local networks, then reports which hosts are up without scanning ports. Port scanning requires different options.

  38. Question 38Reconnaissance and Information Gathering

    Why is reverse DNS considered weak evidence when mapping an IP address to an organization?

    1. PTR records are encrypted and cannot be read by public resolvers
    2. PTR records are generated automatically from WHOIS contact data
    3. The address owner sets them, and they may be stale or generic Correct answer.
    4. PTR records exist only for IPv6 reverse zones

    Explanation

    PTR records are published by whoever controls the address block, not by the organization using the hostname, and they are often left as generic provider names or outdated entries. They are a lead to corroborate with forward lookups and routing data.

  39. Question 39Reconnaissance and Information Gathering

    What does a passive DNS database such as SecurityTrails or VirusTotal provide?

    1. Live zone transfers performed against authoritative servers
    2. The private keys used by a domain's TLS certificates
    3. Port scan results for every host in a domain
    4. Historical name-to-address resolutions collected over time Correct answer.

    Explanation

    Passive DNS providers collect and retain responses observed across the internet, showing which addresses a name resolved to in the past and which names shared an address. This history exposes infrastructure that has since been moved.

  40. Question 40Reconnaissance and Information Gathering

    An assessor wants URL data for a target that predates any current crawling. Which approach best fits?

    1. Querying a public web-archive corpus such as Common Crawl Correct answer.
    2. Running a directory brute force with a large wordlist
    3. Querying the Shodan API for open ports on the target
    4. Enumerating subdomains through a certificate transparency tool

    Explanation

    Common Crawl and similar corpora periodically archive large portions of the web, and their URL indexes can be queried for a domain to recover pages and paths from earlier periods. The technique is passive and sends nothing to the target.

  41. Question 41Reconnaissance and Information Gathering

    Two apparently unrelated sites share the same Google Analytics tracking identifier. What does this most likely reveal to an assessor?

    1. A load balancer distributing traffic between the sites
    2. A common owner or operator behind both sites Correct answer.
    3. Identical web server software versions on both hosts
    4. A shared TLS certificate covering both hostnames

    Explanation

    A shared tracking identifier usually means one party operates or administers both sites, so it is a strong pivot for finding related infrastructure. It is a correlation lead that should be confirmed rather than treated as proof.

  42. Question 42Reconnaissance and Information Gathering

    A domain's TXT records include verification strings for several SaaS providers. What is the reconnaissance value of this observation?

    1. They prove the providers are configured securely
    2. They reveal third-party services the organization uses Correct answer.
    3. They disclose the internal IP addresses of the providers
    4. They list the mail servers authorized to send email

    Explanation

    Host-verification TXT records are issued per provider, so their presence shows which SaaS platforms, gateways, or collaboration tools the organization has onboarded. Each one expands the third-party attack surface worth reviewing.

  43. Question 43Reconnaissance and Information Gathering

    Which approach is best suited to discovering subdomains that use names not present in any public source?

    1. Reviewing historical WHOIS records
    2. Querying certificate transparency logs for the domain
    3. Searching a passive DNS database
    4. Brute forcing labels from a curated wordlist Correct answer.

    Explanation

    Labels that no public source has ever recorded must be guessed, and a curated wordlist of common development, staging, and service names finds them. Public sources such as certificate logs, passive DNS, and WHOIS only reveal names that already leaked.

  44. Question 44Reconnaissance and Information Gathering

    JavaScript loaded from the target's site references several storage bucket names. What is the most appropriate follow-up during reconnaissance?

    1. Download their contents on the assumption they are public
    2. Conclude the buckets belong to a third party and ignore them
    3. Report a vulnerability immediately without further checking
    4. Verify ownership and access settings within scope Correct answer.

    Explanation

    Bucket names in client-side code are a lead to check, but confirming ownership and testing access settings must stay within the agreed scope and rules of engagement. Downloading content without authorization can exceed scope and destroy evidence.

  45. Question 45Vulnerability Analysis and Validation

    In CVSS, what does the base score group describe?

    1. the observed exploitation rate in public threat intelligence feeds
    2. traits adjusted for the organization's deployed compensating controls
    3. intrinsic traits that remain constant across time and environments Correct answer.
    4. the residual severity after the vendor releases an official patch

    Explanation

    The CVSS base metrics describe intrinsic properties of a vulnerability that are constant over time and across environments. Temporal metrics account for changing exploit or remediation conditions, and environmental metrics account for a specific organization's context.

  46. Question 46Vulnerability Analysis and Validation

    In CVSS v3.1, which metric captures whether an attacker must first authenticate?

    1. Privileges Required Correct answer.
    2. User Interaction
    3. Scope
    4. Attack Vector

    Explanation

    Privileges Required records the level of access or authentication an attacker needs before exploiting the flaw. Attack Vector describes the path of attack, User Interaction records whether a victim must act, and Scope records whether impact crosses a security boundary.

  47. Question 47Vulnerability Analysis and Validation

    What is a false positive in vulnerability analysis?

    1. A real flaw that the scanner failed to report at all
    2. A flaw scored lower than the organization's risk threshold
    3. A reported flaw that cannot be reproduced or has no impact Correct answer.
    4. A finding that duplicates another team's open ticket

    Explanation

    A false positive is a finding that the tool or analyst reports as a vulnerability but that does not actually exist or cannot be shown to have security impact. False negatives are the opposite: real flaws that are missed.

  48. Question 48Vulnerability Analysis and Validation

    What is the main purpose of a proof of concept during vulnerability validation?

    1. It confirms that the vendor has released an official patch for the flaw
    2. It assigns a unique CVE identifier for public tracking of the flaw
    3. It proves the flaw can actually be triggered and has impact Correct answer.
    4. It calculates the exact CVSS base score for the finding

    Explanation

    A proof of concept demonstrates that a suspected flaw can be triggered under realistic conditions and produces the claimed security impact. This helps separate real vulnerabilities from scanner noise and supports accurate severity assessment.

  49. Question 49Vulnerability Analysis and Validation

    What does CWE classify?

    1. Organizational risk tolerance levels
    2. Public exploit kits sold on criminal markets
    3. Classes of software weaknesses Correct answer.
    4. Individual vulnerability instances reported by vendors

    Explanation

    CWE is a taxonomy of software weakness types, such as injection or buffer overflow. Individual vulnerability instances are identified by CVEs, while exploit kits and risk tolerance levels are separate concepts.

  50. Question 50Vulnerability Analysis and Validation

    What is the first appropriate step when a new scanner finding appears?

    1. Close the report as a false positive without further review
    2. Publish a public advisory describing the suspected flaw
    3. Immediately notify every customer listed in the asset inventory
    4. Reproduce it in a controlled, authorized test environment Correct answer.

    Explanation

    Validation begins by safely reproducing the finding within the authorized scope so the team can confirm exploitability and impact. Notification, closure, and disclosure decisions should follow only after validation and triage.

  51. Question 51Vulnerability Analysis and Validation

    What does reachability analysis help determine in software composition analysis?

    1. Whether the vulnerable function is actually called at runtime Correct answer.
    2. Whether the dependency has a newer version available upstream
    3. Whether the dependency license conflicts with company policy
    4. Whether the dependency appears in the project's manifest file

    Explanation

    Reachability analysis examines whether the vulnerable code path in a dependency can actually be invoked by the application. A vulnerable library that is present but whose affected functions are never used may pose lower practical risk.

  52. Question 52Vulnerability Analysis and Validation

    In CVSS v3.1, what does the temporal metric Exploit Code Maturity describe?

    1. The organization's specific security requirements for the asset
    2. The remediation level such as official fix or workaround
    3. The attack complexity required to trigger the flaw
    4. The availability of exploit code or working techniques Correct answer.

    Explanation

    Exploit Code Maturity reflects the current state of exploit techniques or code availability for a vulnerability. It is a temporal metric because exploit availability can change over time.

  53. Question 53Vulnerability Analysis and Validation

    In CVSS, which Attack Vector value indicates the attacker needs local access?

    1. Adjacent
    2. Physical
    3. Local Correct answer.
    4. Network

    Explanation

    The Local attack vector means the attacker must have local access, such as a shell or console session, to exploit the vulnerability. Network, Adjacent, and Physical describe other reachability paths.

  54. Question 54Vulnerability Analysis and Validation

    Which statement correctly distinguishes SAST from DAST?

    1. DAST reviews source code only after deployment to production
    2. DAST inspects source code without executing the application
    3. SAST attacks a running application through its external interfaces
    4. SAST analyzes source or bytecode without running the program Correct answer.

    Explanation

    SAST examines source code, bytecode, or binaries without executing the application. DAST tests a running application from the outside, typically through its exposed interfaces.

  55. Question 55Vulnerability Analysis and Validation

    What does an EPSS score estimate?

    1. The severity of the flaw after environmental adjustments
    2. The vendor's self-assessed patch release schedule
    3. Probability of exploitation in the wild within 30 days Correct answer.
    4. The number of products affected by the vulnerability

    Explanation

    EPSS provides a probability estimate that a vulnerability will be exploited in the wild within the next 30 days. It is useful alongside severity scoring because it reflects exploitation likelihood rather than intrinsic impact alone.

  56. Question 56Vulnerability Analysis and Validation

    How is the CISA KEV catalog used in vulnerability prioritization?

    1. It provides source code patches for affected open source projects
    2. It catalogs all weaknesses in the CWE taxonomy
    3. It lists vulnerabilities known to be exploited in the wild Correct answer.
    4. It scores vulnerabilities using the CVSS environmental metrics

    Explanation

    The Known Exploited Vulnerabilities catalog identifies vulnerabilities that have been observed exploited in the wild. Presence in KEV is a strong signal for urgent remediation prioritization.

  57. Question 57Vulnerability Analysis and Validation

    In vulnerability triage, what best describes a duplicate report?

    1. A report covering the same root cause as an earlier one Correct answer.
    2. A report that affects a staging environment
    3. A report submitted after the program's deadline
    4. A report that includes a working proof of concept

    Explanation

    A duplicate report describes the same underlying root cause as a report already received or accepted. Triage should deduplicate by root cause rather than by superficial symptom or endpoint.

  58. Question 58Vulnerability Analysis and Validation

    Which inputs can change a CVSS environmental score?

    1. Modified base metrics plus the organization's requirements Correct answer.
    2. Only the passage of time since the vulnerability was published
    3. Only the attack vector and user interaction metrics
    4. Only the vendor's patch release date and advisory wording

    Explanation

    The environmental score adjusts the base and temporal metrics using modified base metrics and security requirements that reflect the organization's context. This allows the same vulnerability to receive different environmental scores across deployments.

  59. Question 59Vulnerability Analysis and Validation

    What does a VEX document communicate?

    1. The CWE identifier assigned to a weakness class
    2. Whether a product is affected by a given vulnerability Correct answer.
    3. The exact CVSS vector string for a vulnerability
    4. The exploit price observed on criminal marketplaces

    Explanation

    VEX, or Vulnerability Exploitability eXchange, communicates a vendor's or supplier's status about whether a product is affected by a vulnerability. Statuses commonly include affected, not affected, fixed, and under investigation.

  60. Question 60Vulnerability Analysis and Validation

    Why can chained vulnerabilities be more serious than each finding alone?

    1. Low-severity flaws can combine into a high-impact attack Correct answer.
    2. Each flaw must be scored independently with no interaction
    3. Only network-reachable flaws can be chained together
    4. Chaining always increases the CVSS base score of each flaw

    Explanation

    Attackers often combine multiple lower-severity weaknesses so that one flaw enables another, producing impact greater than any single finding. Validation should therefore consider plausible exploit chains rather than only isolated reports.

  61. Question 61Vulnerability Analysis and Validation

    What does taint tracking in static analysis do?

    1. It compares dependency versions against a vulnerability database
    2. It reviews compiled binaries for embedded secrets
    3. It tracks untrusted input from sources toward sinks Correct answer.
    4. It executes the application and monitors network traffic

    Explanation

    Taint tracking follows data from untrusted sources, such as request parameters, to sensitive sinks, such as SQL execution or command invocation. This helps identify injection-style flaws along data-flow paths.

  62. Question 62Vulnerability Analysis and Validation

    What is a key limitation of dynamic application security testing?

    1. It cannot detect injection flaws in running services
    2. It requires access to full source code to run at all
    3. It only covers code paths that the tests actually execute Correct answer.
    4. It can prove that an application contains no vulnerabilities

    Explanation

    DAST observes only the behavior exercised by the requests and workflows it sends, so unexercised code paths are not tested. It therefore cannot prove the absence of vulnerabilities and should be combined with other techniques.

  63. Question 63Vulnerability Analysis and Validation

    When validating a potentially destructive vulnerability, what is the safest approach?

    1. Run destructive exploit steps against the production database
    2. Skip validation because the scanner has a low false-positive rate
    3. Share the exploit publicly to confirm it works for others
    4. Use an isolated test environment within the authorized scope Correct answer.

    Explanation

    Validation should occur in an isolated environment that is explicitly within the authorized scope, using the least destructive proof necessary. This confirms impact while avoiding production damage and unauthorized testing.

  64. Question 64Vulnerability Analysis and Validation

    In CVSS v3.1, what does a Scope change indicate?

    1. The CVSS score changes when environmental metrics are applied
    2. Impact crosses a security scope boundary of the component Correct answer.
    3. The attacker must change scope settings in the application
    4. The vulnerability exists only in a shared hosting environment

    Explanation

    A Scope change means the vulnerable component can impact resources beyond its own security scope, such as the host operating system or other components. It captures a boundary crossing in the impact of exploitation.

  65. Question 65Vulnerability Analysis and Validation

    How should exploitability and impact be weighed during remediation prioritization?

    1. Exploitability and impact both inform remediation priority Correct answer.
    2. Impact alone determines priority regardless of exploitability
    3. CVSS base score is the only factor worth considering
    4. Exploitability alone determines priority regardless of impact

    Explanation

    Risk prioritization considers both how likely and how easily a flaw can be exploited and what impact successful exploitation would have. A high-impact but impractical flaw may rank differently from an easily exploited flaw with moderate impact.

  66. Question 66Vulnerability Analysis and Validation

    Why should multiple validation techniques be combined during vulnerability analysis?

    1. Manual review replaces all automated scanning activities
    2. False negatives are harmless because they are never reported
    3. A single commercial scanner is sufficient for full coverage
    4. No single tool proves absence, so combine techniques Correct answer.

    Explanation

    Each technique has blind spots: SAST may miss runtime configuration issues, DAST may miss unexercised paths, and SCA may flag unreachable code. Combining techniques improves coverage and reduces both false positives and false negatives.

  67. Question 67Vulnerability Analysis and Validation

    Given the CVSS v3.1 vector AV:N/AC:L/PR:N/UI:N, what does it indicate?

    1. Physical access with no privileges and no interaction
    2. Adjacent network with low privileges and no interaction
    3. Network reachable, no privileges, no user interaction Correct answer.
    4. Local access with high privileges and user interaction

    Explanation

    AV:N means network reachable, PR:N means no privileges are required, and UI:N means no user interaction is required. AC:L indicates low attack complexity.

  68. Question 68Vulnerability Analysis and Validation

    What does a CVE identifier represent?

    1. A specific publicly disclosed vulnerability instance Correct answer.
    2. A scoring rubric used to rank vulnerabilities
    3. A class of software weakness such as buffer overflow
    4. A vendor advisory describing multiple flaws

    Explanation

    A CVE ID identifies a specific publicly disclosed vulnerability instance, enabling consistent tracking across tools and vendors. Weakness classes are represented by CWE, and scoring is handled by systems such as CVSS.

  69. Question 69Network Exploitation and Pivoting

    An attacker runs: ssh -L 8080:10.10.5.20:80 [email protected]. What does this command accomplish?

    1. It forwards port 8080 on the SSH server back to the attacker's port 80
    2. It opens a SOCKS proxy on port 8080 of the SSH server
    3. It forwards local port 8080 to 10.10.5.20:80 via the SSH server Correct answer.
    4. It binds a shell on 10.10.5.20 port 8080 for the SSH user

    Explanation

    With -L the client opens a listening socket on the local machine, and each connection is carried inside the SSH session and re-originated by the SSH server toward the destination host and port. This lets the attacker reach a service on an internal address that only the SSH server can route to.

  70. Question 70Network Exploitation and Pivoting

    What does the command ssh -D 1080 [email protected] set up on the attacker's machine?

    1. A tunnel interface that carries all IP protocols through the jump host
    2. A static forward of port 1080 to a service already running on the jump host
    3. A local SOCKS proxy on 1080 tunnelled through the jump host Correct answer.
    4. A reverse shell listener bound to the loopback interface on port 1080

    Explanation

    Dynamic forwarding (-D) makes SSH act as a SOCKS proxy on the local host; client applications send SOCKS requests and SSH opens the corresponding TCP connections from the remote side. The proxy handles TCP only and is not a general-purpose VPN.

  71. Question 71Network Exploitation and Pivoting

    Which statement accurately describes a bind shell?

    1. Both endpoints connect to a shared relay to exchange commands
    2. The attacker's host relays traffic between two internal networks
    3. The victim listens on a port that the attacker then connects to Correct answer.
    4. The victim host initiates a connection back to the attacker's listener

    Explanation

    A bind shell makes the target open a listening socket, so exploitation only succeeds if the attacker can reach that port across intervening firewalls and NAT. This makes it fragile in segmented networks compared to a reverse shell.

  72. Question 72Network Exploitation and Pivoting

    In proxychains.conf, which chain type tries each listed proxy in order and skips entries that are not responding?

    1. strict_chain
    2. random_chain
    3. dynamic_chain Correct answer.
    4. round_robin_chain

    Explanation

    dynamic_chain keeps the ordering of the proxy list but tolerates dead proxies by trying the next entry, which suits unstable pivot chains. strict_chain requires every proxy to be alive because all of them are used in series.

  73. Question 73Network Exploitation and Pivoting

    In network exploitation, pivoting is best described as which activity?

    1. Using a compromised host to reach otherwise unreachable systems Correct answer.
    2. Converting a bind shell into a reverse shell on the same host
    3. Rotating source addresses while attacking an authentication service
    4. Escalating privileges on a host to obtain administrator access

    Explanation

    Pivoting uses a foothold host as a relay so that traffic originates from inside a protected segment, bypassing perimeter filtering. The pivot host must be dual-homed or otherwise able to route to the deeper network.

  74. Question 74Network Exploitation and Pivoting

    The chisel tool is most commonly used for which purpose during a pivot?

    1. Enumerating directory objects over LDAP as a domain user
    2. Performing ARP cache poisoning on a switched network
    3. Cracking hashed credentials captured on an internal segment
    4. Tunnelling TCP traffic over HTTP between two hosts Correct answer.

    Explanation

    Chisel wraps a multiplexed TCP stream inside HTTP, so port forwards and SOCKS proxies survive proxies and egress filters that only permit web traffic. It uses a client and server binary pair and does not require SSH on either side.

  75. Question 75Network Exploitation and Pivoting

    An attacker on their own machine runs: ssh -R 9001:localhost:3306 user@pivot. What is the effect?

    1. A SOCKS proxy is created on port 9001 of the pivot host
    2. The pivot's port 3306 is exposed to the attacker on port 9001
    3. The pivot listens on 9001 and forwards to the attacker's localhost:3306 Correct answer.
    4. The attacker's port 3306 connects out to the pivot's port 9001

    Explanation

    Remote forwarding (-R) creates the listening socket on the SSH server and delivers each accepted connection back to the client, which then connects to the specified destination from its own side. It is the standard way to expose attacker-side tooling or services to an internal host.

  76. Question 76Network Exploitation and Pivoting

    Which Nmap scan type can be used when traffic must pass through a SOCKS proxy?

    1. -sT Correct answer.
    2. -O
    3. -sA
    4. -sS

    Explanation

    A SOCKS proxy relays completed TCP connections, so only the connect() scan (-sT) can be proxied. Scans that craft raw packets and read responses at the IP layer cannot traverse a proxy.

  77. Question 77Network Exploitation and Pivoting

    On a compromised dual-homed Linux host, an attacker runs: socat TCP-LISTEN:8080,fork TCP:10.10.5.20:80. What is the result?

    1. An encrypted tunnel is established between port 8080 and port 80
    2. Traffic on port 8080 is mirrored to both internal and external hosts
    3. Connections to port 8080 are relayed to 10.10.5.20:80 Correct answer.
    4. A shell is bound on 10.10.5.20:80 and exposed on port 8080

    Explanation

    socat with TCP-LISTEN and a TCP destination acts as a simple layer-4 relay, and the fork option keeps the listener available for additional clients. This creates a listening port on the pivot that reaches a service the pivot can route to.

  78. Question 78Network Exploitation and Pivoting

    Which capability does SOCKS5 provide that SOCKS4 does not?

    1. Encryption of tunnelled payloads
    2. UDP relay and user/password auth Correct answer.
    3. Support for IPv4 addressing
    4. Relaying of TCP connections

    Explanation

    SOCKS5 adds a negotiation phase for authentication methods, supports UDP association and IPv6 addresses, and allows the proxy to resolve destination names. It still does not encrypt the tunnelled data by itself.

  79. Question 79Network Exploitation and Pivoting

    An attacker who already has a foothold on a VLAN wants to intercept traffic between a workstation and its default gateway. Which technique is directly applicable?

    1. Poisoning the ARP caches of both endpoints Correct answer.
    2. Flooding the CAM table to force the switch into hub mode
    3. Sending spoofed ICMP redirects to the router's upstream peers
    4. Double tagging Ethernet frames with two 802.1Q headers

    Explanation

    ARP has no authentication, so forged replies can convince each endpoint that the attacker's MAC owns the peer's IP address, placing the attacker in the path. Forwarding must then be enabled so the interception stays transparent.

  80. Question 80Network Exploitation and Pivoting

    A switch port on the target network is left in its default dynamic auto-negotiation mode. Which attack does this configuration enable?

    1. DHCP starvation to exhaust the address pool
    2. MAC flooding to overflow the switch's CAM table
    3. VLAN hopping by negotiating a trunk with the switch Correct answer.
    4. ARP spoofing confined to the attacker's own VLAN

    Explanation

    Ports that still negotiate trunking can be convinced to become an 802.1Q trunk, which lets the attacker tag frames for other VLANs. Hardening requires disabling dynamic trunk negotiation and setting access ports explicitly.

  81. Question 81Network Exploitation and Pivoting

    Which record type is most commonly abused to carry stolen data out of a network through DNS?

    1. TXT queries to an attacker-controlled domain Correct answer.
    2. Zone transfer requests against the internal DNS server
    3. MX records of the target's mail domain
    4. PTR lookups against the internal resolver

    Explanation

    TXT records hold arbitrary text and are frequently combined with long, encoded subdomain labels to return data in the response. Because DNS egress is usually permitted, such channels often survive firewall rules that block other outbound traffic.

  82. Question 82Network Exploitation and Pivoting

    How does an ICMP tunnel carry data across a filtered network?

    1. By using ICMP type 3 messages to carry a TCP handshake
    2. By redirecting ICMP to a TCP port via a firewall rule
    3. By sending oversized ping packets that overwrite the target's memory
    4. By encoding data in the payload of echo request and reply packets Correct answer.

    Explanation

    ICMP echo packets can carry arbitrary payload bytes, so a client and server pair can multiplex a data channel inside them. This works where outbound ping is allowed but TCP and UDP are restricted.

  83. Question 83Network Exploitation and Pivoting

    Which Windows built-in feature creates a TCP listener on the pivot that relays to an internal host and port?

    1. net session
    2. netsh interface portproxy Correct answer.
    3. sc create portfwd
    4. route add

    Explanation

    The netsh portproxy subcommand creates v4tov4 listening and forwarding rules that the host applies in the kernel, requiring no extra binaries. It forwards TCP only and is configured with listenport, connectaddress and connectport parameters.

  84. Question 84Network Exploitation and Pivoting

    How does sshuttle differ from ssh -D when pivoting into an internal network?

    1. It runs only on Windows hosts with WinRM enabled
    2. It provides a layer-2 tunnel that carries broadcast traffic
    3. It transparently routes TCP and DNS into the remote network Correct answer.
    4. It requires every client application to be SOCKS-aware

    Explanation

    sshuttle installs local routing and interception rules so applications need no proxy configuration, then ships the TCP streams (and optionally DNS queries) over an SSH connection. It does not carry UDP or non-TCP protocols.

  85. Question 85Network Exploitation and Pivoting

    Why do operators often enable proxy_dns in proxychains.conf?

    1. To block DNS traffic from the target network entirely
    2. To allow proxychains to use UDP-based protocols over SOCKS4
    3. To resolve the proxy hostnames faster using a local cache
    4. To resolve names through the proxy rather than locally Correct answer.

    Explanation

    If names are resolved locally, the attacker's resolver sees queries for internal hostnames and the internal network sees the attacker's DNS traffic, which leaks intent and can break resolution. proxy_dns forces resolution to happen at the far end of the chain.

  86. Question 86Network Exploitation and Pivoting

    Inside a Meterpreter session on a dual-homed host, which command exposes the internal host 10.10.5.20:3389 on the attacker's loopback port 3389?

    1. portfwd add -l 3389 -p 3389 -r 10.10.5.20 Correct answer.
    2. route add 10.10.5.20 255.255.255.255 1
    3. autoroute add 10.10.5.0/24 -s 1
    4. socks add 3389 10.10.5.20

    Explanation

    The portfwd command creates a listening socket on the attacker machine and relays each connection through the session to the specified remote host and port. It handles a single host and port, whereas routing commands handle whole subnets.

  87. Question 87Network Exploitation and Pivoting

    In msfconsole, what does route add 10.10.5.0 255.255.255.0 1 do?

    1. Forwards port 10.10.5.0 to the listener associated with session 1
    2. Starts a SOCKS server bound to the 10.10.5.0 subnet
    3. Sends traffic destined for 10.10.5.0/24 through session 1 Correct answer.
    4. Adds a persistent route to the attacker's operating system routing table

    Explanation

    The route command adds an entry to the framework's internal routing table so that modules targeting that subnet send their packets through the chosen session. Tools that bypass the framework, such as a normal TCP tool run in a shell, are unaffected.

  88. Question 88Network Exploitation and Pivoting

    A public web application is vulnerable to server-side request forgery. What is the most direct way to leverage it for internal network access?

    1. Use it to reach internal services and metadata endpoints Correct answer.
    2. Run an ARP spoofing attack against the DMZ switch
    3. Upload a reverse shell through the file upload endpoint
    4. Send a phishing email to internal administrators

    Explanation

    SSRF makes the server issue requests from its own network position, so the attacker can probe or interact with services that are not otherwise reachable. Cloud metadata endpoints and internal admin interfaces are the classic high-value targets.

  89. Question 89Network Exploitation and Pivoting

    A compromised host is behind an egress filter that permits outbound TCP/443 only to a whitelisted web proxy. Which approach is most likely to keep a command-and-control channel working?

    1. Send ICMP echo requests carrying the C2 payload
    2. Tunnel the C2 traffic inside HTTPS through the permitted proxy Correct answer.
    3. Start a SOCKS listener on port 1080 on the host
    4. Open a bind shell on a random high port on the host

    Explanation

    When egress is limited to a specific destination and port, only traffic shaped like that permitted protocol survives; wrapping the channel in TLS over 443 through the allowed proxy blends with normal web traffic. Inbound initiation and non-permitted protocols are blocked by the filter.

  90. Question 90Network Exploitation and Pivoting

    You gain code execution on a Windows host that has no SSH service, and you need a reverse tunnel from the target back to your Linux attack box. Which option fits best?

    1. A chisel reverse tunnel connecting out to the attack box Correct answer.
    2. Two netsh portproxy rules configured on the target
    3. An OpenVPN server component deployed onto the target
    4. ssh -R against the Windows host with GatewayPorts enabled

    Explanation

    A chisel client on the Windows host dials outward and establishes the tunnel, then exposes the requested forwards or SOCKS proxy. Tools that only need to run a small binary avoid the prerequisite of an SSH daemon on the target.

  91. Question 91Network Exploitation and Pivoting

    When running Nmap through proxychains against an internal subnet, which limitation should the tester expect?

    1. OS fingerprinting works while service detection fails
    2. SYN scans work while version detection fails
    3. UDP scans work only when proxy_dns is enabled
    4. Only TCP connect scans work; host discovery is unreliable Correct answer.

    Explanation

    A SOCKS chain relays TCP connections, so Nmap must fall back to a full connect scan with -sT and treat hosts as up with -Pn. Raw-packet techniques and ICMP-based discovery cannot traverse the proxy, which changes timing and accuracy.

  92. Question 92Network Exploitation and Pivoting

    An ssh -D SOCKS tunnel is limited to which type of traffic?

    1. TCP streams only Correct answer.
    2. Raw IP packets including ICMP
    3. UDP datagrams only
    4. TCP and UDP flows alike

    Explanation

    Dynamic forwarding carries individual TCP connections requested by the SOCKS client; there is no mechanism for datagrams or raw IP. To move UDP or ICMP across a pivot, a different tool or an IP-level tunnel is needed.

  93. Question 93Network Exploitation and Pivoting

    You add a netsh portproxy rule on a Windows pivot and it appears in 'netsh interface portproxy show v4tov4', but connections to the listening port are refused. What is the most likely cause?

    1. The IP Helper service is not running Correct answer.
    2. The Windows firewall silently drops all portproxy traffic
    3. Portproxy only supports UDP traffic on Windows
    4. The portproxy rule needs a reboot to take effect

    Explanation

    Portproxy rules are implemented through the IP Helper service, so a stopped or disabled iphlpsvc leaves the rule listed but no listener in place. Starting the service makes the configured listener active immediately without rebooting.

  94. Question 94Network Exploitation and Pivoting

    A tester compromises a host that has one interface on the DMZ segment and another on the internal server VLAN. The tester uses that host to reach servers on the internal VLAN, which are not routable from the tester's own machine. Which role does the compromised host fill?

    1. A passive tap
    2. A sinkhole
    3. A canary token
    4. A pivot host Correct answer.

    Explanation

    A pivot host is a compromised system with a foothold on a second, otherwise unreachable network; traffic is relayed through it so the attacker can operate against hosts they cannot route to directly. Its value comes from sitting on two segments at once.

  95. Question 95Network Exploitation and Pivoting

    An attacker on a switched network sends a large volume of frames, each with a different fabricated source MAC address. What is the primary effect on the switch?

    1. It converts the access port into a routed port and strips VLAN tags
    2. It resets spanning tree and elects a new root bridge for the VLAN
    3. Its CAM table fills and unknown traffic is flooded Correct answer.
    4. It forwards the frames only to ports assigned to the management VLAN

    Explanation

    Every new source MAC normally creates a CAM table entry; flooding enough of them exhausts the table's capacity. Once the table is full, the switch can no longer look up destinations and falls back to flooding unknown-unicast frames out of all ports in the VLAN, which lets the attacker capture traffic.

  96. Question 96Network Exploitation and Pivoting

    A tester with a foothold runs: ssh -L 8080:10.10.5.20:80 user@pivot. What does this achieve?

    1. A listener on the tester's localhost:8080 relays to 10.10.5.20:80 Correct answer.
    2. The tester's port 8080 is opened on the pivot so other internal hosts can reach the tester
    3. The pivot starts a SOCKS proxy on port 8080 that the tester can point tools at
    4. The pivot listens on port 8080 and forwards each connection back to the tester's port 80

    Explanation

    Local forwarding (-L) binds the listening socket on the attacker's own machine and tunnels each connection over SSH to the destination, which the pivot resolves and connects to. The internal server therefore sees the connection coming from the pivot's address.

  97. Question 97Network Exploitation and Pivoting

    An attacker on the native VLAN of an 802.1Q trunk link sends a frame carrying two tags: an outer tag for the native VLAN and an inner tag for the victim VLAN. What result should the tester expect?

    1. A bidirectional session is established with hosts on the victim VLAN
    2. A one-way injection reaches the victim VLAN; replies do not return Correct answer.
    3. The frame is dropped because the trunk strips both tags
    4. The frame is flooded to every VLAN configured on the switch

    Explanation

    The first switch strips the outer (native VLAN) tag and forwards the frame on the trunk; the second switch then reads the still-present inner tag and delivers the frame into the victim VLAN. Because the victim has no route back to the attacker's native VLAN, only one-way injection is normally achievable, making the technique useful for blind or spoofed traffic rather than interactive sessions.

  98. Question 98Network Exploitation and Pivoting

    A tester creates a remote forward with ssh -R and finds the new port is reachable only from the pivot itself, not from other hosts on the pivot's network. Which directive on the pivot's sshd_config governs whether the forward may bind a non-loopback address?

    1. AllowTcpForwarding yes
    2. PermitOpen any
    3. PermitTunnel yes
    4. GatewayPorts yes Correct answer.

    Explanation

    Remote forwards bind to the loopback interface by default; GatewayPorts controls whether the server will honour a request to bind the forwarded port to an address other than loopback. Setting it to yes (or clientspecified) lets other hosts on the pivot's network reach the forwarded port.

  99. Question 99Web Application Exploitation

    Which practice most effectively prevents SQL injection in a login form?

    1. Escaping single quotes before string concatenation
    2. Blocklisting the UNION keyword at the web server
    3. Binding user input through parameterized queries Correct answer.
    4. Hashing the submitted password with MD5 before use

    Explanation

    Parameterized queries send the query structure and the user values over separate channels, so input can never change the SQL grammar. Escaping and blocklists work on syntax patterns and are routinely defeated by alternate encodings, comments, or numeric contexts.

  100. Question 100Web Application Exploitation

    An attacker posts a script inside a comment field, and the script runs for every visitor who later views that comment page. Which category of cross-site scripting is this?

    1. Stored XSS Correct answer.
    2. Self-XSS
    3. Reflected XSS
    4. DOM-based XSS

    Explanation

    Stored XSS occurs when the payload is persisted server-side and served to other users without proper output encoding. The defining trait is that the malicious content survives the original request and affects later visitors.

  101. Question 101Web Application Exploitation

    Which mechanism most directly defends a state-changing request against cross-site request forgery?

    1. Validating the Referer header as the only check
    2. Requiring a per-session unpredictable anti-CSRF token Correct answer.
    3. Setting the session cookie with the HttpOnly flag
    4. Encrypting the session cookie value

    Explanation

    A per-session token that the attacker cannot read or predict proves the request originated from the application's own pages. Because the browser does not automatically attach the token to cross-site requests, forged requests fail validation.

  102. Question 102Web Application Exploitation

    Which input sequence is classically used to escape the web root in a path traversal attack?

    1. A null byte (%00)
    2. A hash fragment (#)
    3. A pipe character (|)
    4. A dot-dot-slash sequence (../) Correct answer.

    Explanation

    Dot-dot-slash tells the filesystem to move to the parent directory, so repeating it climbs above the intended base directory. The application then reads or writes files outside the directory it meant to restrict.

  103. Question 103Web Application Exploitation

    A logged-in user changes the URL from /invoice/1042 to /invoice/1043 and views another customer's invoice. What is the root cause?

    1. The endpoint accepts only GET requests
    2. The application lacks a Content-Security-Policy header
    3. Missing server-side authorization on the requested object Correct answer.
    4. The session cookie is not encrypted

    Explanation

    The server confirms the user is authenticated but never checks whether that user owns the requested record. This is an insecure direct object reference: the identifier is trusted as proof of entitlement.

  104. Question 104Web Application Exploitation

    What does a server-side request forgery vulnerability primarily allow an attacker to do?

    1. Forge a valid TLS certificate for the target domain
    2. Make the vulnerable server issue requests to internal systems Correct answer.
    3. Execute shell commands on the database host
    4. Read the victim browser's local storage contents

    Explanation

    SSRF abuses a server-side fetch primitive so that requests originate from the trusted server's network position. That lets the attacker reach internal services, cloud metadata endpoints, and otherwise unreachable hosts.

  105. Question 105Web Application Exploitation

    A page builds the shell command ping -c 1 $host and passes it to a shell. Which input is most likely to append a second command?

    1. example.com?x=1
    2. example.com; id Correct answer.
    3. example.com
    4. 192.168.1.1

    Explanation

    A semicolon terminates the first command and starts a new one when the string is interpreted by a shell. The other values are ordinary host or URL syntax that does not create a command separator.

  106. Question 106Web Application Exploitation

    For a classic XXE attack to succeed, the XML parser must do which of the following?

    1. Resolve external entities declared in the document Correct answer.
    2. Use UTF-8 encoding for input
    3. Reject DTDs by default
    4. Run under an unprivileged account

    Explanation

    XXE relies on the parser honoring the DOCTYPE declaration and fetching or expanding external entities. Many hardened parsers disable DTD processing entirely, which removes the attack surface.

  107. Question 107Web Application Exploitation

    An application returns an identical generic error page for all input, but a request with id=1 AND 1=1 returns content while id=1 AND 1=2 returns an empty page. Which technique is being used?

    1. Time-based blind SQL injection
    2. Error-based SQL injection
    3. Boolean-based blind SQL injection Correct answer.
    4. Union-based SQL injection

    Explanation

    The attacker infers one bit of information per request by turning a true/false condition into a visible difference in the response. No error output or timing signal is required, only a distinguishable true versus false state.

  108. Question 108Web Application Exploitation

    An application safely parameterizes every INSERT statement, but a later admin report concatenates a stored username into a dynamic query. What does this describe?

    1. Second-order SQL injection Correct answer.
    2. A SQL truncation attack against a column limit
    3. Cross-site scripting persisted in the database
    4. A false positive because input was sanitized on entry

    Explanation

    Second-order injection occurs when data is stored safely but later reused in an unsafe query, so the payload fires at a different time and place than it was submitted. Defending the entry point alone is insufficient; every query that consumes the data must also be parameterized.

  109. Question 109Web Application Exploitation

    Which client-side sink most directly enables DOM-based XSS when it receives attacker-controlled data?

    1. document.createElement('div')
    2. JSON.parse(value)
    3. localStorage.getItem('theme')
    4. element.innerHTML = value Correct answer.

    Explanation

    Assigning to innerHTML hands a string to the HTML parser, so any markup in the value is interpreted and script can execute. Sinks that treat input as text or as structured data do not build executable markup.

  110. Question 110Web Application Exploitation

    A site serves the policy default-src 'self'; script-src 'self' 'unsafe-inline'. What is the practical security impact?

    1. Inline injection still runs thanks to 'unsafe-inline' Correct answer.
    2. Any CDN script may run because default-src is 'self'
    3. Only scripts from any HTTPS origin may run
    4. Inline scripts are blocked by the 'self' directive

    Explanation

    The script-src directive overrides default-src for scripts, and 'unsafe-inline' re-permits inline script blocks and event handlers. That removes the main XSS mitigation the policy would otherwise provide.

  111. Question 111Web Application Exploitation

    A site uses cookie-based sessions and accepts POST bodies with Content-Type text/plain, with no anti-CSRF token and no Origin check. What is the primary consequence?

    1. Any website can read the victim's session cookie value
    2. Cross-site requests can perform state changes without a form submission Correct answer.
    3. The application becomes vulnerable to SQL injection
    4. Passwords are transmitted to the server in clear text

    Explanation

    Simple-request content types avoid preflight, and without a token or Origin validation the server cannot tell a forged request from a legitimate one. The browser still attaches the session cookie, so the action executes as the victim.

  112. Question 112Web Application Exploitation

    A server blocks SSRF by checking an allowlisted hostname before DNS resolution occurs. Which technique defeats that control?

    1. DNS rebinding so the allowlisted name later resolves to an internal address Correct answer.
    2. Sending the request over HTTP/2 instead of HTTP/1.1
    3. Encoding the URL twice as UTF-16
    4. Appending a fragment identifier to the URL

    Explanation

    If validation happens on the name and resolution happens afterwards, a short-TTL record can return a public address during the check and an internal address when the request is actually made. The fix is to resolve first, validate the resolved address, and pin the connection to it.

  113. Question 113Web Application Exploitation

    A server reads the alg field from a JWT header and skips signature verification when it is none. What is the impact?

    1. Tokens can no longer be replayed after expiry
    2. An attacker can forge tokens carrying arbitrary claims Correct answer.
    3. The session cookie is forced to be HttpOnly
    4. The token is encrypted but not authenticated

    Explanation

    If the algorithm is attacker-selected and none is accepted, the integrity guarantee disappears entirely. The attacker can mint a token for any user or role and the server will accept it as authentic.

  114. Question 114Web Application Exploitation

    Which condition turns an insecure deserialization bug into remote code execution?

    1. The serialized payload is sent over HTTPS
    2. Untrusted data reaches a deserializer with gadget chains Correct answer.
    3. The serializer emits a binary format
    4. The payload is stored in a database column

    Explanation

    Deserialization itself only rebuilds objects; execution requires that rebuilding trigger methods along a chain of classes already present in the application, known as a gadget chain. Without a usable chain the impact is usually limited to data tampering or denial of service.

  115. Question 115Web Application Exploitation

    An upload endpoint validates only the Content-Type header sent by the client. Which approach is most likely to bypass it?

    1. Uploading the same file twice in one request
    2. Base64-encoding the file contents
    3. Compressing the file into a .tar archive
    4. Sending a PHP file while declaring image/jpeg Correct answer.

    Explanation

    Client-supplied content types are fully attacker-controlled, so a script named with an executable extension can be submitted with an image MIME type. Robust defenses validate the extension, inspect magic bytes, rename the file, and store it outside the web root.

  116. Question 116Web Application Exploitation

    A response includes Access-Control-Allow-Origin echoing the request Origin together with Access-Control-Allow-Credentials: true. Why is this dangerous?

    1. Any origin can read authenticated responses from the site Correct answer.
    2. It forces cookies to be sent over plain HTTP only
    3. It exposes the server's private key to browsers
    4. It disables TLS for cross-origin requests

    Explanation

    Reflecting arbitrary origins while allowing credentials means any site the victim visits can issue authenticated cross-origin reads and inspect the response. The correct pattern is a strict allowlist of trusted origins, never a reflection.

  117. Question 117Web Application Exploitation

    An OAuth authorization server fails to strictly validate redirect_uri. What can an attacker do with a stolen authorization code flow?

    1. Redirect the code to their own host and exchange it for tokens Correct answer.
    2. Brute-force the client secret offline
    3. Decrypt the user's encrypted session cookie
    4. Downgrade the authorization server to HTTP/1.0

    Explanation

    Loose redirect_uri matching lets the attacker register a lookalike or path-confused URI so the authorization code is delivered to a host they control. They then redeem the code at the token endpoint before the legitimate client does.

  118. Question 118Web Application Exploitation

    HTTP request smuggling most commonly arises from which condition?

    1. Disagreement between front end and back end on message boundaries Correct answer.
    2. Session cookies issued without the Secure flag
    3. Missing HSTS headers on the front-end server
    4. Servers that log request bodies in plaintext

    Explanation

    When two servers in a chain parse Content-Length and Transfer-Encoding differently, one attacker-controlled request can be split so a second request is prefixed onto the next user's traffic. Consistent parsing, or rejecting ambiguous requests, prevents it.

  119. Question 119Web Application Exploitation

    A template engine renders {{7*7}} as 49. Which payload most directly escalates a Jinja2 server-side template injection to command execution?

    1. Submitting %00 to truncate the template name
    2. Walking __class__.__mro__ to reach os.popen Correct answer.
    3. Sending a <script> tag in the same parameter
    4. Appending -- to comment out the rest of the template

    Explanation

    Jinja2 exposes Python object internals through attributes, so traversing __class__ and __mro__ reaches classes such as subprocess or os that can spawn commands. The arithmetic evaluation already proves the input is being interpreted as template code rather than data.

  120. Question 120Web Application Exploitation

    A library deep-merges JSON.parse(body) into a configuration object without filtering the __proto__ key. What is the primary risk?

    1. The JSON parser consumes excessive CPU time
    2. Cookies are automatically attached to the outgoing request
    3. Attacker-set properties appear on all objects and can alter logic Correct answer.
    4. The merged response is stored by shared caching proxies

    Explanation

    Prototype pollution writes to Object.prototype, so every object in the process inherits the injected property. That can flip authorization checks, override configuration values, or feed a gadget that ends in code execution.

  121. Question 121Web Application Exploitation

    A CDN caches responses keyed only on the URL path, and the origin reflects the X-Forwarded-Host header into absolute links on the page. What attack results?

    1. Session fixation through an injected Set-Cookie header
    2. Reflected XSS that only affects the attacker's own browser
    3. A denial of service caused by aggressive cache eviction
    4. Web cache poisoning that serves malicious links to other users Correct answer.

    Explanation

    The unkeyed header lets the attacker place a crafted response into the shared cache, which is then served to every subsequent visitor of that path. Because links are built from the injected header, victims are steered toward attacker infrastructure.

  122. Question 122Web Application Exploitation

    A login handler runs users.find({user: req.body.user, pass: req.body.pass}) against MongoDB with a JSON body parser. Which body authenticates as admin without the password?

    1. {"user":"admin","pass":"*"}
    2. {"user":"admin","pass":"admin"}
    3. {"user":"admin","pass":{"$ne":null}} Correct answer.
    4. {"user":"admin' --","pass":"x"}

    Explanation

    Passing an operator object instead of a string makes the query match any document whose pass field is not null, which the admin record satisfies. The fix is to reject non-scalar types for credential fields before building the query.

  123. Question 123Web Application Exploitation

    Which GraphQL capability most directly helps an attacker map a target schema before crafting exploitation queries?

    1. Introspection queries returning types and fields Correct answer.
    2. Complexity-based rate limiting per field
    3. Persisted queries sent as GET requests
    4. Automatic batching of mutation operations

    Explanation

    Introspection exposes the type system, including fields, arguments, and mutations, effectively documenting the API for an attacker. Production deployments often disable introspection to raise the cost of enumeration, though the schema can still be inferred.

  124. Question 124Web Application Exploitation

    A Node.js API copies every property of an incoming JSON body onto a user record before saving, so any field the client sends is persisted. Which control most directly stops a client from setting fields such as role or isAdmin?

    1. Require a valid session cookie on the update endpoint
    2. Escape every string field with HTML entity encoding before saving
    3. Copy only fields named in an explicit allowlist Correct answer.
    4. Store the model in a table whose columns are all nullable

    Explanation

    Binding request data straight onto a model lets clients set fields the UI never exposes, which is mass assignment. The fix is to build the record from an explicit allowlist of client-writable fields, typically via a DTO or schema validation. Encoding, session checks, and column nullability do not restrict which properties are written.

  125. Question 125Web Application Exploitation

    An application must stop other sites from embedding its authenticated pages inside a frame. Which response header addresses this requirement directly?

    1. X-Frame-Options: DENY Correct answer.
    2. Referrer-Policy: no-referrer
    3. Strict-Transport-Security: max-age=31536000
    4. Content-Security-Policy: default-src 'self'

    Explanation

    Framing is controlled by X-Frame-Options or the CSP frame-ancestors directive; X-Frame-Options: DENY forbids all framing of the page. A default-src restriction governs resource loading, not who may embed the document. Referrer and HSTS policies affect URL leakage and transport security respectively.

  126. Question 126Web Application Exploitation

    A PHP script passes a user-supplied value directly to include(). Which server setting is required for that code to be exploitable as remote file inclusion rather than local file inclusion?

    1. open_basedir restricted to the web root
    2. session.use_strict_mode set to 1
    3. disable_functions containing system and exec
    4. allow_url_include set to On Correct answer.

    Explanation

    Remote file inclusion requires the runtime to fetch include targets over network wrappers, which PHP gates behind allow_url_include. Without it, attacker-supplied URLs are not fetched, limiting the bug to local files. The other settings harden the environment but do not enable remote inclusion.

  127. Question 127Web Application Exploitation

    A download endpoint removes every occurrence of the substring ../ from the requested filename once, then opens the resulting path. Which filename is most likely to retrieve /etc/passwd from a Linux host?

    1. ..../..../etc/passwd
    2. ..//..//etc/passwd
    3. ..\..\etc\passwd
    4. ....//....//etc/passwd Correct answer.

    Explanation

    The filter deletes the substring ../ once per match, so a sequence that contains ../ only after its own removal survives the filter as a usable traversal. In ....//....//etc/passwd each ....// collapses to ../, leaving ../../etc/passwd. The other inputs either contain no ../ at all or collapse to a path that does not traverse upward.

  128. Question 128Web Application Exploitation

    A login page sends users onward using a next parameter, and the code accepts the value only if it begins with a forward slash. Which value most likely still sends the victim to an attacker-controlled site?

    1. /login
    2. //attacker.example Correct answer.
    3. //dashboard
    4. https://attacker.example

    Explanation

    Browsers interpret a value starting with // as a scheme-relative absolute URL, so //attacker.example resolves to https://attacker.example while still satisfying a naive leading-slash check. Validation must compare the resolved host or restrict the value to a fixed path map. The other values either stay on the original host or fail the leading-slash check.

  129. Question 129Web Application Exploitation

    An attacker has found a UNION-based injection point where query results are rendered on the page. Which action determines how many columns the injected UNION SELECT must return?

    1. Appending a comment sequence to truncate the rest of the original query
    2. Increasing ORDER BY n until the database reports an error Correct answer.
    3. Replacing the parameter value with a boolean true condition
    4. Submitting a quote to force a syntax error message

    Explanation

    A UNION requires both branches to project the same number of columns, and ORDER BY n fails once n exceeds the column count, revealing the exact width. Comments are useful for trimming trailing query text but do not count columns. Error forcing and boolean tests help confirm injectability but do not measure the projection width.

  130. Question 130Web Application Exploitation

    An application returns the same generic page for every id value, surfaces no database errors, and never reflects query results. Which technique best confirms an injection point in that parameter?

    1. Injecting a stacked query that drops a temporary test table
    2. Appending a UNION SELECT statement that returns a large number of NULL columns
    3. Injecting a conditional time delay and comparing response times Correct answer.
    4. Sending an XSS payload and watching for script execution in the browser

    Explanation

    When no output or error is observable, inference must come from a side channel the attacker controls, such as a database sleep function gated on a condition. Timing differences confirm both injectability and the truth of the condition. UNION output and cross-site scripting both depend on data being reflected, and stacked queries risk destructive, uncontrolled effects.

  131. Question 131Web Application Exploitation

    An application copies a user-supplied value into a Location response header without filtering it. Which payload most clearly demonstrates CRLF injection that sets a cookie?

    1. javascript:alert(document.cookie)
    2. %0d%0aSet-Cookie: a=1 Correct answer.
    3. %00Set-Cookie: session=attacker
    4. ..%2f..%2f..%2fSet-Cookie: a=1

    Explanation

    Header injection requires carriage return and line feed characters to terminate the current header and start a new one; %0d%0a supplies them once the value is decoded. A script URL or path traversal payload changes the header value rather than splitting it. A null byte truncates a string but does not create additional headers.

  132. Question 132Web Application Exploitation

    A store validates a one-use discount code by first checking whether it has already been redeemed, then marking it redeemed in a separate statement. Which action best exploits this design?

    1. Changing the submitted code to an administrative coupon identifier
    2. Replaying one captured request after the response has already been received
    3. Issuing many redemption requests in parallel before the update commits Correct answer.
    4. Submitting the request over plain HTTP instead of HTTPS

    Explanation

    The check and the update are not atomic, so concurrent requests can all observe the unredeemed state before any write lands, allowing the code to be applied several times. This is a race condition rather than a replay issue. Sequential replay, identifier guessing, and transport downgrade do not exploit the gap between check and write.

  133. Question 133Web Application Exploitation

    An importer extracts an uploaded ZIP archive into a fixed temporary directory, using each entry's stored name as the destination path without canonicalizing the result. Which entry name yields a write outside that directory?

    1. ftp://attacker.example/shell.jsp
    2. shell.jsp.exe
    3. shell.jsp%00.png
    4. ....//....//etc/passwd Correct answer.

    Explanation

    Archive entry names are not URL-encoded and are used literally, so a name containing ../ sequences-here doubled to survive a single-pass filter-escapes the extraction root when the destination is joined without canonicalization. This is the Zip Slip pattern. Remote URLs, double extensions, and null-byte suffixes affect how a stored file is later interpreted, not where it is written.

  134. Question 134Web Application Exploitation

    An origin server returns an authenticated user's profile for /account/profile and ignores any extra path segments after it. A CDN in front caches every response whose URL path ends in .css. How can an attacker retrieve a victim's cached profile page?

    1. Send the victim a link to /account/profile;x.css, then request that same URL Correct answer.
    2. Request /account/profile with an X-Forwarded-Host header naming the attacker's server
    3. Request /account/profile with a Cache-Control: no-store request header
    4. Forge the victim's session cookie and request the profile page directly

    Explanation

    The cache keys on the extension while the origin routes on the prefix, so a request for /account/profile;x.css both returns the victim's profile and is stored as a static asset; the attacker then fetches the same URL from the cache. This is web cache deception. Host-header tricks, client cache directives, and session forgery do not place the victim's authenticated response into a shared cache.

  135. Question 135Identity and Directory Service Attack Paths

    Kerberoasting targets which attribute of a domain account?

    1. servicePrincipalName Correct answer.
    2. ms-Mcs-AdmPwd
    3. userAccountControl
    4. msDS-AllowedToDelegateTo

    Explanation

    Any authenticated domain user can request a service ticket for an account that has an SPN, and the KDC encrypts part of that ticket with the service account's long-term key. The ticket can therefore be cracked offline, so accounts with servicePrincipalName populated are exposed to any low-privileged user.

  136. Question 136Identity and Directory Service Attack Paths

    Which account setting makes a user vulnerable to AS-REP roasting?

    1. DONT_REQUIRE_PREAUTH Correct answer.
    2. TRUSTED_FOR_DELEGATION
    3. PASSWD_NOTREQD
    4. DONT_EXPIRE_PASSWORD

    Explanation

    With pre-authentication disabled (UF_DONT_REQUIRE_PREAUTH), the KDC returns an AS-REP encrypted with the user's key to any requester without proof of identity. That response can be cracked offline to recover the account password.

  137. Question 137Identity and Directory Service Attack Paths

    Which permission on the domain object is most directly abused to perform DCSync?

    1. WriteDACL on the domain object
    2. GenericRead on the domain NC
    3. DS-Replication-Get-Changes-All Correct answer.
    4. DS-Replication-Get-Changes alone

    Explanation

    DCSync impersonates a domain controller through the DRSUAPI replication interface. The DS-Replication-Get-Changes-All right permits replication of secret data, which is what allows password hashes to be pulled over the network.

  138. Question 138Identity and Directory Service Attack Paths

    A golden ticket is forged and signed with which key?

    1. The NTLM hash of the krbtgt account Correct answer.
    2. The private key of the domain's root CA
    3. The NTLM hash of the targeted service account
    4. The password hash of a Domain Admin

    Explanation

    All TGTs issued by the domain are encrypted with the krbtgt account's key. An attacker holding its NTLM hash can forge a TGT for any identity with any group memberships, and the KDC will accept it.

  139. Question 139Identity and Directory Service Attack Paths

    In BloodHound, what does a shortest-paths-to-Domain-Admins query return?

    1. Domain controllers that still allow NTLM authentication
    2. Abusable paths from owned principals to Domain Admins Correct answer.
    3. GPOs linked to OUs that grant write access to Domain Admins
    4. Kerberoastable accounts sorted by password age

    Explanation

    BloodHound models the directory as a graph in which permissions, group links and session data become edges. The query returns the minimum-hop chains of abusable edges linking principals the operator controls to Domain Admins.

  140. Question 140Identity and Directory Service Attack Paths

    An attacker has 500 usernames and a policy that locks accounts after five failed attempts. Which approach best avoids lockout?

    1. Submitting the username as the password ten times per account
    2. Submitting 400 candidate passwords against one high-value account
    3. One candidate password per account, then waiting out the window Correct answer.
    4. Submitting five candidate passwords to every account in turn

    Explanation

    Spraying spreads a very small number of attempts across many accounts so that no single account approaches the lockout threshold. Pausing before the next round prevents failed attempts from accumulating on any account.

  141. Question 141Identity and Directory Service Attack Paths

    Which technique captures NetNTLMv2 hashes from users who mistype a share name?

    1. DNS zone transfer from an internal resolver
    2. DHCP starvation of the local subnet
    3. LLMNR/NBT-NS spoofing of name queries Correct answer.
    4. ARP spoofing of the default gateway

    Explanation

    When normal name resolution fails, Windows falls back to LLMNR and NBT-NS broadcasts. A rogue responder answers those broadcasts and induces the client to authenticate to it, exposing a NetNTLMv2 challenge response.

  142. Question 142Identity and Directory Service Attack Paths

    Which mechanism re-applies the AdminSDHolder ACL to protected accounts?

    1. The Knowledge Consistency Checker running on each DC
    2. SDProp, which runs every 60 minutes
    3. The File Replication Service Correct answer.
    4. Group Policy refresh, every 90 minutes

    Explanation

    SDProp runs on the PDC emulator every 60 minutes by default and overwrites the security descriptors of AdminCount=1 objects with the AdminSDHolder template. Attacker ACEs planted on those objects disappear unless the template itself is modified.

  143. Question 143Identity and Directory Service Attack Paths

    An attacker has GenericWrite on a user object that currently has no SPN. Which technique produces a crackable Kerberos artifact?

    1. AS-REP roasting after clearing the pre-auth flag
    2. Silver ticket forgery using the user's SID
    3. Requesting a TGT with the user's AES key
    4. Adding an SPN and requesting a service ticket Correct answer.

    Explanation

    GenericWrite permits writing the servicePrincipalName attribute, creating a service identity out of an ordinary account. The attacker can then request a service ticket for that SPN and crack its encrypted portion offline without ever knowing the password.

  144. Question 144Identity and Directory Service Attack Paths

    A server is trusted for unconstrained delegation. What does coercing a domain controller to authenticate to it give an attacker?

    1. The DC's local SAM database
    2. The DC's cached TGT, usable for DCSync Correct answer.
    3. An NTLM relay channel to LDAP over TLS
    4. A service ticket encrypted with the krbtgt key

    Explanation

    Unconstrained delegation causes the server to cache a complete TGT for every account that authenticates to it. Forcing a domain controller to connect therefore yields the DC machine account's TGT, which can be replayed to replicate directory secrets.

  145. Question 145Identity and Directory Service Attack Paths

    Which attribute defines the services a constrained-delegation account may access on behalf of a user?

    1. msDS-AllowedToDelegateTo Correct answer.
    2. msDS-KeyCredentialLink
    3. msDS-AllowedToActOnBehalfOfOtherIdentity
    4. userAccountControl

    Explanation

    Constrained delegation is configured through msDS-AllowedToDelegateTo, which lists the SPNs the account may present to on behalf of a user. If protocol transition is enabled, an attacker holding the account can impersonate arbitrary users to those listed services.

  146. Question 146Identity and Directory Service Attack Paths

    Which access does an attacker need on a computer object to configure resource-based constrained delegation against it?

    1. Local administrator rights on the target host
    2. Domain Admin group membership
    3. GenericWrite on the target computer object Correct answer.
    4. SeEnableDelegationPrivilege on a domain controller

    Explanation

    RBCD is driven entirely by the target's msDS-AllowedToActOnBehalfOfOtherIdentity attribute, so write access to that object is sufficient. No domain-wide privilege or local administrator access is required.

  147. Question 147Identity and Directory Service Attack Paths

    An attacker holds GenericAll over a group that is nested inside Domain Admins. What is the most direct next step?

    1. Change the group's GPO link order
    2. Reset the krbtgt password twice
    3. Request a TGS for the group's SID
    4. Add a controlled account Correct answer.

    Explanation

    GenericAll over a group includes the right to modify its membership. Adding an account the attacker controls to that group confers Domain Admins privileges through the existing nesting.

  148. Question 148Identity and Directory Service Attack Paths

    Which extended right on a user object lets an attacker set a new password without knowing the current one?

    1. WriteDACL
    2. User-Force-Change-Password Correct answer.
    3. Replicating Directory Changes All
    4. Send-As

    Explanation

    The User-Force-Change-Password control access right allows a principal to set a new password without supplying the old one. It is commonly delegated to helpdesk staff and is a frequent privilege escalation edge.

  149. Question 149Identity and Directory Service Attack Paths

    An attacker has WriteDACL on the domain object and no other rights. How can this be turned into credential theft?

    1. Grant their account DCSync rights and replicate Correct answer.
    2. Enable unconstrained delegation on their own account
    3. Read the LAPS password of every domain controller
    4. Forge a silver ticket for the LDAP service

    Explanation

    WriteDACL lets the attacker rewrite the domain object's security descriptor. Granting their own account DS-Replication-Get-Changes-All then enables DCSync replication of password hashes over the network.

  150. Question 150Identity and Directory Service Attack Paths

    In an ESC1 vulnerable template, what does the attacker supply in the enrolment request to impersonate a privileged user?

    1. A forged krbtgt signature
    2. A trusted publisher signature
    3. Any user principal name in the request Correct answer.
    4. The domain controller's machine account hash

    Explanation

    ESC1 templates allow the requester to define the subject and carry a client authentication EKU. An attacker with enrolment rights submits a subjectAltName containing a privileged UPN and receives a certificate usable for PKINIT.

  151. Question 151Identity and Directory Service Attack Paths

    Which AD CS service is the target of an ESC8 NTLM relay attack?

    1. The HTTP enrollment endpoint accepting NTLM Correct answer.
    2. The certificate revocation list distribution point
    3. The offline root CA's hardware security module
    4. The LDAP service on the issuing CA server

    Explanation

    ESC8 relays NTLM authentication to an enrolment endpoint that accepts NTLM over HTTP. Relaying a machine or domain controller account yields a certificate that can be used with PKINIT to obtain a TGT for that identity.

  152. Question 152Identity and Directory Service Attack Paths

    An attacker gains write access to a GPO linked to the Domain Controllers OU. What is the most direct consequence?

    1. A change to the forest functional level
    2. Issuance of certificates for arbitrary users
    3. SYSTEM-level code execution on domain controllers Correct answer.
    4. Replication of directory secrets without touching a DC

    Explanation

    Policy settings apply to every computer under the linked OU, and startup scripts or scheduled tasks run as SYSTEM. Write access to that GPO therefore produces code execution on the domain controllers themselves.

  153. Question 153Identity and Directory Service Attack Paths

    Which attribute stores the LAPS-managed local administrator password?

    1. unixUserPassword
    2. msDS-ManagedPassword
    3. ms-Mcs-AdmPwd Correct answer.
    4. ms-Mcs-AdmPwdExpirationTime

    Explanation

    LAPS writes the rotated local administrator password to ms-Mcs-AdmPwd on the computer object. Read access to that attribute is a common escalation edge because it exposes the host's local administrator credential.

  154. Question 154Identity and Directory Service Attack Paths

    Which attribute holds the blob from which a group managed service account's password is derived?

    1. msDS-GroupMSAMembership
    2. msDS-ManagedPassword Correct answer.
    3. userAccountControl
    4. ms-Mcs-AdmPwd

    Explanation

    A gMSA stores a managed password blob in msDS-ManagedPassword, readable only by principals named in msDS-GroupMSAMembership. An attacker with read access to the blob can derive the service account password.

  155. Question 155Identity and Directory Service Attack Paths

    An attacker with a child domain's krbtgt hash forges an inter-realm TGT carrying the parent's Enterprise Admins SID. Why does this succeed?

    1. Intra-forest trusts do not filter same-forest SIDs Correct answer.
    2. The parent domain enforces SID filtering
    3. Kerberos armoring is disabled on the child DCs
    4. The child domain holds the parent's trust key

    Explanation

    SID filtering is applied to external and forest trusts, but SIDs belonging to the same forest are not stripped on intra-forest referral tickets. A forged ticket carrying a parent-domain privileged SID is therefore honoured.

  156. Question 156Identity and Directory Service Attack Paths

    Why must the krbtgt password be reset twice when remediating a golden ticket?

    1. To re-enable Kerberos pre-authentication
    2. To invalidate pre-authentication for all users
    3. To remove SID history from every account
    4. To purge the current and previous key versions Correct answer.

    Explanation

    Active Directory retains both the current and the previous krbtgt key, and the KDC accepts tickets signed with either. A single reset leaves the old key valid, so a second reset after replication is required to invalidate both keys an attacker may hold.

  157. Question 157Identity and Directory Service Attack Paths

    Which capability does membership in Protected Users remove?

    1. Access to LAPS-managed passwords
    2. Membership in Domain Admins
    3. Kerberos AES support
    4. Credential delegation to other services Correct answer.

    Explanation

    Protected Users members cannot delegate credentials through unconstrained or constrained delegation, cannot authenticate with NTLM or DES/RC4, and do not receive cached credentials. This blocks several credential-theft paths against privileged accounts.

  158. Question 158Identity and Directory Service Attack Paths

    Which setting lets ordinary domain users create up to ten computer accounts by default?

    1. msDS-AllowedToActOnBehalfOfOtherIdentity
    2. userAccountControl
    3. MaxTokenSize
    4. ms-DS-MachineAccountQuota Correct answer.

    Explanation

    The ms-DS-MachineAccountQuota attribute on the domain object allows authenticated users to join up to ten computers. Attackers abuse this to create a machine account whose credentials they control, a common prerequisite for resource-based constrained delegation.

  159. Question 159Identity and Directory Service Attack Paths

    Which mechanism keeps Domain Admins privileges in a user's token after the user is removed from the group?

    1. Setting primaryGroupID to 512 Correct answer.
    2. Adding the account to Protected Users
    3. Deleting the group's member attribute
    4. Forcing an NTLM logon instead of Kerberos

    Explanation

    A user's primary group SID is written into the access token separately from the enumerated group list. Setting primaryGroupID to 512 while the account is still a member leaves the Domain Admins SID in the token after removal, until an administrator resets the attribute.

  160. Question 160Identity and Directory Service Attack Paths

    During an assessment, a tester reads a Groups.xml preference file from a domain controller's SYSVOL share and finds a cpassword value alongside a local account name. What does that value represent?

    1. It is an integrity signature that validates the preference item and holds no credential material
    2. It is a one-way hash that must be cracked offline before any plaintext password can be recovered
    3. It is encrypted with a publicly documented key, so any domain user who can read the file can decrypt it Correct answer.
    4. It is a randomly generated AES key that encrypts the local administrator password, so it is useless to an attacker without the local machine key

    Explanation

    Group Policy Preferences encrypted stored passwords with a single AES key that Microsoft published in its documentation, so the cpassword field is reversible rather than protecting the secret. Any principal able to read the XML in SYSVOL can therefore recover the plaintext and reuse it against the referenced local account across the estate. Remediation requires deleting the preference item and rotating the exposed password.

  161. Question 161Privilege Escalation (Windows and Linux)

    On a Linux host, which command lists regular files that have the SUID bit set, while suppressing permission-denied noise?

    1. find / -user root -type f 2>/dev/null
    2. grep -R 'suid' /etc 2>/dev/null
    3. find / -perm -2000 -type f 2>/dev/null
    4. find / -perm -4000 -type f 2>/dev/null Correct answer.

    Explanation

    The SUID bit is represented by the octal mode 4000 in find's -perm test. Enumerating files with that mode reveals binaries that execute with the file owner's privileges, a core Linux escalation checklist step. The 2000 mode is SGID, and ownership or configuration greps do not identify the SUID bit.

  162. Question 162Privilege Escalation (Windows and Linux)

    A Linux binary has the file capability cap_setuid+ep. What can an attacker who can execute it do?

    1. Set its process effective UID to an arbitrary value such as 0 Correct answer.
    2. Capture raw packets from any interface
    3. Load unsigned modules into the running kernel
    4. Bind to TCP ports below 1024

    Explanation

    cap_setuid permits a process to call setuid() with any UID, so a binary carrying it effectively behaves like a SUID-root program. The +ep flags mean the capability is in the permitted and effective sets when the file runs. The other capabilities listed are CAP_NET_BIND_SERVICE, CAP_NET_RAW and CAP_SYS_MODULE.

  163. Question 163Privilege Escalation (Windows and Linux)

    An IIS worker process runs as a low-privileged account that holds SeImpersonatePrivilege. Which class of attack does this privilege enable?

    1. Replicating directory credentials over DRSUAPI
    2. Impersonating a higher-privileged token obtained through coerced local authentication Correct answer.
    3. Requesting service tickets for offline cracking
    4. Pass-the-hash using the machine account hash

    Explanation

    SeImpersonatePrivilege allows a process to adopt a token that another principal has authenticated to it. Attackers coerce a SYSTEM service into authenticating to a named pipe or local endpoint they control and then impersonate the resulting token. Pass-the-hash, Kerberoasting and DCSync rely on credential or ticket material, not on token adoption.

  164. Question 164Privilege Escalation (Windows and Linux)

    What does running `sudo -l` as an unprivileged user accomplish?

    1. It prints the installed sudo version and its compile-time flags
    2. It lists the commands the invoking user is permitted to run through sudo Correct answer.
    3. It prints the effective UID of the current shell
    4. It opens the sudoers file in the visudo safe editor

    Explanation

    The -l flag asks sudo to display the invoking user's sudoers entries, including allowed commands, runas targets and NOPASSWD tags. This output is the starting point for identifying GTFOBins-style escapes. Version printing, UID display and sudoers editing are performed by other flags or tools.

  165. Question 165Privilege Escalation (Windows and Linux)

    A Windows service has the ImagePath value C:\Program Files\My App\svc.exe written without quotation marks. Why is this exploitable?

    1. An unquoted path lets any authenticated user rewrite the ImagePath registry value
    2. The Service Control Manager refuses to start unquoted paths, leaving the service disabled and its files unprotected
    3. Windows tries each space-delimited path segment in turn, so a planted binary earlier in the search runs with the service's privileges Correct answer.
    4. An unquoted path causes the service to run at a lower integrity level than intended

    Explanation

    The Service Control Manager parses an unquoted path by trying successive prefixes, so C:\Program.exe is attempted before C:\Program Files\My App\svc.exe. If an attacker can create such a file in a writable location, it executes under the service account. Quoting the full path removes the ambiguity.

  166. Question 166Privilege Escalation (Windows and Linux)

    /etc/passwd is world-writable on a Linux host that uses shadow passwords. Which change most directly yields root access?

    1. Changing an existing account's UID field to 0 and authenticating as that account Correct answer.
    2. Setting the root account's login shell to /bin/false
    3. Copying the contents of /etc/shadow into the writable file
    4. Inserting a new password hash into the root entry's password field

    Explanation

    With shadow passwords in use, the password field in /etc/passwd is ignored, but the UID field is authoritative. Editing a known account's UID to 0 grants that account root-level file access once it logs in with its existing credentials. Writing a hash into the passwd field has no effect while a shadow entry exists.

  167. Question 167Privilege Escalation (Windows and Linux)

    What is GTFOBins used for during privilege escalation?

    1. Cataloguing ways legitimate Unix binaries can be abused to escalate privileges or escape restricted shells Correct answer.
    2. Translating Linux file capabilities into equivalent SUID permission sets
    3. Enumerating Windows services whose DACLs grant write access to low-privileged users
    4. Generating candidate passwords from leaked credential databases and policy rules

    Explanation

    GTFOBins documents, per binary, the arguments and side effects that let an attacker read files, spawn shells or otherwise break out of a constrained context. It is consulted after enumeration shows which binaries are runnable through sudo, SUID or capabilities. It is not a Windows service scanner, a wordlist generator or a capability converter.

  168. Question 168Privilege Escalation (Windows and Linux)

    A root cron job runs `cd /var/backups && tar -cf backup.tar *` every minute, and /var/backups is writable by your user. Which file name lets you inject a command through tar?

    1. tar.conf
    2. --checkpoint-action=exec=sh shell.sh Correct answer.
    3. backup.tar
    4. tarignore.tmp

    Explanation

    The shell expands the wildcard into file names, and tar interprets any name beginning with dashes as an option. A file named with a tar option such as --checkpoint-action=exec= runs the given command when the checkpoint feature fires. This is the classic wildcard injection pattern affecting tar, rsync, chown and similar tools.

  169. Question 169Privilege Escalation (Windows and Linux)

    A backup operator account holds SeBackupPrivilege but is not a local administrator. Which action leads to credential extraction?

    1. Running sekurlsa::logonpasswords from a medium-integrity shell
    2. Requesting a Kerberos service ticket for the krbtgt account
    3. Copying the SAM and SYSTEM registry hives with reg save and parsing them offline Correct answer.
    4. Dumping LSASS memory through Task Manager

    Explanation

    SeBackupPrivilege bypasses file ACLs for read access, allowing the account to save protected registry hives such as SAM, SYSTEM and SECURITY. Local account hashes can then be extracted offline from those hives. Reading LSASS memory or dumping its process requires SeDebugPrivilege, which this account lacks.

  170. Question 170Privilege Escalation (Windows and Linux)

    Which condition must hold for the AlwaysInstallElevated policy to allow running an MSI package as SYSTEM?

    1. Setting only the HKLM policy value is sufficient for msiexec to run elevated
    2. The policy applies only to per-user installs that Explorer launches interactively
    3. The policy forces every installer to run at medium integrity regardless of caller
    4. Both the HKLM and HKCU policy values must be set to 1 Correct answer.

    Explanation

    Windows only honours AlwaysInstallElevated when the value is set to 1 under both HKLM\SOFTWARE\Policies\Microsoft\Windows\Installer and the matching HKCU path. With both set, msiexec installs packages with SYSTEM privileges. Setting only the machine-wide value is not sufficient.

  171. Question 171Privilege Escalation (Windows and Linux)

    A root-owned SUID program calls `system("ps aux")` with a relative command name. Your user controls a directory that appears in the inherited PATH. Which action escalates privileges?

    1. Adding your directory to /etc/ld.so.conf and running ldconfig
    2. Replacing the SUID binary with your own copy in its own directory
    3. Placing an executable named ps in a PATH directory that is searched before /bin Correct answer.
    4. Creating a symlink named ps that points at /bin/ps in your directory

    Explanation

    Because the program resolves the command through PATH, a file with the same name in an earlier-writable directory is executed in place of the intended binary. Since the parent runs SUID-root, the planted program runs as root. Symlinking to the real binary merely runs the real command.

  172. Question 172Privilege Escalation (Windows and Linux)

    A low-privileged user is granted SERVICE_CHANGE_CONFIG on a service that runs as LocalSystem. What is the direct escalation path?

    1. Stopping and restarting the service to inherit its token
    2. Granting the account SeLoadDriverPrivilege through the service
    3. Repointing the service binary path to a payload and restarting the service Correct answer.
    4. Renaming the service so it is started earlier during boot

    Explanation

    SERVICE_CHANGE_CONFIG allows rewriting the service configuration, including its binary path. Once binPath points at attacker-controlled code, restarting the service executes that code under the service account, here LocalSystem. Display names, restart timing and driver privileges are unrelated to this right.

  173. Question 173Privilege Escalation (Windows and Linux)

    A helper binary carries cap_dac_read_search+ep and is executable by your user. What does that capability grant?

    1. Reading any file on the system regardless of its permission bits, including /etc/shadow Correct answer.
    2. Full write access to files owned by other users
    3. Changing the owning group of arbitrary files
    4. Mounting filesystems without supplying a password

    Explanation

    CAP_DAC_READ_SEARCH bypasses discretionary access control for read and directory-search operations, so protected files such as /etc/shadow become readable. It does not confer write access, ownership changes or mount rights. Those map to CAP_DAC_OVERRIDE, CAP_CHOWN and CAP_SYS_ADMIN respectively.

  174. Question 174Privilege Escalation (Windows and Linux)

    A low-privileged service account holds SeImpersonatePrivilege on a patched, current Windows host. Which technique obtains a SYSTEM token?

    1. Dumping cached domain credentials from the SECURITY registry hive
    2. Coercing the Print Spooler into authenticating to a named pipe you control, then impersonating the resulting token Correct answer.
    3. Creating a scheduled task that runs under the SYSTEM account
    4. Extracting the machine account hash and forging a silver ticket

    Explanation

    Printer and spooler based coercion tricks force a SYSTEM service to authenticate to an attacker-controlled named pipe, producing an impersonation token that the attacker can adopt. This works even on fully patched systems because it abuses a legitimate privilege rather than a memory corruption bug. The other options require different privileges or credentials.

  175. Question 175Privilege Escalation (Windows and Linux)

    Which sudoers configuration enables LD_PRELOAD-based privilege escalation?

    1. An env_keep entry for LD_PRELOAD plus permission to run a dynamically linked command Correct answer.
    2. Defaults secure_path pointing at /usr/local/sbin
    3. Defaults env_reset together with a command whitelist of /bin/ls
    4. Defaults requiretty enabled for the invoking user

    Explanation

    sudo normally strips LD_PRELOAD, but an env_keep entry preserves it into the privileged child process. The preloaded shared object's constructor then runs with the target privileges of the sudo command. Environment reset, secure_path and requiretty do not by themselves allow library injection.

  176. Question 176Privilege Escalation (Windows and Linux)

    An NFS export is mounted on a client with no_root_squash. How does this enable root access on the server?

    1. Reading the server's protected files directly through the export with your normal user
    2. Mounting the export as root and placing a root-owned SUID shell on it, then running that shell on the target Correct answer.
    3. Placing a script on the export and waiting for the server's scheduled jobs to run it
    4. Overwriting the server's /etc/sudoers file across the export as an unprivileged user

    Explanation

    With no_root_squash, files created by a root client process keep UID 0 on the server's backing filesystem. An attacker who can mount the export locally creates a SUID-root binary whose ownership is preserved, and executing it on the target grants root. Root squashing normally defeats this by mapping remote root to nobody.

  177. Question 177Privilege Escalation (Windows and Linux)

    Which description matches CVE-2021-4034 in pkexec?

    1. pkexec trusts XDG_SESSION_TYPE and loads a library from the current directory
    2. pkexec follows symbolic links in /etc/polkit-1/rules.d, letting a user rewrite policy files
    3. pkexec fails to drop supplementary groups, exposing a local container socket
    4. pkexec mishandles an empty argument vector, letting a crafted GCONV_PATH reintroduce attacker-controlled environment variables Correct answer.

    Explanation

    The flaw arises when pkexec is invoked with no arguments, causing it to process an out-of-bounds argv entry and re-import environment values it had cleared. Injecting GCONV_PATH causes a crafted shared object to be loaded with root privileges. Exploiting it requires only local execution of the installed pkexec binary.

  178. Question 178Privilege Escalation (Windows and Linux)

    An administrator-level shell holds SeDebugPrivilege. Which action produces interactive logon credential material?

    1. Decrypting SYSVOL cpassword values with the published AES key
    2. Reading the LSA secrets key from the registry with no other privilege
    3. Opening the LSASS process and dumping its memory for offline credential extraction Correct answer.
    4. Opening the SAM registry key interactively to read local account hashes

    Explanation

    SeDebugPrivilege permits opening protected processes such as LSASS, whose memory holds credential material for logged-on sessions. Dumping that memory and parsing it offline recovers plaintext or hash material. The privilege is what makes the process handle obtainable; registry-only reads do not require it.

  179. Question 179Privilege Escalation (Windows and Linux)

    A sudoers rule reads `user ALL=(ALL, !root) NOPASSWD: /usr/bin/vi`. Which invocation exploits CVE-2019-14287 in this configuration?

    1. sudo -u#-1 /usr/bin/vi Correct answer.
    2. sudo -g root /usr/bin/vi
    3. sudo -E /usr/bin/vi
    4. sudo -u root /usr/bin/vi

    Explanation

    The flaw let a user bypass the runas denial because sudo compared the requested user name against the exclusion list while resolving -u#-1 to UID 4294967295, which wrapped to 0. The command therefore ran as root despite the !root restriction. Explicitly naming root or another group is still blocked by the rule.

  180. Question 180Privilege Escalation (Windows and Linux)

    A Groups.xml file in SYSVOL contains a cpassword attribute. What is the key fact about recovering that password?

    1. The AES key Microsoft published makes the value decryptable to plaintext Correct answer.
    2. It is readable only by Domain Admins because of SYSVOL ACLs on the file
    3. It is a one-way hash that must be cracked offline with a wordlist
    4. It is encrypted with the domain krbtgt key and requires DCSync to decrypt

    Explanation

    Group Policy Preferences encrypted passwords used a fixed AES key that Microsoft publicly documented, so any authenticated user who can read SYSVOL can decrypt them. The vendor patch prevents new cpassword values but does not invalidate existing ones. This makes SYSVOL review a standard escalation step.

  181. Question 181Reporting, Evidence Handling and Remediation Guidance

    What is the primary purpose of the executive summary in a penetration test report?

    1. To list every tool, command, and payload used during the testing window
    2. To convey overall risk and business impact to non-technical leaders Correct answer.
    3. To document the tester's certifications, methodology, and reference sources
    4. To reproduce packet captures for each confirmed vulnerability finding

    Explanation

    The executive summary is written for decision makers who need to understand the risk posture, headline findings, and strategic direction rather than exploit mechanics. Technical detail such as tooling, payloads, and packet captures belongs in the body or appendices. Keeping the summary business-focused lets leadership prioritize budget and remediation ownership.

  182. Question 182Reporting, Evidence Handling and Remediation Guidance

    During an authorized test you gain domain administrator access to a production system earlier than planned. What should you do first?

    1. Notify the client's point of contact per the escalation plan Correct answer.
    2. Fix the misconfiguration yourself to reduce the client's exposure
    3. Wait and include it only in the final written report at the end of the project
    4. Continue exploiting the environment until the agreed engagement window closes

    Explanation

    Rules of engagement normally define an escalation path for critical or unexpected access so the client can respond while the exposure is live. Out-of-band notification is expected when a test crosses into high-impact territory such as production domain control. Testers should never remediate client systems themselves unless the contract explicitly authorizes it.

  183. Question 183Reporting, Evidence Handling and Remediation Guidance

    A chain-of-custody record for digital evidence primarily documents which of the following?

    1. The list of remediation steps recommended to the client's engineers
    2. The tester's hourly billing against the signed statement of work
    3. The vulnerability severity ratings assigned to each affected host
    4. Who handled the evidence, when, and for what purpose Correct answer.

    Explanation

    Chain of custody is an unbroken record of every transfer and access of evidence, including the identity of the handler, the date and time, and the reason for handling. It exists to show that the evidence was not altered or substituted from acquisition to presentation. Severity ratings and remediation advice are unrelated to custody.

  184. Question 184Reporting, Evidence Handling and Remediation Guidance

    Which mechanism best demonstrates that a captured evidence file has not been altered since acquisition?

    1. Storing the file on the tester's own encrypted laptop hard drive
    2. Renaming the file with the acquisition date and tester initials
    3. A hash recorded at acquisition and re-verified later Correct answer.
    4. Compressing the file into a password-protected archive

    Explanation

    A cryptographic hash taken at the moment of acquisition acts as a fingerprint; recomputing it later and matching the stored value proves the content is unchanged. Encryption and naming conventions protect or describe a file but do not detect modification. Archiving changes the container, not the assurance of integrity.

  185. Question 185Reporting, Evidence Handling and Remediation Guidance

    Which part of a penetration test report is written primarily for the administrators and developers who will implement fixes?

    1. Technical findings with reproduction steps and remediation detail Correct answer.
    2. Distribution list and document classification markings page
    3. Statement of scope and rules of engagement
    4. Executive summary

    Explanation

    The technical findings section carries the evidence, reproduction steps, affected components, and fix guidance that remediation owners need to act. The executive summary targets leadership and the scope statement governs the engagement. Administrative pages serve governance and handling, not repair.

  186. Question 186Reporting, Evidence Handling and Remediation Guidance

    A finding's evidence screenshot shows a valid user password in a login form. What is the best practice for including it in the report?

    1. Redact the secret while keeping the context that shows the issue Correct answer.
    2. Omit the finding entirely because its evidence contains a credential
    3. Replace the screenshot with a short text description and no supporting image
    4. Include the credential unchanged so the client's engineers can verify it quickly

    Explanation

    Evidence should preserve proof of the issue while removing unnecessary sensitive values such as passwords, tokens, or personal data. Redaction keeps the artifact usable for verification without spreading a working secret through email and document stores. Dropping the finding or the image weakens the report's evidentiary value.

  187. Question 187Reporting, Evidence Handling and Remediation Guidance

    A finding carries a high CVSS base score, but the affected host is an isolated lab system with no data and no path to production. How should the report's risk rating be presented?

    1. Use whatever rating the client's sponsor prefers, regardless of the technical facts
    2. Combine technical severity with context such as exposure, data sensitivity, and compensating controls Correct answer.
    3. Use the raw CVSS base score alone, because it is the recognized industry standard
    4. Assign the lowest possible rating, because CVSS scores are known to be unreliable

    Explanation

    CVSS base scores describe intrinsic technical severity and do not account for environment. A defensible report rating explains the technical severity and then adjusts for real exposure, business criticality, and existing controls. Showing both the base score and the contextual rating keeps the assessment transparent.

  188. Question 188Reporting, Evidence Handling and Remediation Guidance

    The engagement contract says nothing about how long test evidence should be kept. What is the most defensible approach?

    1. Keep all evidence indefinitely on tester laptops in case it is ever requested
    2. Upload the raw evidence to a public file-sharing link for convenient client access
    3. Apply the firm's documented retention policy and agree a period with the client in writing Correct answer.
    4. Delete everything immediately after the final report has been delivered

    Explanation

    Evidence retention should follow a documented policy and be confirmed with the client so both parties know when secure destruction occurs. Indefinite hoarding increases breach exposure, while immediate deletion can destroy records needed for retests or disputes. Any legal hold overrides the normal retention schedule.

  189. Question 189Reporting, Evidence Handling and Remediation Guidance

    A screenshot is captured as proof of a time-sensitive vulnerability. Which practice most strengthens its evidentiary value?

    1. Rely on the operating system's file creation date as the only timestamp
    2. Take the screenshot at the lowest possible resolution to save storage space
    3. Capture the system clock and target URL alongside the content, and note the time zone Correct answer.
    4. Crop out the browser address bar so the vulnerable field is more visible

    Explanation

    Evidence is far stronger when it shows the target identity and the moment of capture, including the time zone so the record can be correlated. File system dates are metadata that can change when files are copied or edited. Contextual framing makes the artifact reproducible and harder to challenge.

  190. Question 190Reporting, Evidence Handling and Remediation Guidance

    Which remediation recommendation is most useful to a client's engineering team?

    1. Apply vendor best practices to the application where appropriate
    2. Upgrade to the patched version and verify the vulnerable setting is disabled Correct answer.
    3. Consider commissioning a further security review of the application
    4. Improve the overall security posture of the affected server as soon as practical

    Explanation

    Actionable guidance names the specific change and the verification step so engineers can close the issue and prove it. Vague appeals to best practice or posture give no target for implementation or retest. References to the root cause help prevent recurrence rather than treating symptoms.

  191. Question 191Reporting, Evidence Handling and Remediation Guidance

    After the client says all findings have been fixed, what should a retest do before closing them?

    1. Resolve only the findings that were originally rated as critical or high
    2. Close the findings once the client's written confirmation of fixes arrives
    3. Re-rate each finding's CVSS score without re-testing the environment
    4. Re-run the original exploit and verify the fix, including attempts to bypass it Correct answer.

    Explanation

    Closure requires verified evidence that the vulnerability is no longer exploitable, not a client assertion. Retesting should repeat the original technique and try reasonable variants, because partial fixes often leave bypasses. The result, including failed exploit attempts, is recorded as retest evidence.

  192. Question 192Reporting, Evidence Handling and Remediation Guidance

    During testing you discover a vulnerable host that is not listed in the agreed scope. What is the appropriate action?

    1. Ignore it completely and leave it out of the final report
    2. Test it fully, since the vulnerability is already visible from in-scope hosts
    3. Patch the host yourself to remove the exposure before reporting
    4. Stop, document the observation, and notify the client contact Correct answer.

    Explanation

    Authorization limits which systems may be tested, and interacting with an out-of-scope asset can be unlawful even if the weakness is obvious. The professional response is to halt interaction, record what was incidentally observed, and let the client decide whether to expand scope. Unauthorized remediation is also prohibited.

  193. Question 193Reporting, Evidence Handling and Remediation Guidance

    How should collected penetration test evidence be stored between acquisition and report delivery?

    1. Encrypted at rest in an access-controlled repository with logged access Correct answer.
    2. On removable media carried in the tester's laptop bag
    3. In a shared team folder with the provider's default permissions
    4. In the client's public-facing web portal for easy retrieval

    Explanation

    Test evidence frequently contains credentials, personal data, and exploitable detail, so it should be treated as sensitive client data. Encryption at rest plus access control and audit logging limits exposure and supports accountability. Default permissions and portable media create avoidable loss and disclosure paths.

  194. Question 194Reporting, Evidence Handling and Remediation Guidance

    Before writing up a suspected SQL injection, what should the tester do?

    1. Report it immediately as critical so that a real issue is never missed
    2. Ask the client to confirm whether the behavior is actually a real issue
    3. Reproduce it reliably, rule out false positives, and capture minimal proof Correct answer.
    4. Wait for the automated scanner to flag it again in the next scan run

    Explanation

    Findings should be validated, reproducible, and supported by the minimum evidence needed to prove the issue without damaging data. False positives erode the report's credibility and waste remediation effort. Confirming manually also rules out scanner artifacts and application error handling quirks.

  195. Question 195Reporting, Evidence Handling and Remediation Guidance

    A finished report contains exploitable detail about unpatched systems. How should distribution be handled?

    1. Classify it, deliver over an agreed secure channel, and track recipients Correct answer.
    2. Email it to every employee at the client so everyone is aware
    3. Post it on the company's public blog so affected parties can act quickly
    4. Attach it unencrypted to the kickoff meeting invitation for convenience

    Explanation

    Reports containing attack detail are sensitive documents and should carry a classification marking that governs handling. Delivery over an agreed secure channel with a recipient list keeps distribution controlled and auditable. Broad or unencrypted distribution turns the report itself into a security risk.

  196. Question 196Reporting, Evidence Handling and Remediation Guidance

    Which set of elements best allows a reader to reproduce a reported finding?

    1. A summary of the client's industry and a generic recommendation
    2. Affected endpoint, preconditions, exact request or steps, and observed result Correct answer.
    3. The tool name and the tester's personal opinion of the severity level
    4. Screenshots only, with no narrative or request detail at all

    Explanation

    Reproduction requires knowing where the issue lives, what must be true beforehand, precisely what was sent, and what came back. Together these let a defender confirm the flaw and verify the fix independently. Missing any element forces guesswork and slows remediation.

  197. Question 197Reporting, Evidence Handling and Remediation Guidance

    Which combination most supports the use of digital evidence in a later legal or disciplinary proceeding?

    1. Screenshots stored on a personal phone with no timestamps recorded
    2. Compressed copies produced separately by each analyst who reviewed the file
    3. A verbal account of the acquisition given later by the lead tester
    4. Unchanged originals, hashes recorded at acquisition, and an unbroken custody log Correct answer.

    Explanation

    Admissibility rests on showing that the evidence is what it claims to be and has not been tampered with. Preserving originals, hashing at acquisition, and maintaining an unbroken custody record together establish authenticity and integrity. Each analyst re-copying or re-processing material introduces doubt about which version is authentic.

  198. Question 198Reporting, Evidence Handling and Remediation Guidance

    While testing, you obtain a database dump containing personal data of real customers. What is the most appropriate handling?

    1. Publish anonymized samples from the dump in the final report appendix
    2. Keep minimal proof, protect it as sensitive, and notify the client Correct answer.
    3. Share the dump with the client's marketing team for further analysis
    4. Retain the full dump so the appendix is complete for the client's records

    Explanation

    Testers should collect only the data needed to demonstrate the issue and treat anything else as sensitive personal information. The client must be told because they may have regulatory notification obligations and because the dump should be secured or destroyed promptly. Redistribution to other teams or into report appendices expands the exposure.

  199. Question 199Reporting, Evidence Handling and Remediation Guidance

    A critical finding cannot be fixed for six months because of a vendor dependency. What remediation guidance is most appropriate?

    1. Insist that the affected system be shut down immediately in every case
    2. Advise the client to accept the risk quietly until the vendor patch arrives
    3. Recommend compensating controls, a documented risk decision, and a re-check date Correct answer.
    4. Remove the finding from the report because it cannot be fixed right away

    Explanation

    Where a permanent fix is delayed, the report should offer interim measures that reduce exposure and record that leadership accepted the residual risk with a review date. This keeps the decision accountable and time-bound. Deleting the finding or demanding shutdown ignores operational reality and the client's authority over risk.

  200. Question 200Reporting, Evidence Handling and Remediation Guidance

    Shortly before delivery, a client sponsor asks you to delete a validated critical finding so it will not affect an upcoming audit. What is the correct response?

    1. Decline, explain the implications, and escalate through agreed channels Correct answer.
    2. Relabel the finding as informational so the sponsor is satisfied
    3. Delete it but mention it verbally to one of the client's engineers
    4. Delete it, because the client owns the report and its contents

    Explanation

    Reports must accurately reflect validated results; removing or downgrading a confirmed finding misrepresents the security posture to auditors and other stakeholders. The tester should decline, explain the professional and evidentiary consequences, and raise the request through the governance path defined in the engagement. Documenting the request protects both parties.

CSPT flashcards

32 cards. Read the term and the definition here. The deck at the top of this page is optional.

Engagement Letter
Formal document authorizing penetration testing and defining parties, scope, and limitations; an authorized representative signs it before active testing begins.
Non-Disclosure Agreement
Legal contract that prohibits disclosure of confidential information learned during testing and protects client data, credentials, and findings from unauthorized release.
Escalation Contact
Named individual and backup reachable during testing for emergencies, scope questions, and incident response; confirm both before active testing starts.
In-Scope Assets
Explicit list of systems, networks, applications, cloud accounts, or people authorized for testing, expressed as CIDR ranges, hostnames, URLs, or account identifiers.
Out-of-Scope Assets
Assets and techniques explicitly excluded from testing, such as third-party SaaS, shared infrastructure, production data, or partner networks, to prevent unauthorized impact.
Testing Window
Approved dates, times, and time zone during which active testing may occur, including blackout periods and maximum request rates.
Passive Reconnaissance
Information gathering from public sources without sending packets to the target, including WHOIS, DNS records, certificate logs, search engines, and breach data.
Certificate Transparency
Public logs of TLS certificates that can reveal subdomains and hostnames; search by domain or organization using services such as crt.sh.
DNS Zone Transfer
AXFR query to an authoritative name server that returns all zone records if misconfigured; reveals hostnames, IPs, and mail servers.
Autonomous System Number
Identifier for a network's routing domain; ASN and BGP lookups map IP ranges to organizations for external reconnaissance.
CVSS Base Score
Numerical severity from 0 to 10 derived from exploitability and impact metrics; it excludes temporal and environmental factors.
False Positive
Scanner finding that does not represent an exploitable vulnerability; common causes include version-only matches, backported patches, and misread banners.
Authenticated Scan
Vulnerability scan that logs into hosts with credentials to inspect installed patches and configuration, reducing false positives compared with unauthenticated scanning.
CWE Identifier
Taxonomy code classifying the type of software weakness, such as CWE-79 for cross-site scripting, independent of any specific product vulnerability.
SSH Local Forwarding
Command ssh -L localport:target:targetport user@pivot binds a local listener that forwards traffic through the pivot to the target service.
SOCKS Proxy
Proxy protocol that relays arbitrary TCP connections; ssh -D or Chisel can create a SOCKS proxy for proxychains to route tools through a pivot.
Pass-the-Hash
Authentication technique using an NTLM hash directly instead of a plaintext password; tools such as Impacket and Mimikatz support it.
Insecure Direct Object Reference
Access control flaw where a user-controlled identifier, such as an account number or file ID, exposes another user's object without authorization checks.
Server-Side Request Forgery
Vulnerability where the server fetches a user-supplied URL, enabling access to internal services or cloud metadata such as 169.254.169.254.
Time-Based SQL Injection
Blind SQL injection technique that infers query results from deliberate database delays, using conditional sleeps or heavy queries.
Cross-Site Scripting
Injection of client-side script into pages viewed by other users; variants are reflected, stored, and DOM-based, and impact depends on session and CSP controls.
Kerberoasting
Attack that requests Kerberos service tickets for accounts with SPNs and cracks them offline to recover service account passwords.
DCSync
Credential theft technique that impersonates domain controller replication to request password hashes, requiring replication permissions such as DS-Replication-Get-Changes.
BloodHound
Graph analysis tool that ingests Active Directory collection data to reveal privilege escalation paths through ACLs, sessions, and group memberships.
SeImpersonatePrivilege
Windows privilege allowing a process to impersonate another user's token; service accounts often hold it, enabling Potato-family privilege escalation.
SUID Binary
Linux executable with the set-user-ID bit set; it runs with the file owner's privileges, often root, and can be abused via GTFOBins.
Unquoted Service Path
Windows service path containing spaces without quotes; the system may execute a malicious binary placed in an earlier writable directory.
Executive Summary
Non-technical report section describing overall security posture, key risks, and business impact for decision-makers, without exploit-level detail.
Finding Reproduction Steps
Ordered instructions, requests, and payloads that let a third party independently verify a reported vulnerability and confirm the affected asset.
Chain of Custody
Chronological record of evidence collection, handling, transfer, and storage, with signatures and timestamps, to preserve admissibility and integrity.
SHA-256 Hash
Cryptographic digest used to verify evidence integrity; compute at collection and before analysis, and record values in the report.
Remediation Guidance
Specific, prioritized actions to fix a finding, including configuration changes, patches, validation criteria, and residual risk acceptance.

CSPT study guide

Engagement Management, Scoping and Rules of Engagement

Engagement Management

The contract stack

Every engagement rests on three documents: the NDA (confidentiality both ways), the master services agreement (liability, insurance, indemnity), and the statement of work (the actual task, deliverables, timeline, and fee). The SOW is what defines your authority. Separately, the authorization to test letter names the client signatory, the in-scope assets, and the test window - it is the document you produce if law enforcement or a network defender challenges you mid-test.

Running the engagement

Competence in this domain is really about operational discipline: named points of contact on both sides, an escalation path for critical findings (who do you call at 02:00 when you find ransomware?), an agreed emergency stop procedure, a communication cadence, and deconfliction so the client's blue team knows your traffic is sanctioned. Track time and scope consumption; a scope creep discovered midway requires a written amendment, not a phone call.

Traps in exam questions

  • Confusing the MSA (commercial terms) with the SOW (technical scope) - questions often ask which document authorizes a specific action.
  • Assuming verbal approval is authorization. It is not.
  • Assuming a signature dated today covers testing that started yesterday.
  • Forgetting that third-party hosted assets (SaaS, shared hosting, cloud) need their own permission.

Rule of thumb

If an action is not written into the SOW or the rules of engagement, treat it as out of scope until you have written approval.

Worked example: mid-test you discover an undocumented /24 adjacent to the client's subnet. You do not scan it. You document the discovery, notify the client contact, and request a written scope amendment; only after the ROE is re-issued do you touch it.

Scoping and Rules of Engagement

Scope is an inventory, not an idea

A defensible scope lists assets precisely: CIDR ranges, exact FQDNs or explicitly bounded wildcards, application URLs, credentials, and source IPs to be allowlisted. Equally important is the exclusion list: third-party hosted services, production databases, partner networks, and anything the client cannot legally authorize you to test. Record whether the engagement is black box, grey box, or white box - this drives effort estimates and what a "finding" even means.

Rules of engagement content

The ROE converts scope into permitted behaviour: testing windows and time zones, allowed and forbidden techniques (denial of service, social engineering, physical access, data exfiltration), rate limits, whether destructive exploitation is permitted, how credentials may be used, how evidence is stored, and notification thresholds for critical findings.

Traps in exam questions

  • Wildcard scope (*.example.com) where some subdomains are hosted by a third party with no authorization.
  • Production vs staging confusion - the client says "test the app" and the candidate picks the production cluster.
  • Source IP not allowlisted, so your scan trips a WAF and gets the engagement suspended.
  • Time zone drift - a 22:00–06:00 window in the client's local time, not yours.
  • Cloud provider policy: most CSPs permit testing of your own instances but prohibit attacking the provider's infrastructure, and some services require prior notification.

Rule of thumb

Explicit inclusion beats reasonable assumption. When scope is ambiguous, pause and confirm in writing; the cost of a delay is trivial compared with an out-of-scope incident.

Worked example: an acquired legacy domain resolves to a marketing vendor's CDN. The client owns the brand, not the infrastructure. Testing it is out of scope even though the name is on the target list.

Reconnaissance and Information Gathering

Passive before active

Passive reconnaissance gathers information without touching client infrastructure: WHOIS and RIR records, ASN and netblock mapping, DNS records (A, MX, NS, TXT/SPF/DMARC), certificate transparency logs, search engine and code repository dumps, cloud storage bucket naming, and metadata embedded in published documents. Commercial and open sources (Shodan, Censys) index hosts you have not scanned, which keeps you quiet but does not make the data authoritative.

Active reconnaissance

Once the ROE permits, move to host discovery, port scanning, service and version detection, banner grabbing, virtual host enumeration, zone transfer attempts, and protocol-specific queries (SNMP, LDAP, SMB). Rate control matters: aggressive scans trip IDS and can destabilise fragile devices. Always correlate multiple sources - a port found on one tool and confirmed by another.

Traps in exam questions

  • The open port is not a vulnerability. It is an observation that may lead to one.
  • Wildcard DNS makes brute-force subdomain enumeration return thousands of false positives; compare response bodies and sizes against a random name.
  • Testing a CDN edge IP rather than the origin asset behind it.
  • Scope drift: discovery of a related domain does not authorize testing it.
  • Mistaking certificate transparency log queries (public data) for active scanning - close enough that the distinction is a common question.

Rule of thumb

Confirm ownership and scope before you touch a discovered host. Discovery creates a question, not a permission.

Worked example: subdomain enumeration returns dev-portal.example.com pointing to a residential IP. You verify via WHOIS and certificate data that it belongs to a contractor, not the client, and you exclude it while documenting the exposure as an observation.

Vulnerability Analysis and Validation

Scanner output is a hypothesis

Automated scanners generate both false positives (version-based guesses, backported patches, banner spoofing) and false negatives (business logic flaws, chained conditions, authenticated-only issues). Your job is validation: reproduce the condition manually, confirm the vulnerable code path, and record minimal, non-destructive proof.

Scoring and prioritisation

CVSS base metrics describe the vulnerability in the abstract - attack vector, complexity, privileges required, user interaction, scope, and confidentiality/integrity/availability impact. Environmental and temporal modifiers, plus asset criticality, determine actual risk. A high base score on an isolated lab-grade host may rank below a medium score on an internet-facing authentication service. Exploitability (public exploit, weaponised, authenticated?) is a separate axis from impact.

Proving without damaging

Use the least intrusive payload that demonstrates the condition: id rather than a reverse shell; SELECT @@version rather than data modification; a benign out-of-band DNS callback rather than file exfiltration. Never run destructive verification against production unless the ROE explicitly says so.

Traps in exam questions

  • Equating CVSS 9.8 with "fix first" without considering exposure and business context.
  • Reporting a banner version as proof of a CVE without confirming the patch level.
  • Treating a second scanner's agreement as validation - two tools reading the same banner agree about nothing.
  • Reporting theoretical or unexploited findings as confirmed.

Rule of thumb

No finding enters the report without a reproducible, minimal-impact verification step and the evidence to prove it.

Worked example: a scanner flags an outdated TLS library. Manual testing shows the host negotiates a deprecated cipher but the vulnerable function is unreachable. You downgrade severity and describe the configuration weakness accurately instead of citing the CVE.

Network Exploitation and Pivoting

From access to position

Network exploitation starts with an exposed service or valid credential set. Credential attacks must respect lockout policy - password spraying with a small candidate list across many accounts, not brute force against one account. Relay attacks depend on missing SMB signing or LDAP signing/channel binding. After initial access, establish situational awareness before acting: identity, host name, network interfaces and routes, running services, and the local security posture.

Pivoting mechanics

A pivot turns a compromised host into a route into otherwise unreachable segments. Core techniques:

  • SSH local forwarding (-L) exposes a remote service on your local port.
  • SSH remote forwarding (-R) exposes your service on the remote host.
  • SSH dynamic forwarding (-D) creates a SOCKS proxy.
  • chisel / ligolo-ng / socat for environments without SSH.
  • Proxychains to force tool traffic through a SOCKS proxy - note it handles TCP only, and UDP or ICMP will leak around it.

Plan pivot chains (double pivots) before you build them, and record every hop, as pivots are the most common cause of lost access and lost evidence.

Traps in exam questions

  • Reversing -L and -R under time pressure.
  • Running a scan without the proxy wrapper, sending traffic from your own address into a segmented network.
  • Triggering account lockout by ignoring the observed threshold.
  • Assuming a SOCKS proxy carries DNS; often it does not.

Rule of thumb

-L is Local to you and forwards out; -R is Remote and forwards back to you. Verify a pivot with curl ifconfig style checks against an internal host before scanning.

Worked example: a foothold in the DMZ reaches an internal /24 over TCP 8080 only. You build a SOCKS proxy through that port, tune proxychains, and confirm reachability with a single targeted request before running any enumeration.

Web Application Exploitation

Map before you break

Effective web testing begins with attack surface mapping: authentication mechanisms, session handling, roles and privilege boundaries, API endpoints, input sinks, file handling, and third-party integrations. Test both authenticated and unauthenticated states, and with at least two accounts of different privilege so that authorization flaws (IDOR/BOLA, forced browsing, function-level access control) can be proven rather than suspected.

Core vulnerability classes

Injection (SQL, command, template, LDAP), cross-site scripting (reflected, stored, DOM), server-side request forgery, XML external entity, insecure deserialisation, file upload and path traversal, JWT weaknesses (algorithm confusion, missing signature verification, weak secret), CSRF, and business logic abuse such as price manipulation or workflow bypass. Each needs a payload that demonstrates real impact on that specific application.

Traps in exam questions

  • Claiming stored XSS from a reflected-only vector, or claiming XSS impact where no session or sensitive action is reachable.
  • Reporting IDOR from a single account without demonstrating access to another user's object.
  • SSRF behind an allowlist - the answer is usually a redirect, DNS rebinding, or alternate URL parser, not brute force.
  • Treating a 403 as proof of safety rather than a hint about routing.
  • Time-based blind injection judged from one slow response instead of a repeated, controlled baseline.

Rule of thumb

Prove the maximum impact with the minimum invasiveness. SELECT @@version, a single alert-equivalent callback, or reading one file - never data modification, never a mass defacement payload.

Worked example: an API returns another customer's invoice when the numeric id is incremented. You note both accounts, capture the request and the returned object with the other account's identifier, and stop - you do not enumerate the full id space.

Identity and Directory Service Attack Paths

The mechanics you must know

Active Directory attacks are built on Kerberos and NTLM fundamentals. Kerberos pre-authentication protects AS-REQ; accounts with it disabled can be AS-REP roasted. Kerberoasting requests a service ticket for an account with a registered SPN and cracks it offline - RC4-encrypted tickets crack far more readily than AES. Delegation misconfigurations (unconstrained, constrained with protocol transition, and resource-based constrained delegation) each yield different paths to impersonation. ADCS misconfigurations, notably templates that allow the requester to supply a subject name combined with client-authentication EKU and enrol rights, allow certificate-based impersonation. DCSync depends on replication rights (DS-Replication-Get-Changes and related) granted to a principal. NTLM relay succeeds where signing and channel binding are absent.

Path thinking

Enumerate group memberships, ACLs, sessions, and delegation with tooling such as BloodHound, then reason about the shortest path to Tier-0: Domain Admin, krbtgt, or a domain controller. Every hop needs its own evidence.

Traps in exam questions

  • Kerberoasting assumed to work on any account - an SPN is required, and crackability depends on the encryption type and password strength.
  • Confusing golden ticket (krbtgt hash plus domain SID) with silver ticket (service account hash).
  • Assuming unconstrained delegation alone grants access without a privileged user authenticating to that host.
  • Treating password spraying as safe without checking the lockout threshold and observation window.
  • Ignoring that DCSync rights can be delegated to a non-admin principal.

Rule of thumb

Enumerate the graph before attacking a node. Identify the target Tier-0 asset, find the shortest edge sequence, and validate each edge with its own captured evidence.

Privilege Escalation (Windows and Linux)

Linux

Enumerate before exploiting: kernel and distribution version, SUID/SGID binaries, sudo -l output including allowed arguments, file capabilities, cron and systemd timers, writable service unit files, PATH abuse, NFS exports with no_root_squash, group memberships such as docker or lxd, readable credential files, and exposed SSH keys. GTFOBins classifies binaries that can break out of their intended function - but presence alone is not a finding; you need a path to execution in the current context.

Windows

Focus on service and task misconfiguration: unquoted service paths containing spaces, weak service ACLs allowing binPath rewrite, writable directories in a service's execution path (DLL search order), scheduled tasks running as SYSTEM in writable locations, AlwaysInstallElevated, stored credentials, and registry autoruns. Token privileges matter enormously: SeImpersonate/SeAssignPrimaryToken enables potato-style attacks, SeBackup/SeRestore enables file-level access to protected hives, SeDebug enables process memory access. Remember that an administrative shell is not SYSTEM - UAC separates the two.

Traps in exam questions

  • Running a kernel exploit on a production host: it can panic the machine and breach the ROE.
  • Assuming SUID bit or file capability equals escalation without a command that actually invokes it.
  • Confusing local administrator with SYSTEM or with domain privilege.
  • Ignoring that a writable path only matters if the privileged process resolves a binary there.

Rule of thumb

Prove the new privilege level by doing something that was previously denied - read /etc/shadow, or dump a protected registry hive - rather than relying on whoami output alone.

Reporting, Evidence Handling and Remediation Guidance

Reporting

Two audiences, one document

A penetration test report has an executive layer and a technical layer, and questions frequently probe the difference. The executive summary states overall risk posture, the most material themes, and business consequences in plain language - no CVE identifiers, no payloads. The technical body carries the reproducible detail.

Anatomy of a finding

Each finding should contain: a descriptive title, severity with the scoring rationale, affected assets, a clear description of the weakness and its root cause, reproduction steps precise enough for an administrator who was not present, captured evidence, realistic business impact, a specific remediation recommendation, and references. Deduplicate the same weakness across many hosts into one finding with an affected-asset list.

Quality control

Score consistently across findings; a report where severity ordering contradicts the executive summary destroys credibility. Separate technical severity from business risk. Remove unsupported claims, vague language, and anything you did not personally verify. Include scope, methodology, limitations, and testing dates so the reader knows what was and was not covered.

Traps in exam questions

  • Severity inflation to impress the client, which undermines the whole report.
  • Findings with no reproduction path or no remediation.
  • Screenshots with no hostname, timestamp, or surrounding context.
  • Mixing testing limitations into a footnote rather than declaring them explicitly.
  • Reporting an observation as a vulnerability.

Rule of thumb

Write every finding so that a system administrator who was not on the engagement can reproduce it unaided. If they cannot, the finding is incomplete regardless of how good the evidence is.

Evidence Handling and Remediation Guidance

Evidence integrity

Evidence must be attributable, unaltered, and retained per the contract. Practical controls: record hostname, source IP, timestamp and tester identity in every capture; save raw tool output alongside screenshots; hash exported evidence sets (SHA-256) and record the hash; store evidence encrypted at rest and transfer it over secure channels; and maintain a chain-of-custody record for anything that leaves your environment. Collect only what you need - pulling whole databases or user mailboxes creates privacy exposure with no reporting benefit. Redact credentials, tokens, personal data and customer records in the delivered report, retaining the unredacted originals in the secured evidence store.

Remediation guidance

Good remediation is specific, prioritised, and verifiable. It addresses root cause, not just the symptom: "apply the vendor patch and add a regression test to the monthly cycle," not "patch the server." Where a full fix is slow, propose compensating controls (network segmentation, WAF rule, monitoring alert) and state their residual risk. Define retest criteria so the client knows what "fixed" will look like, and offer defence-in-depth where one control failure would re-open the path.

Traps in exam questions

  • Pasting live credentials, password hashes, or full personal records into the report body.
  • Deleting evidence at engagement close without documented destruction and client confirmation.
  • Recommending only a point fix that leaves the vulnerable pattern in place elsewhere.
  • Remediation advice that contradicts the discovered architecture (blocking a port the business depends on).

Rule of thumb

Every piece of evidence should answer: what, where, when, and who captured it - and every recommendation should state how the client will prove it worked.

CSPT cheat sheet

Quick reference

Engagement Management

  • Written authorization from an officer with authority must precede active testing; verbal approval does not create legal coverage.

  • An SOW defines deliverables, timeline, fees, assumptions, and acceptance criteria; the RoE defines permitted techniques and constraints.

  • Engagement models: black-box (no internal knowledge), grey-box (partial knowledge or credentials), white-box (full source, configuration, and credentials).

  • Confidentiality terms specify handling of client data, PII, credentials, and findings; NDAs bind testers and subcontractors.

  • Emergency escalation contacts and stop conditions must be reachable for the full test window.

  • Scope changes require written change control approved by both parties before testing new assets.

  • Deliverables typically include executive summary, technical findings, evidence, risk ratings, and remediation guidance.

  • Liability, indemnification, and insurance requirements are negotiated in the contract, not the technical RoE.

Quick reference

Scoping and Rules of Engagement

  • Scope enumerates assets by CIDR, hostname, URL, cloud account, API, mobile app, physical site, or personnel.

  • Out-of-scope must list third-party SaaS, shared hosting, partner networks, production data, and excluded techniques.

  • RoE specifies allowed and forbidden techniques: DoS, social engineering, physical entry, data exfiltration, brute force, and destructive tests.

  • Testing windows include time zone, blackout periods, and maximum request rates.

  • Source IPs and VPN endpoints are allowlisted and documented; spoofed or rotating IPs may violate scope.

  • Emergency stop procedure names the contact, channel, and expected response time.

  • Data handling rules cover encryption, storage location, retention period, destruction method, and breach notification.

  • Legal references include the authorization letter, applicable CFAA or Computer Misuse Act, GDPR or DPA, and contractual limits.

Quick reference

Reconnaissance and Information Gathering

  • Passive recon uses public sources without directly touching the target: WHOIS, DNS records, CT logs, ASN and BGP data, search engines, and breach data.

  • Active recon sends packets to the target: port scans, service probes, web crawling, and banner grabs; it must be in scope.

  • DNS record types to enumerate: A, AAAA, CNAME, MX, NS, SOA, TXT, SRV, and PTR.

  • Zone transfer attempt: dig AXFR @nameserver domain; success reveals all zone records if the server is misconfigured.

  • Certificate Transparency logs reveal subdomains via crt.sh and Censys; search by domain and organization.

  • ASN lookup maps IP ranges to organizations; use whois, BGP.he.net, or RDAP to find netblocks.

  • Search dorks: site:, filetype:, inurl:, intitle:, and intext:; combine them to find exposed files and admin pages.

  • Email format discovery supports password spraying and phishing scope; verify the format from public sources.

Quick reference

Vulnerability Analysis and Validation

  • CVSS v3.1 base score uses AV, AC, PR, UI, S, C, I, and A; base score excludes temporal and environmental metrics.

  • CVSS vector example: CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H.

  • Authenticated scans use credentials and report host-level patch state; unauthenticated scans infer from network exposure.

  • Banner and version checks can mislead when vendors backport patches without changing version strings.

  • Validate findings by reproducing the issue with minimal impact, capturing request and response, and confirming the affected component.

  • Common false positives: TLS cipher claims, missing headers, ICMP responses, and version-only matches; manual verification is required.

  • CWE classifies weakness type; CVE identifies a specific vulnerability; CVSS scores severity, not business risk.

  • Risk equals likelihood times impact adjusted for asset criticality, exposure, and compensating controls.

Quick reference

Network Exploitation and Pivoting

  • SSH local forward: ssh -L localport:target:targetport user@pivot binds a local listener through the pivot.

  • SSH remote forward: ssh -R remoteport:target:targetport user@pivot exposes the target to the remote side.

  • SSH dynamic forward: ssh -D 1080 user@pivot creates a SOCKS proxy; proxychains routes tools through it.

  • Meterpreter route add target/netmask session_id and autoroute enable pivoting at the network layer.

  • Chisel and Ligolo-ng create TCP or UDP tunnels over HTTP or WebSocket; useful when SSH is blocked.

  • Pass-the-Hash uses an NTLM hash directly with tools such as Impacket or Mimikatz; no plaintext password is required.

  • Lateral movement protocols: SMB (445), WinRM (5985/5986), RDP (3389), WMI (135), and SSH (22).

  • Egress filtering may block reverse shells; DNS, ICMP, and HTTPS are common allowed channels.

Quick reference

Web Application Exploitation

  • SQLi types: error-based, union-based, boolean blind, time-based blind, and out-of-band; test with quote and comment sequences.

  • XSS types: reflected, stored, and DOM-based; impact depends on session cookies, CSRF tokens, and CSP.

  • SSRF targets internal services and cloud metadata; the AWS IMDSv1 endpoint is 169.254.169.254/latest/meta-data/.

  • IDOR occurs when object identifiers are user-controlled; test by swapping IDs between two authenticated accounts.

  • File upload bypass: content-type, extension blacklists, magic bytes, path traversal, and server-side execution checks.

  • XXE requires an XML parser with external entities enabled; test with file://, http://, and parameter entities.

  • JWT flaws: alg:none, weak HMAC secret, kid injection, and missing signature verification.

  • Access control tests must use at least two roles; compare vertical and horizontal privilege differences.

Quick reference

Identity and Directory Service Attack Paths

  • Kerberoasting requests service tickets for accounts with SPNs and cracks them offline; no elevated rights are needed.

  • AS-REP roasting targets accounts with Do not require Kerberos preauthentication and returns crackable AS-REP responses.

  • DCSync requires DS-Replication-Get-Changes and Get-Changes-All rights on the domain object; it mimics domain controller replication.

  • BloodHound ingests SharpHound data to map ACL edges: GenericAll, WriteDACL, ForceChangePassword, and AddMember.

  • Unconstrained delegation stores TGTs of connecting users; constrained delegation limits use to specified services.

  • Resource-Based Constrained Delegation abuses msDS-AllowedToActOnBehalfOfOtherIdentity on a target computer.

  • Golden Ticket forges a TGT with the krbtgt hash; Silver Ticket forges a service ticket with a service account hash.

  • Entra ID attacks: consent phishing, device code phishing, token theft, and service principal credential abuse.

Quick reference

Privilege Escalation (Windows and Linux)

  • Windows privilege check: whoami /priv, whoami /groups, systeminfo, and accesschk for service and ACL permissions.

  • SeImpersonatePrivilege enables token impersonation; Potato-family exploits target Windows service accounts.

  • Unquoted service paths containing spaces allow binary planting in writable directories earlier in the path.

  • AlwaysInstallElevated in HKLM and HKCU allows MSI installation with SYSTEM privileges when both keys are set to 1.

  • Linux enumeration: sudo -l, find / -perm -4000, getcap -r /, cron jobs, and writable PATH directories.

  • A SUID binary with a known GTFOBins entry can spawn a root shell or read and write files as the owner.

  • Kernel exploits require matching kernel version and distribution; verify build and patch level before use.

  • Credential hunting locations: .bash_history, config files, /etc/shadow if readable, Windows registry, SAM, and LSASS.

Quick reference

Reporting

  • Report structure: cover, confidentiality, scope, methodology, executive summary, findings, remediation, and appendices.

  • Each finding includes title, severity, affected assets, description, reproduction steps, evidence, impact, and remediation.

  • Severity ratings use CVSS plus business context; critical findings require immediate client notification.

  • Reproduction steps must be specific enough for a third party to verify without tester assistance.

  • Executive summary states overall risk posture in business language and avoids exploit-level jargon.

  • Evidence references link findings to screenshots, request and response pairs, and log excerpts.

  • Retest section documents remediation verification and residual risk.

  • Report classification and distribution list are agreed in the engagement contract.

Quick reference

Evidence Handling and Remediation Guidance

  • Chain of custody records who collected, handled, transferred, and stored evidence; each transfer is signed and timed.

  • Hash evidence with SHA-256 at collection and verify before analysis; record hash values in the report.

  • Work on copies; preserve original media and capture volatile data first, including memory, connections, and processes.

  • Evidence storage is encrypted, access-controlled, and logged; retention period follows contract and law.

  • Screenshots include timestamp, URL or host, user context, and visible proof of impact.

  • Remediation guidance is specific, prioritized by risk, and includes validation criteria.

  • Retest verifies fixes and documents residual or accepted risk.

  • Destroy or return evidence per agreement; record destruction date and method.

CSPT key terms

Statement of Work

The commercial contract for an engagement that sets out deliverables, schedule, level of effort, assumptions and cost. It defines what the client is buying and what the tester is obligated to produce.

Related: Rules of Engagement, Point of Contact
Point of Contact

The named person on the client side and on the tester side responsible for coordination, approvals and emergency escalation. Testers must know how to reach them at any hour during testing.

Related: Rules of Engagement, Authorization Letter
Non-Disclosure Agreement

A contract obligating the tester to keep client information confidential and restricting how findings, credentials and data encountered during testing may be used, stored or disclosed.

Related: Statement of Work
Rules of Engagement

A written agreement describing how testing will be conducted: allowed techniques, in-scope targets, timing, prohibited actions such as denial of service, and the escalation path when something goes wrong.

Related: Scope Definition, Authorization Letter, Testing Window
Scope Definition

The explicit list of addresses, hostnames, applications, accounts and environments authorized for testing, together with what is excluded. Anything not listed must not be touched, even if reachable.

Related: Rules of Engagement, Authorization Letter
Authorization Letter

A signed statement from someone with authority at the client confirming that testing of the listed assets is permitted during stated dates. It is the document produced if activity is questioned.

Related: Rules of Engagement, Scope Definition
Testing Window

The agreed dates and hours during which testing activity may take place, often set to avoid business-critical periods. Activity outside this window is unauthorized even if the target is in scope.

Related: Rules of Engagement, Scope Definition
Open Source Intelligence

Collecting and correlating publicly available information such as registered domains, employee roles, technology job adverts and leaked credentials to map an organization's exposure without touching its systems.

Related: Passive Reconnaissance, Subdomain Enumeration, Certificate Transparency
Passive Reconnaissance

Gathering information about a target from third-party sources, cached pages and public records without sending traffic to the target's own systems, making it very hard for the target to detect.

Related: Open Source Intelligence, Active Reconnaissance
Active Reconnaissance

Directly querying and probing the target's systems to discover hosts, open ports, services and versions. It produces logs and can trigger monitoring, so timing must respect the rules of engagement.

Related: Passive Reconnaissance, Vulnerability Scanning
Subdomain Enumeration

Discovering hostnames under a parent domain using wordlists, DNS records, certificate logs and search engines. Forgotten development, staging and VPN portals found this way often carry weaker controls.

Related: Certificate Transparency, Open Source Intelligence
Certificate Transparency

Public append-only logs of issued TLS certificates. Searching them reveals hostnames an organization has provisioned, including internal or staging systems that were never linked from public pages.

Related: Subdomain Enumeration, Passive Reconnaissance
Vulnerability Scanning

Automated checks that compare hosts and applications against known signatures and configuration baselines to produce candidate weaknesses. Results are leads, not findings, until a human confirms them.

Related: False Positive, Manual Validation
False Positive

A reported weakness that either does not exist or cannot be exploited in the target's actual configuration. Reporting these without checking damages credibility and wastes the client's remediation effort.

Related: Manual Validation, Vulnerability Scanning
Manual Validation

Human confirmation that a suspected weakness is genuine and reachable, usually by reproducing the condition carefully and recording the exact request, response and result.

Related: False Positive, Proof of Concept
CVSS

A standardized framework that scores vulnerability severity from metrics such as attack vector, complexity, privileges required and impact. Scores help compare issues but do not replace business context.

Related: Risk Rating
Proof of Concept

A minimal, controlled demonstration that a weakness can actually be exploited, showing the precise steps or request and the resulting evidence while avoiding damage or data loss.

Related: Manual Validation, Evidence Capture
Pivoting

Using a compromised host as a relay point to reach network segments and services that are not directly reachable from the tester's own machine.

Related: Port Forwarding, SOCKS Proxy, Lateral Movement
Port Forwarding

Redirecting a local port through a compromised host to a service on an internal network, so ordinary tools can interact with the internal target as though it were local.

Related: Pivoting, SOCKS Proxy
SOCKS Proxy

A proxy tunnel built over an existing foothold that relays arbitrary TCP connections through the compromised host, letting scanning and client tools reach many internal addresses at once.

Related: Pivoting, Port Forwarding
Lateral Movement

Moving from one compromised system to other hosts in the environment using stolen credentials, remote execution services or trust relationships, with the aim of reaching higher-value assets.

Related: Pivoting, Kerberoasting
SQL Injection

Supplying database syntax through unsanitized input so the application executes altered queries, which can bypass authentication, disclose data or modify records.

Related: Cross-Site Scripting, Web Shell
Cross-Site Scripting

Injecting script into content that other users will view, so it runs in their browsers and can steal sessions, perform actions as them or manipulate what they see.

Related: SQL Injection, Web Shell
Insecure Direct Object Reference

Accessing another user's records by changing an identifier in a request, because the application verifies that a user is logged in but not that they own the object.

Related: Proof of Concept
Server-Side Request Forgery

Causing the server to issue requests to destinations the attacker chooses, such as internal services or cloud metadata endpoints, exploiting the server's own network position and trust.

Related: Pivoting
Web Shell

A small script placed on a web server that accepts commands over HTTP, giving the attacker remote execution in the context of the web service account.

Related: SQL Injection, Token Impersonation
Kerberoasting

Requesting Kerberos service tickets for accounts that have service principal names, then cracking the encrypted material offline to recover the service account's password.

Related: AS-REP Roasting, Lateral Movement
AS-REP Roasting

Requesting authentication response data for accounts configured without Kerberos pre-authentication, then cracking it offline to obtain those accounts' passwords without any prior access.

Related: Kerberoasting, BloodHound
BloodHound

A graph analysis tool that maps users, computers, sessions and permissions in Active Directory to reveal the shortest relationship paths from a starting account to highly privileged targets.

Related: ACL Abuse, Lateral Movement
ACL Abuse

Exploiting delegated Active Directory permissions, such as GenericAll or WriteDACL, to reset passwords, add group members or grant rights without cracking any credentials.

Related: BloodHound, DCSync
DCSync

Impersonating a domain controller's replication request to pull stored password hashes for domain accounts, including the krbtgt account, from Active Directory.

Related: ACL Abuse, Lateral Movement
SUID Binary

A Linux executable marked to run with the permissions of its owner rather than the caller. Certain binaries can be abused to execute commands as root.

Related: Sudo Misconfiguration
Sudo Misconfiguration

A sudoers entry that lets a user run particular commands as root, where those commands can be leveraged to read files, spawn a shell or edit permissions.

Related: SUID Binary
Unquoted Service Path

A Windows service whose executable path contains spaces and lacks quotation marks, allowing an attacker who can write to a parent directory to plant a binary that the service runs as its own account.

Related: Token Impersonation
Token Impersonation

Abusing Windows access tokens held by services or logged-on users, often through the SeImpersonate privilege, to execute code in a higher-integrity or more privileged security context.

Related: Unquoted Service Path
Executive Summary

A short, non-technical opening section describing the engagement, the overall risk posture and the most important themes, written for decision makers who will not read the technical detail.

Related: Risk Rating, Remediation Recommendation
Risk Rating

Assigning severity to each finding by combining technical impact with business context such as data sensitivity and exposure, so the client can prioritize remediation sensibly.

Related: CVSS, Remediation Recommendation
Chain of Custody

A documented record of who collected each piece of evidence, when, from where, and every transfer or storage location since, so its integrity can be defended later.

Related: Evidence Capture
Evidence Capture

Recording proof of exploitation, such as timestamped screenshots, command output and HTTP requests, stored securely and referenced in the report so findings can be reproduced.

Related: Chain of Custody, Proof of Concept
Remediation Recommendation

Specific actionable advice on how to fix a finding, covering code or configuration changes, compensating controls where a full fix is delayed, and how the fix can be verified.

Related: Risk Rating, Evidence Capture

CSPT study plan

8 weeks · About 6 hours per week

Week 1

Engagement Management, Scoping and Rules of Engagement & Web Application Exploitation

  • Draft a rules-of-engagement and scoping document for a fictional client that specifies in-scope CIDRs and hosts, testing windows, escalation contacts, data-handling limits, and explicitly prohibited techniques

  • Build an isolated attack-and-defense lab (attacker VM, intentionally vulnerable web application, and a Windows host suitable for later directory work) and verify that snapshot rollback restores a clean state

  • Enumerate the lab web application's authentication, session, and access-control surfaces, mapping each user-controlled input to the trust boundary it crosses

  • Demonstrate three access-control failures (insecure direct object reference, privilege escalation by parameter tampering, forced browsing to a privileged page) and capture the raw request and response for each

Your checkpoint

A two-to-three page scope and RoE document plus a lab build sheet, with three access-control entries logged in a findings register (title, evidence, impact, suggested fix).

Week 2

Web Application Exploitation

  • Exploit injection flaws in the lab (SQL injection, OS command injection, server-side template injection) and record the exact payload, the server response, and the root cause for each

  • Test and document SSRF, unrestricted file upload, and insecure deserialization paths, including any filter or allowlist bypasses that were required

  • Chain a server-side request path to reach an internal-only service and capture the transition from an external web request to an internal response

  • Write remediation guidance for each web finding that names the specific control to implement (parameterized queries, allowlisting, content-type and signature validation) instead of generic advice

Your checkpoint

A web exploitation summary containing five validated findings, each with reproducible proof-of-concept steps and a justified severity rating.

Week 3

Network Exploitation and Pivoting

  • Differentiate bind, reverse, and staged payload behaviour by executing each in the lab and capturing the resulting traffic in a packet capture

  • Demonstrate one network-level credential attack in the lab (relay or hash-based authentication) and document the protocol weakness that made it possible

  • Configure local port forwarding and a SOCKS proxy through a compromised host to prove reachability into a simulated internal segment that was previously unreachable

  • Compare three pivoting methods (SSH tunnelling, proxychains, and a lightweight tunnel such as Chisel) and document the stability, throughput, and detection tradeoffs of each

Your checkpoint

A pivot-chain diagram with the exact commands used, plus a packet capture or proxy log proving traffic reached an internal-only host.

Week 4

Identity and Directory Service Attack Paths

  • Enumerate an Active Directory lab domain: users, groups, computers, trusts, SPNs, and delegation settings, using both native utilities and a third-party collection tool

  • Execute at least two credential-based attack paths (Kerberoasting, AS-REP roasting, or ACL abuse) and capture the artefacts obtained from each

  • Trace a path from a low-privilege domain user to Domain Admin and annotate every edge with the specific misconfiguration that enables it

  • Write a targeted remediation for each edge in the traced path (delegation removal, ACL cleanup, password policy change, administrative tiering)

Your checkpoint

An annotated attack-path graph from low-privilege user to Domain Admin, plus a table mapping each hop to its misconfiguration and its fix.

Week 5

Reconnaissance and Information Gathering & Vulnerability Analysis and Validation

  • Perform passive and active reconnaissance against a practice range: DNS and OSINT collection, subdomain enumeration, service and version scanning, and content discovery

  • Build a consolidated asset and service inventory from the collected output and rank it by exposure and business context

  • Triage automated scanner findings by manually confirming or refuting at least five items, and record the false-positive rate of the scanner used

  • Write a reusable validation checklist that defines reproducibility steps, required evidence, affected scope, and explicit false-positive exit criteria

Your checkpoint

A reconnaissance inventory plus a validation worksheet covering at least ten candidate findings with confirmed or refuted status and supporting evidence.

Week 6

Privilege Escalation (Windows and Linux)

  • Identify and exploit three Windows privilege escalation vectors (service misconfiguration, unquoted service paths, and token or impersonation privilege abuse) and document each successful escalation

  • Identify and exploit three Linux privilege escalation vectors (SUID binaries, sudo misconfiguration, and scheduled-task or writable-path abuse)

  • Run at least two local enumeration tools and manually verify every candidate they surface before acting on it

  • Document the detection signal each escalation technique would generate for a defender monitoring the host

Your checkpoint

A privilege escalation playbook containing six verified techniques (three per operating system), each with the pre-check command, exploitation steps, evidence, and a detection note.

Week 7

Reporting, Evidence Handling and Remediation Guidance & Engagement Management, Scoping and Rules of Engagement

  • Convert the lab findings into a client-ready report containing an executive summary, technical findings, severity ratings, and a prioritized remediation roadmap

  • Apply a consistent evidence-handling process across the engagement: hashing, timestamping, redaction of sensitive data, and storage that preserves integrity

  • Write remediation guidance for every finding that names a responsible owner, an alternative compensating control, and a verifiable retest criterion

  • Review your own report against an engagement-close checklist covering scope coverage, evidence completeness, and justification of each rating

Your checkpoint

A complete redacted sample report for the lab engagement with an executive summary, eight to ten findings with evidence, and a remediation roadmap.

Week 8

Engagement Management, Scoping and Rules of Engagement & Reconnaissance and Information Gathering & Vulnerability Analysis and Validation & Network Exploitation and Pivoting & Web Application Exploitation & Identity and Directory Service Attack Paths & Privilege Escalation (Windows and Linux) & Reporting, Evidence Handling and Remediation Guidance

  • Complete timed practice sets across all eight domains and log every missed item with its root cause (knowledge gap, misread question, or tooling error)

  • Re-demonstrate the three weakest objectives from the log until each is performed twice without reference material

  • Deliver a fifteen-minute verbal walkthrough of one full engagement lifecycle, from scoping through reporting, without notes

  • Audit your study notes for accuracy on any fact you flagged as uncertain and mark unresolved items for verification before exam day

Your checkpoint

A scored timed mock assessment (target 80 percent or higher) plus a one-page weak-area log listing a remediation action for each missed objective.

Milestones

Week 2: Explains and demonstrates an injection finding end to end, from payload to root cause to remediation, without consulting notes.

Week 4: Demonstrates at least one complete chain in the lab, from initial web access through a network pivot to a domain-level objective, without reference material.

Week 6: Completes a full lab chain, reconnaissance through privilege escalation, inside a single three-hour sitting while capturing evidence as you go.

Week 8: Scores 80 percent or higher on the timed mock assessment and has a written remediation action for every missed objective.

CSPT exam day

Your checklist

Government-issued photo ID

Bring the original ID (not a photocopy) whose name matches your registration exactly, including middle names and suffixes. If the name differs, resolve it with the provider before test day rather than at check-in.

Appointment confirmation / admission details

Have your confirmation email, registration number, and the exact start time (with time zone) accessible on your phone and printed as backup.

Platform login credentials

Know your username and password and have any second-factor method available. Do not rely on a browser's saved password if you are testing on a different machine.

Completed system / equipment readiness check

Run whatever compatibility or equipment test the provider offers at least 72 hours ahead, and repeat it the day before on the same device, network, and location you will use for the exam. If no such tool is offered, confirm requirements with the provider.

Private, quiet, well-lit room (remote delivery)

Clear the desk of papers, phones, and second screens. Tell household members your exam window so they do not enter. Keep the door closed.

Webcam, microphone, and stable internet

If your exam is remotely proctored, confirm the camera and mic are working and that you can position the camera as directed. A wired connection or a backup hotspot reduces risk.

Charged laptop and power adapter

Start at full charge with the charger plugged in, and confirm the cable cannot be kicked loose. For in-person delivery, confirm whether power outlets are available.

Permitted materials only

Most certification exams allow no notes, books, or reference sheets, and some allow a provided or blank scratch material. Verify the exact permitted-items list with the provider before you sit; do not assume.

Water and a light snack for scheduled breaks

Only if the delivery format permits it. For remotely proctored exams, confirm with the provider whether leaving the camera view during a break is allowed.

Plan for technical failure

Save the provider's support phone number and your appointment number offline (written down, not only in email) so you can reach support if the session drops.

Arrival/check-in buffer

Plan to be at the test center or logged in well before your start time to absorb ID verification and check-in. Confirm the specific early-arrival window with the provider.

Rest and logistics

Protect sleep in the 48 hours before the exam, confirm your route or workspace, and avoid scheduling anything demanding immediately after the test.

Plan the day

72 hours before

Re-confirm appointment details, delivery mode, permitted items, and the cancellation/reschedule terms. Run the equipment or system check on the exact device and network you will use. Identify your two or three weakest topic areas and do focused practice on those only.

72 hours before (evening)

Take one timed practice set under realistic conditions to calibrate your pacing, then stop. Use the results to decide what to review tomorrow, not to open new material.

24 hours before

Light review only: skim your own summary notes, terminology, and any formulas or frameworks you keep forgetting. Do not start a new topic. Prepare ID, charger, and login details. Set two alarms. Get a normal night's sleep rather than a late cram session.

Morning of

Eat a normal meal, limit caffeine to your usual amount, and avoid last-minute quizzing that can shake confidence. Close background apps and notifications, plug in power, place your ID within reach, and log in early enough to complete check-in without rushing.

First minutes of the test

Read the instructions and any interface tutorial carefully, confirm the timer and question count, and note how to flag and revisit items. Take a few slow breaths before the first question.

During the test

Answer in two passes: first pass handles questions you know quickly and flags anything slow or uncertain; second pass returns to flagged items. Budget roughly your average per-question time and check the clock at quarter points rather than after every question. Answer every item unless the scoring rules tell you otherwise.

Final 10 minutes

Stop opening new reasoning on long items, verify that no questions were left blank due to skipping, and confirm each flagged item has a selection. Submit deliberately and confirm the submission screen before closing the browser.

Rules to check

  • Bring the original government-issued photo ID you registered with. The name must match your registration; if it does not, contact the provider before test day to correct it.

  • Assume no outside materials, notes, phones, smart watches, or other people are allowed in the testing environment unless the provider explicitly states otherwise. Confirm the permitted-items list with the provider in advance.

  • Breaks are governed by the delivery format and proctor instructions. Do not leave the camera view, the testing room, or the test center without permission; confirm the break policy with the provider beforehand.

  • Scoring rules vary by exam. Do not assume that wrong answers are penalized or that partial credit exists; confirm the scoring and pass-mark policy with the provider before the exam, and answer every question if there is no penalty.

  • A technical failure is not automatically your fault or automatically excused. Note the time, capture a screenshot if possible, contact proctor or support immediately, and follow their instructions rather than restarting the session on your own.

  • Exam content is confidential. Do not record, screenshot, reproduce, or discuss specific questions afterward; doing so typically violates the candidate agreement and can void your result.

  • If you need an accommodation (extra time, separate room, medical or language needs), request it well before test day and get written confirmation. Accommodations are rarely granted at check-in.

  • Read each question fully and watch for qualifiers such as 'best', 'primary', 'except', and 'least'. If you must guess, eliminate options first and commit to one answer rather than leaving the item blank.

CSPT FAQ

How many CSPT practice questions are available?

The practice set has 200 CSPT questions. Each one includes the choices, the correct answer, and an explanation, and the set is free to read with no signup.

How many CSPT flashcards are available?

The flashcard set has 32 CSPT cards. Each card shows the term and its definition.

Which topics does the CSPT study material cover?

The CSPT study material has 8 sections, including Engagement Management, Scoping and Rules of Engagement, Reconnaissance and Information Gathering, and Vulnerability Analysis and Validation.

What format facts does the catalog list for the CSPT exam?

The catalog lists these item types: multiple choice for the CSPT exam.