• SOC 2 Readiness
  • 24/7 incident line
  • 24/7 security operations team
GuardsArm security services

Penetration TestingThink Like an Attacker, Defend Like a Pro

Our penetration testers simulate real-world attacks to identify vulnerabilities in your systems before malicious actors do. Comprehensive testing with actionable remediation guidance.

Call (587) 821-5997
Scoped to your environment24/7 security operationsWorldwide coverage

Penetration Testing Challenges Facing Modern Organizations

Your compliance framework (SOC 2, PCI DSS, ISO 27001) mandates annual penetration testing and the audit is coming

You've patched systems and deployed tools, but you don't know if attackers can still get in

Your last pen test report identified vulnerabilities that were never fully remediated

Comprehensive Penetration Testing

Everything you need to protect your organization, delivered directly by senior security practitioners.

External Penetration Testing

Test your perimeter defenses from the attacker perspective. Identify vulnerabilities in internet-facing systems.

Internal Penetration Testing

Simulate insider threats and compromised credentials to test internal network segmentation and controls.

Application Penetration Testing

Deep security testing of web and mobile applications using OWASP methodology and advanced techniques.

Wireless Penetration Testing

Comprehensive WiFi security assessment including WPA2/WPA3 testing and rogue access point detection.

Vulnerability Assessment

Systematic scanning and identification of security vulnerabilities across your infrastructure.

Social Engineering Testing

Test your human defenses with phishing simulations and pretexting exercises.

Watch

Scope first. Quote second.

Scope First, Quote Second: 5 Decisions Before You Request a Penetration Test

Most penetration test quotes are priced before anyone agrees what is being tested. Five decisions to settle first: objective and scope, starting position, constraints and more.

What You Get With Penetration Testing

Tell us what you need tested, monitored or evidenced, and we will scope it with you.

Identify vulnerabilities before attackers do

Evidence for frameworks that require testing, such as PCI DSS, SOC 2 and HIPAA

Detailed remediation guidance with priorities

Senior testers working to agreed rules of engagement

A written quote based on the agreed scope

Retest of fixed findings, with retest terms agreed in the statement of work

Executive and technical reports included

Testing windows scheduled around your operations

Deliverables and proof

What you receive from a penetration test

Every engagement is scoped and run to written rules of engagement. These are the deliverables, and the published engagements you can read before a call.

Deliverables

  • Executive summary for leadership
  • Technical report: each finding with severity, evidence and remediation guidance
  • Remediation tracker you can work from
  • Debrief with your technical and non-technical staff
  • Verification report after the retest

Retest: after you remediate, we retest the fixed findings and confirm in writing which are closed. Retest timing and terms are agreed in the statement of work.

Choosing a test

Which penetration test do you need?

Penetration testing is not one service. Each type answers a different question about a different part of your environment. Pick a type to see what we test, what it usually finds, and when it is the right choice.

The view from the internet. We start with what anyone can see about you, then try to get in through it.

What we test

  • Discovery of every internet-facing host, subdomain and certificate, including ones missing from your asset list
  • VPN, remote desktop and firewall management interfaces
  • Mail, web and file transfer servers, and their TLS configuration
  • Password spraying against single sign-on and remote access, within agreed limits

What it often finds

  • Forgotten test servers and old marketing sites that are still online
  • Remote access appliances behind on critical patches
  • Admin panels reachable from anywhere, sometimes with default credentials
  • Accounts with weak passwords and no MFA

Choose it when

Every organization needs this at least yearly. It is the test insurers and PCI DSS ask about first, and the one most closely matching how intrusions start.

What we need from you

Your IP ranges and domains, confirmation you own or control them, and a contact during testing.

More on external penetration testing

The view from inside: what someone can do after a phishing click, a stolen laptop or a rogue network port.

What we test

  • Active Directory and Entra ID: privilege paths, delegation, certificate services, password policy
  • Lateral movement between workstations and servers
  • Unpatched internal services and legacy protocols
  • Segmentation between user, server, management and OT networks

What it often finds

  • Paths from a normal user account to domain admin, often in a few steps
  • The same local administrator password on many machines
  • Credentials in scripts, shares and old group policy files
  • Flat networks where a laptop can reach backup servers and payroll

Choose it when

You have an office, data centre or plant network. Run it after external testing, or together with it, at least yearly.

What we need from you

A network port or a small virtual machine on your network, plus one standard user account if you want an assumed breach start.

More on internal penetration testing

Hands-on testing of an application you build or host, following the OWASP Web Security Testing Guide.

What we test

  • Authentication, session handling and password reset flows
  • Access control between users, roles and tenants
  • Injection, file upload, server-side request forgery and other OWASP Top 10 classes
  • Business logic: what the application lets you do that it should not

What it often finds

  • One customer able to read another customer’s records by changing an ID
  • Admin functions that check the menu but not the request
  • Password reset links that can be reused or guessed
  • Sensitive data in responses that the page never displays

Choose it when

Before a major release, at least yearly for customer-facing applications, and whenever a customer’s security questionnaire asks for an independent test.

What we need from you

A stable test or staging environment that mirrors production, and two accounts for each role.

More on application penetration testing

The interfaces behind mobile apps, single page applications and partner integrations, tested against the OWASP API Security Top 10.

What we test

  • Object and function level authorization on every endpoint
  • Authentication tokens: issuance, expiry, scope and revocation
  • Mass assignment, excessive data exposure and rate limiting
  • Undocumented endpoints and older API versions still in service

What it often finds

  • Endpoints that return every field in the database record, not just the ones the app shows
  • Version 1 of an API still running without the fixes made in version 2
  • Fields such as role or account balance that the client can set
  • No limit on login or one-time code attempts

Choose it when

You expose an API to customers, partners or a mobile app, or your web application is a JavaScript front end over an API.

What we need from you

An OpenAPI file or Postman collection, test credentials or API keys per role, and a non-production environment.

A configuration review of AWS, Azure or Google Cloud, followed by targeted testing of anything that looks exploitable.

What we test

  • Identity and access: roles, policies, keys, cross-account trust
  • Storage buckets, snapshots and databases reachable from the internet
  • Network rules, exposed management ports and public endpoints
  • Logging, alerting and key management settings

What it often finds

  • Long-lived access keys belonging to people who have left
  • Roles that let a developer escalate to full administrator
  • Public snapshots or storage containing backups
  • Audit logging turned off in some regions or accounts

Choose it when

Your workloads or customer data live in the cloud, or you are preparing for SOC 2 with a cloud-hosted system.

What we need from you

Read-only access to each account or subscription, and a list of which ones are production.

More on cloud security assessment

On-site testing of corporate and guest Wi-Fi, to answer one question: can someone outside reach your internal network?

What we test

  • Encryption and authentication settings for each network
  • Guest network isolation from internal systems
  • Rogue and look-alike access points
  • Captive portals and pre-shared key strength

What it often finds

  • Guest networks that can reach printers, cameras or servers
  • A shared Wi-Fi password that has not changed in years
  • Unapproved access points plugged into the corporate network
  • Devices that connect to any network with a familiar name

Choose it when

You have offices, clinics, stores or plants with wireless, especially where the car park is within range.

What we need from you

Site access, a facilities contact, and a list of the networks you expect to see.

More on wireless penetration testing

Tests the controls around people, not the people themselves: reporting, verification and what a stolen session can reach.

What we test

  • Targeted phishing that captures credentials or a session token
  • Help desk pretext calls asking for password or MFA resets
  • Response: how quickly reports reach IT and what happens next

What it often finds

  • MFA resets granted on the strength of a name and a manager’s name
  • Phishing reports that sit in a shared mailbox nobody watches
  • MFA that can be bypassed by a proxy phishing kit

Choose it when

Leadership wants to know whether the human side of your controls holds up, or your insurer asks about phishing resilience.

What we need from you

Written approval from leadership, a heads-up to HR, and agreement on what happens to the results. We do not name individuals in the executive report.

More on social engineering testing

We start where a successful phishing email would leave an attacker: a standard laptop and user account. Then we see how far it goes, and whether anyone notices.

What we test

  • What a normal account can read, change and reach
  • Privilege escalation on the endpoint and in the directory
  • Movement towards an agreed objective, such as the finance system or backups
  • Whether your SOC, MDR or EDR raises an alert and what happens next

What it often finds

  • File shares open to every employee that hold HR or finance data
  • Detections that fire but never reach a person
  • Service accounts with far more access than they use

Choose it when

You already test the perimeter regularly and want to know how well detection and response work. It skips the phishing step so the budget goes on what happens after.

What we need from you

A standard build laptop and account, an agreed objective, and a decision on who in your team or SOC knows the test is running.

Black, grey or white box

How much should the testers know?

The less we are told, the more time goes on discovery an attacker would do anyway. The more we are told, the more of your budget goes on finding real weaknesses.
Black box

No prior knowledge

We get a company name or a URL and nothing else. It is the most realistic view of an outside attacker, but a fixed budget spent mostly on reconnaissance covers less ground.
  • Best for: external tests where discovery is the point
  • Trade-off: less depth for the same days
Grey box

Some knowledge and access

We get scope lists, user accounts and basic documentation, the same as a contractor or a customer would have. This is our default, because it reflects how most breaches progress.
  • Best for: applications, APIs and internal tests
  • Trade-off: the usual balance
White box

Full knowledge

We get architecture diagrams, configuration and sometimes source code. It gives the deepest coverage per day and suits critical applications and cloud reviews.
  • Best for: critical apps, cloud, pre-launch
  • Trade-off: less like a real attacker

Scan, test or red team

Vulnerability scan, penetration test or red team?

All three are useful. They answer different questions, at different costs, and they are often confused in quotes.
Comparison of vulnerability scanning, penetration testing and red teaming
AspectVulnerability scanPenetration testRed team
Question it answersWhich known weaknesses exist on these hosts?Which weaknesses can actually be exploited, and how far do they lead?Could a determined attacker reach a goal without us noticing?
How it worksAutomated tool, compared against a database of known issuesPeople, using tools, chaining weaknesses within an agreed scopeA small team emulating a real adversary across people, process and technology
CoverageBroad: everything in range, shallowlyFocused: the agreed scope, in depthNarrow: whatever path leads to the objective
Tests detection?NoOnly if agreed (for example, assumed breach)Yes, that is the point
Typical cadenceMonthly or continuousYearly and after significant changeOnce a program is mature, every year or two
OutputA long list, with false positives to triageConfirmed findings with evidence, impact and fixes, plus a retestA narrative of the attack path and the detection gaps
Best forHygiene between testsCompliance, assurance and finding what mattersTesting a mature SOC

Inside the report

What a finding looks like

The report is what you are paying for, so it should be usable by the developer who fixes the issue and by the executive who funds it. Here is a fictional example, annotated.
Fictional example for illustrationGA-2026-APP-04

1Customer invoices readable by any logged-in customer

2Severity
High7.7 (CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N)
Affected
portal.example-client.ca, /api/v2/invoices/{id}

Description

The invoice endpoint returns any invoice whose numeric ID is supplied, without checking that it belongs to the logged-in customer (broken object level authorization, OWASP API1:2023). IDs are sequential, so every invoice can be enumerated.

3Evidence

GET /api/v2/invoices/104882 HTTP/1.1
Host: portal.example-client.ca
Authorization: Bearer eyJ...[customer A, redacted]

HTTP/1.1 200 OK
{"id":104882,"customer":"[customer B, redacted]",
 "address":"[redacted]","total":4120.00, ...}

4Impact

Any customer, or anyone who registers a free account, can download every customer’s invoices, including names, addresses and amounts. This would likely be a reportable privacy breach.

5Remediation

Enforce ownership server side: load the invoice by ID and the customer ID from the session together, and return 404 when they do not match. Apply the same check in the shared data access layer so new endpoints inherit it. Consider non-sequential identifiers as defence in depth, not as the fix.

6Status:Open. Retest pending.
  1. A title a manager can understand

    It says what can happen, not the name of the vulnerability class. The technical label (broken object level authorization) sits in the body.

  2. Severity with its working shown

    The CVSS vector lets your team check our scoring. We adjust the narrative for business context, for example when the data is health or payment information.

  3. Evidence you can reproduce

    The exact request and response, redacted, so a developer can reproduce the issue in minutes rather than asking us how we found it.

  4. Impact in plain terms

    Who could exploit it, what they would get, and what it would mean for you: privacy notification, client trust, contractual exposure.

  5. A fix, not a link

    Specific to how the application is built, with a defence-in-depth step and how to check the fix. References to OWASP are there for further reading.

  6. Retest status

    After remediation we retest and update the status in writing. That letter is what auditors and insurers ask to see.

Interactive tool

Build your test scope

Answer nine questions about your environment. You will get the test types we would recommend and why, what can wait, an indicative testing effort, and a summary you can paste into a quote request.

Describe your environment

Rough numbers are fine. Nothing is sent until you choose to send it.

Internet-facing targets

Public IP addresses and domains: websites, VPN gateways, mail servers, remote desktop portals.

Internal network size

Servers, workstations and network devices on office or data centre networks.

Web applications

Customer portals, booking systems, internal apps you build or host. Not SaaS you subscribe to.

APIs

Mobile app back ends, partner integrations, public or internal REST and GraphQL APIs.

Cloud accounts

AWS accounts, Azure subscriptions or Google Cloud projects that run your workloads.

Sites with Wi-Fi

Offices, clinics, stores or plants with their own corporate or guest wireless.

Staff for social engineering

People with company email who could be included in phishing or phone pretext tests.

Last penetration test

When an independent tester last looked at this environment.

Why are you testing?

Pick all that apply.

The engagement

How a GuardsArm penetration test runs

The same five stages for every test type. What changes is the scope and the techniques.
  1. Scope and rules of engagement

    Targets, testing windows, out-of-scope systems and emergency contacts are agreed in writing before any traffic is sent.
  2. Reconnaissance and mapping

    We map what is actually reachable: hosts, services, applications and identities, including anything the asset list missed.
  3. Testing and exploitation

    Findings are confirmed by hand, not copied from a scanner. Exploitation stays inside the agreed rules and stops before it risks production.
  4. Reporting

    An executive summary for the people who fund the fixes and a technical report for the people who make them, each finding with evidence and a remediation path.
  5. Retest

    After remediation we retest the findings and confirm in writing which are closed, so the report can be used as evidence.

How to prepare your team

Most delays in a test come from access that was not ready on day one. Two weeks before the window opens:

  • Name one technical contact who can answer questions and pause testing if needed.
  • Tell your SOC, MDR or IT provider, unless the rules of engagement say detection is being tested.
  • Whitelist nothing by default. If you do allow-list our source IPs, tell us, so the report says so.
  • Prepare test accounts (two per role for applications) and check they work before day one.
  • Take or confirm backups of any fragile system in scope, and tell us which systems are fragile.
  • Freeze major changes to in-scope systems during the testing window where possible.
  • Book the debrief and a remediation owner now, so findings do not wait for a meeting.

Rules of engagement checklist

Agreed in writing before any traffic is sent. If a provider does not ask you about these, ask them.

  • Targets in scope: IP ranges, hostnames, applications, accounts and cloud accounts
  • Explicitly out of scope: systems, third parties and techniques
  • Testing windows, including any after-hours work, and time zone
  • Source IP addresses we will test from
  • Emergency contacts on both sides, with phone numbers
  • Stop conditions: what makes us pause and call you
  • Handling of critical findings: we report them immediately, not at the end
  • Data handling: what evidence we keep, how it is stored, and when it is destroyed
  • Written authorization from someone entitled to give it, and from hosting or cloud providers where their policy requires it
  • Whether detection is being tested, and who knows the test is running

Methodology aligned with NIST SP 800-115, Technical Guide to Information Security Testing and Assessment; OWASP Web Security Testing Guide (application and API testing); Penetration Testing Execution Standard (PTES); CVSS for finding severity.

Plan the engagement

How Penetration Testing delivery works

Prepare for a focused scoping conversation. The statement of work confirms the final scope, timing, responsibilities and deliverables.

Scoping inputs

Target inventory
Applications, APIs, domains, IP ranges, environments and the owner of each target.
Test perspective
External or internal access, authenticated roles, test accounts and assumed attacker knowledge.
Rules of engagement
Permitted techniques, prohibited actions, third-party constraints and immediate stop conditions.
Change context
Release freezes, fragile components, backups and any production safety constraints.

Before work starts

Record these decisions with the project owner so delivery does not depend on assumptions.

Authorization owner
Name the person who can approve the work, access and testing methods.
In and out of scope
List systems, locations, accounts, third parties and explicit exclusions.
Data handling
Agree what may be collected, where it is stored, who can access it and when it is deleted.
Escalation contacts
Provide technical, business and emergency contacts, including backups.
Maintenance and testing windows
Confirm permitted times, blackout periods, change constraints and stop conditions.
Report audience
Identify technical, executive, auditor or customer recipients and confidentiality needs.
Acceptance and retesting
Where applicable, record report review, acceptance criteria and retest scope in the statement of work.

Delivery sequence

  1. 1

    Authorize and scope

    Confirm targets, exclusions, test perspective, rules of engagement and communications.

  2. 2

    Validate access

    Check reachability, credentials and test accounts before the agreed testing window.

  3. 3

    Test and verify

    Exercise approved attack paths, validate findings and record reproducible evidence.

  4. 4

    Report and debrief

    Walk through findings, affected assets, remediation priorities and open questions.

  5. 5

    Retest when included

    Check selected fixes against the original findings and record their retest status.

Client responsibilities

  • Maintain written authorization and notify relevant internal and third-party system owners.
  • Provide agreed test accounts, access routes and a technical contact during testing.
  • Keep backups and rollback procedures available, and use the agreed stop mechanism if needed.
  • Route findings to remediation owners and provide fixes or context for any included retest.

Written outputs

  • Agreed scope, exclusions and rules of engagement.
  • Findings with affected assets, evidence, reproduction steps and risk context.
  • Prioritized remediation guidance and an executive-readable summary.
  • Retest results and remaining exposure for findings included in retesting.

Related Services

Vulnerability Scanning

Continuous vulnerability identification and management.

Learn More

Security Risk Assessment

Comprehensive evaluation of organizational security risks.

Learn More

Managed Security Services

Ongoing protection and monitoring of your infrastructure.

Learn More

Penetration Testing Buyer Checklist

Free checklist: scope a test and evaluate providers before you sign.

Learn More

The GuardsArm platform

XDR, EDR, ITDR and email protection in one platform

GuardsArm helps organizations detect, investigate and contain threats around the clock. These are the capabilities our analysts work with, available as a managed service or for your own team to operate.

See the full platform

Penetration Testing: Frequently Asked Questions

What types of penetration testing do you offer?

We test web applications, APIs, external and internal networks, wireless networks and cloud environments. We can also include social engineering scenarios, such as help desk pretext calls, when leadership approves them and the rules of engagement allow it.

How is a penetration test different from a vulnerability scan?

A vulnerability scan is automated and lists known weaknesses. A penetration test is performed by people who chain weaknesses together the way an attacker would, confirm what is actually exploitable, and show the business impact. Both are useful; they answer different questions.

How often should we run a penetration test?

At least annually and after significant changes to your environment, such as a new application, a major infrastructure change or a merger. Some frameworks set their own requirements; PCI DSS, for example, requires penetration testing at least once every 12 months and after significant changes.

Will testing disrupt our production systems?

Testing is planned to avoid disruption. Scope, timing, out-of-bounds systems and emergency contacts are agreed in the rules of engagement before testing starts, and higher-risk activities can be scheduled outside business hours.

What do we receive at the end?

You receive an executive summary for leadership and a technical report with each finding, its severity, evidence and remediation guidance. After you remediate, we retest the fixed findings and confirm in writing which are closed; retest timing and terms are agreed in the statement of work.

How long does a penetration test take?

It depends on scope. A small external test can take two or three tester days; an engagement covering external, internal and a customer-facing application is more often two to four weeks of effort, with modules running in parallel. Reporting follows testing, and the retest is scheduled once you have remediated. The scope builder on this page gives an indicative range, which we confirm after a scoping call.

Should we tell our IT provider, SOC or MDR before the test?

Usually yes, so that alerts are not escalated as a real incident and nobody blocks our source addresses halfway through. The exception is when detection is part of what you want tested, for example in an assumed breach scenario. In that case a small number of named people know, and the rules of engagement say how a real incident is told apart from the test.

Do we need permission from our cloud or hosting provider?

AWS, Microsoft Azure and Google Cloud allow customers to test their own resources without prior approval for most services, within their published policies. Some hosting and managed service providers do require notice. We check the relevant policy during scoping, and the rules of engagement record any approvals.

What is assumed breach testing?

An assumed breach test starts from a standard employee laptop and account, as if a phishing email had already worked. It skips the part attackers usually succeed at anyway and spends the time on what matters next: how far that account can get, and whether your monitoring notices. It suits organizations that already test their perimeter regularly.

Ready to Secure Your Organization?

Request a security assessment and discover how Penetration Testing can protect your business.

Security assessment scoped to your organization
No obligation, no pressure
A written summary of findings and next steps

The request form is at the top of this page. A senior practitioner replies within one business day to scope it with you. For an active incident, call us instead.