Physical Security Testing Boundaries

A client has authorized a penetration test and mentions physical security during the kickoff meeting. You arrive at the building and notice an employee holding a controlled door open for several people. Nearby, a badge reader protects another entrance, and a locked cabinet sits…

A client has authorized a penetration test and mentions physical security during the kickoff meeting. You arrive at the building and notice an employee holding a controlled door open for several people. Nearby, a badge reader protects another entrance, and a locked cabinet sits behind reception. What can you test, and what can you conclude from what you see?

The answer doesn’t begin with whichever technique seems most interesting. It begins with the written scope. A penetration test might include networks, applications, facilities, or some combination of them.

General authorization to conduct a penetration test doesn’t automatically authorize you to enter controlled spaces, leave removable media, copy badge data, or manipulate locks. Those activities create safety, legal, privacy, and property risks that can differ from the risks of ordinary network scanning. They need explicit constraints.

CompTIA PenTest+ objective 3.3 includes tailgating, site surveys, USB drops, badge cloning, and lock picking as physical security concepts. Knowing the names is only the beginning. You also need to understand what each concept tests, what evidence would support a finding, and when an action would cross an engagement boundary.

Tailgating means entering a controlled area by following an authorized person without independently presenting the required authorization. Imagine that an employee uses a valid badge, opens a door, and allows someone without a presented credential to follow. The physical barrier worked mechanically. The weakness involves the way access was granted or challenged. Depending on the environment, relevant controls could include visitor handling, individual credential presentation, employee awareness, guards, or monitoring.

An observation of that behavior isn’t automatically permission to repeat it. Watching a door from an authorized reception area may be within scope while attempting to follow an employee through it may not be. The first activity collects evidence about normal behavior. The second attempts to enter a controlled facility. They have different consequences, so the scope must distinguish them.

A site survey is a structured examination of the physical environment within the authorized boundaries. It can consider entrances, reception procedures, visible access controls, visitor movement, credential use, and other relevant conditions. The purpose isn’t to wander wherever curiosity leads. It’s to compare expected protection with observable conditions and identify where stronger validation may be justified.

Suppose the client says visitors must report to reception, receive authorization, and remain controlled while inside. From the approved observation area, you might look for evidence that the process is being followed. Are visitors directed to reception? Are credentials used at controlled entrances?

Are exceptions challenged or recorded? One brief observation can support a limited statement about what happened at that time. It usually can’t establish how the facility behaves every day or at every entrance.

A USB drop involves placing removable USB media in an authorized location to evaluate how people or processes respond to it. The assessment question might concern whether unknown media is reported, ignored, or handled contrary to organizational expectations. It isn’t permission to scatter devices across a client’s property or to expose people and systems to uncontrolled content. Placement locations, device ownership, timing, labeling, recovery, monitoring, and permitted interactions all need to be settled in advance.

Badge cloning means copying or reproducing access-badge data for an authorized assessment scenario. The badge is part of a facility access decision, so copying its data is not a harmless collection exercise. Authorization should identify whose credential may be involved, which reader or area is in scope, when testing may occur, how copied data must be protected, and what happens after the test. Merely noticing that employees carry badges doesn’t establish that the badges can be cloned or that a copied badge would open a door.

Lock picking tests a different part of physical protection. It concerns manipulating a lock rather than presenting the expected key or credential. Even when a lock is an approved target, the tester must consider ownership, the specific door or cabinet, possible damage, safety, occupancy, alarms, and operational effects.

A successful manipulation can affect property and access. An unsuccessful attempt can also cause disruption. That makes explicit permission and a defined stopping condition essential.

These concepts are related, but they don’t produce interchangeable evidence. Seeing one person follow another through a door is evidence relevant to tailgating behavior. It doesn’t prove that the badge system can be cloned.

Observing a mechanical lock doesn’t prove it can be picked. Finding unattended USB media doesn’t show that an employee connected it to a system. A defensible assessment keeps each observation tied to the mechanism it actually supports.

A useful decision process starts with the requirement. What physical control or expected behavior is the client asking you to evaluate? Next, predict the evidence that should exist if that protection is working. Then choose a bounded action that can test the prediction without exceeding the authorization. Finally, compare the result with the prediction and document both the conclusion and its limits.

Consider a statement of work that authorizes observation from the public reception area. It names physical security testing in general, but it doesn’t authorize lock picking, badge cloning, USB drops, or entry beyond reception. The expected outcome is evidence about visible access and visitor practices in that approved area. You can observe whether visitors are directed appropriately, whether controlled doors appear to require credentials, and whether people in the area follow the defined process.

Now suppose you see an employee badge through a door and hold it open for another person. That observation may indicate a possible weakness in individual authorization or visitor control. It does not authorize you to join them. It also doesn’t prove that an unknown person could enter successfully, because you haven’t validated how employees, guards, or monitoring systems would respond to an actual attempt.

The least-disruptive valid action is to record what was observed, including the location, time, applicable procedure, and limits of the observation. If stronger validation is necessary, you request clarification or escalation through the engagement process. The client might authorize a controlled tailgating attempt with a designated tester, entrance, time window, safety contact, and stop condition. Until that approval exists, observation remains observation.

This distinction is part of technical correctness. Choosing not to perform an unauthorized action isn’t a failure to test. It’s the correct result of checking the engagement constraints. A technically possible action can still be the wrong professional decision when the target, timing, ownership, safety conditions, or permitted impact haven’t been established.

Before any physical action, identify the exact target. A building name may be too broad. One entrance might be authorized while another serves a sensitive or safety-critical area.

A cabinet might belong to the client, a tenant, or a service provider. An access badge might be issued to an employee but remain organizational property. If ownership or authority is unclear, don’t infer permission from physical proximity.

Timing matters as well. A test that is acceptable during a controlled window could be dangerous or disruptive during an emergency, a shift change, maintenance, or a high-traffic event. Dependencies matter too. A door can lead to an ordinary office, an equipment room, or an operational environment where availability and human safety carry special weight. The written plan should identify contacts and escalation procedures so the tester knows when to pause and whom to notify.

Prediction keeps the test connected to its purpose. If the control requires every person to present an individual credential, the predicted state is that a person without one won’t be admitted merely by following someone else. If the organization expects unknown USB media to be reported, the prediction concerns that response. If a badge is meant to authorize only specified areas, the prediction concerns whether the approved credential use matches those restrictions. Clear predictions make it easier to select evidence without performing unrelated actions.

Verification then asks whether the collected evidence genuinely supports the proposed finding. Imagine that a tester observes an apparently unsecured door. Before reporting a failed access control, the tester should consider whether the door was actually a controlled entrance, whether it was temporarily authorized to remain open, and whether another control governed entry. The observation is real, but its meaning may still be unresolved.

The same care applies when the expected evidence is missing. A badge doesn’t open an approved test door. That result doesn’t necessarily demonstrate strong security, and it doesn’t necessarily demonstrate a testing failure.

The badge might lack access by design. The wrong door, credential, time window, or prerequisite may have been selected. The tester compares the setup with the authorization and expected conditions before drawing a conclusion or requesting a more intrusive action.

Conflicting evidence is a reason to revisit assumptions, not a reason to expand the test silently. Suppose reception staff appear to enforce visitor controls during one observation, while an employee later holds a controlled door for someone else. Those observations may reflect different entrances, different people, or inconsistent enforcement. Report the distinction. Don’t compress them into a claim that visitor control always works or never works.

Evidence quality also depends on freshness and context. An old photograph, an outdated floor plan, or a description from an employee may help form a hypothesis, but it may not describe current conditions. A direct observation can be stronger for the specific time and place observed, yet it still has limits. Corroboration might come from approved records, access-control information, monitoring evidence, or another authorized observation. Which sources are available depends on the engagement.

Now return to the dangerous assumption that a general penetration-test authorization covers all of these actions. The verification step is simple in principle. Read the governing documents and look for explicit permission, named targets, timing, operational limits, and escalation procedures. A casual reference to physical testing doesn’t fill in missing authority.

Neither does an objective on a certification exam. Exam coverage tells you which concepts to understand. It doesn’t grant permission to use them against a facility.

If the statement of work permits reception-area observation only, lock manipulation remains outside scope. Badge cloning remains outside scope. Leaving USB devices remains outside scope.

Attempting to enter behind an employee remains outside scope. The appropriate next action is clarification, not creative interpretation. That decision protects the client, the tester, employees, visitors, property, and the integrity of the assessment.

Lower-impact validation should come before escalation when it can establish the needed evidence. If approved observation already shows that multiple people routinely enter on one credential presentation, the client may have enough evidence to address the behavior without authorizing an attempted entry. If existing access records can confirm whether individual credentials are required, reviewing those approved records may reduce the need for a more intrusive test. The goal is evidence adequate for the assessment, not maximum physical interaction.

Sometimes stronger validation is justified because observation alone leaves an important uncertainty. In that case, the request should identify exactly what the additional action would answer. For example, an authorized tailgating attempt might test whether an uncredentialed person is challenged. A controlled USB drop might test the response to unknown media.

A badge-cloning scenario might test a defined access decision. A lock-picking test might assess a specific approved lock. Each action needs its own boundaries because each creates different risks.

Results should distinguish what was observed, what was attempted, and what was demonstrated. If an employee held a door open but the tester didn’t cross the threshold, report the behavior that was observed. Don’t report successful unauthorized entry.

If a controlled USB device was recovered untouched, report that outcome and the authorized observation period. Don’t claim that every employee would avoid unknown media. If a copied badge was created in an approved scenario but never tested at a reader, don’t claim that it granted facility access.

A finding becomes more useful when it connects the weakness to the relevant protection. Tailgating evidence may lead the client to examine individual access authorization, visitor procedures, employee awareness, monitoring, or the way exceptions are handled. Badge-related evidence may lead to a review of credential issuance, facility permissions, and credential protection. Site-survey observations may show that a documented process isn’t consistently visible in practice. Lock-related evidence may support a review of the specific physical barrier and the assets behind it.

Remediation should fit the validated condition rather than the most dramatic interpretation. If the evidence concerns employees allowing others through a controlled door, replacing every badge may not address the observed behavior. Visitor control, awareness, monitoring, or enforcement may be more directly connected. If the issue is an access credential being accepted where it shouldn’t be, the client needs to examine facility access authorization and credential permissions. The evidence guides the response.

Reporting also needs operational context. Describe the authorized location and time, expected control, observed behavior, validation performed, and any limitations. Record whether the test was observational or interactive. State whether entry occurred, whether a credential was presented, whether property was manipulated, and whether the result was corroborated. Avoid turning uncertainty into certainty merely to make the finding sound stronger.

Cleanup is part of the engagement decision too. Any approved test items, copied data, temporary credentials, or other assessment artifacts need to be handled according to the agreed constraints. The central point isn’t a particular cleanup procedure. It’s that physical testing can leave objects, data, and access consequences behind, so the engagement must account for them rather than treating the action as complete the moment evidence is collected.

Suppose you must choose between two next steps after seeing inconsistent visitor handling at reception. One choice is to attempt entry through a controlled door immediately because physical testing was mentioned at kickoff. The other is to document the observation, compare it with the expected visitor process, and seek authorization for a defined validation if the existing evidence is insufficient. The second choice fits the evidence and the scope. The first relies on an authority that hasn’t been established.

What additional information could change that decision? Explicit written authorization for the attempted entry could change it, provided the target, timing, safety limits, monitoring expectations, stop conditions, and escalation contacts were also clear. A broader technical goal alone wouldn’t be enough. The permission must match the actual physical action.

That reasoning applies across all five concepts. First identify the control being assessed. Predict what successful protection would look like.

Select the least-impactful authorized action capable of testing it. Verify that the result supports the conclusion, and preserve the distinction between an observation and a demonstrated outcome. The resulting report can then explain not only what happened, but why the evidence matters and what remains uncertain.

Physical security testing isn’t defined by how far a tester can get. It’s defined by whether the evidence answers an authorized question. Tailgating, site surveys, USB drops, badge cloning, and lock picking examine different controls, so an observation about one can’t prove a weakness in another. Start with the written boundary, predict the expected protection, and use the least-disruptive valid action.

Then report exactly what you observed or demonstrated, without turning uncertainty into a stronger claim. When physical access, credentials, people, and property are involved, authorization and safety aren’t administrative details added to the technical work. They’re part of what makes the assessment technically sound and professionally defensible.

Pause and decide: What can you test, and what can you conclude from what you see? Consult the companion episode reference for exact terminology, named tools or standards, source-derived evidence checkpoints, and review material for Physical Security Assessment Boundaries. This lesson was produced by Bare Metal Cyber dot com.

Physical Security Testing Boundaries

headphones Listen Anywhere

More Options »
Broadcast by