A procurement team at a Bengaluru fintech asks three vendors for a “VAPT quote”. One sends a price for an automated scan of 200 IP addresses. Another quotes ten person-days of manual testing on the mobile app. The third proposes a six-week objective-based exercise against the whole company. All three called it VAPT. The confusion is common in India, where Vulnerability Assessment and Penetration Testing are routinely bundled into one acronym, and it hides the fact that these activities answer different questions.

Vulnerability scanning: breadth, repeatedly

A vulnerability scanner is software that inventories what it can see and compares it with a database of known weaknesses. It identifies a service, works out the product and version from banners, responses or installed package lists, and checks whether that version is affected by published CVEs. It also flags configuration issues such as weak TLS settings, default pages, or missing security headers.

The strengths are coverage and repeatability. A scanner can check thousands of hosts every week without getting bored, and it gives a consistent baseline so you can see whether things are improving. Its weaknesses follow from the method:

  • It mostly infers vulnerability from version numbers. If a Linux distribution backports a security fix without changing the main version string, the scanner may still report the flaw, which is a false positive.
  • It only knows what has been published. A logic flaw unique to your application, such as letting a customer change the account number in a request and view someone else’s statement, will not appear in any signature database.
  • It judges findings one at a time. It cannot see that three medium issues chain into a full compromise.

Scans can be unauthenticated, seeing the system as an outsider would, or authenticated, logging in with credentials to read installed software and settings directly. Authenticated scans find far more and produce fewer guesses, which is why mature programmes use them internally.

Penetration testing: depth, with a human

A penetration test is a time-boxed, authorised attempt by skilled people to actually exploit weaknesses within an agreed scope. Testers often start with automated tools, but the value comes from what they do next: confirming whether a finding is genuinely exploitable, combining small weaknesses into a meaningful path, and probing business logic that no tool understands.

Penetration tests are usually described by how much the tester knows in advance. In a black-box test the tester starts with almost nothing, like an outside attacker. In a grey-box test they get some information, typically user accounts at different privilege levels. In a white-box test they get architecture documents and sometimes source code. Grey-box is the usual compromise: it avoids spending paid days on discovery an attacker would eventually do anyway.

The output is not just a list of CVEs but evidence: what was done, what it led to, and how the business would be affected. A good report might say that a low-privilege support account could reach an internal admin panel, that the panel reused a database credential, and that the credential allowed reading the full customer table. That narrative is something a scanner cannot produce.

The limits are equally real. A pentest is a snapshot from one period and one team’s skill set. It covers the agreed scope and nothing else. A clean report means the testers did not find a way in during those days, not that none exists.

Red teaming: testing the defenders

A red team exercise moves the question from “what is vulnerable” to “would we notice and respond if a capable adversary came after a specific goal”. The red team is given an objective, such as reaching the payment approval system, and is allowed to use realistic techniques including phishing and physical entry if agreed. Typically only a few senior people know it is happening. The detection and response team, the blue team, is being tested as much as the technology.

Red teaming only makes sense once the basics are in place. If routine scans still show hundreds of unpatched critical findings, a red team will simply walk through the obvious holes and the organisation will learn what it already knew, expensively.

Side by side

AspectVulnerability scanPenetration testRed team
Main questionWhat known weaknesses exist?Can these weaknesses be exploited, and how far?Would we detect and stop a real attack?
MethodMostly automatedManual, tool-assistedManual, goal-driven, stealthy
FrequencyWeekly to monthly, or continuousAnnually or per major releaseOccasionally, for mature organisations
CoverageWideAgreed scope, deepNarrow path to an objective
Typical outputList of findings with severityProven attack paths with evidenceDetection and response gaps

How each feeds detection

These activities are not only about fixing things. They also tell the monitoring team what to watch. Scan results show which systems carry known exploitable flaws that cannot be patched quickly, so alerts on those systems deserve a lower threshold. Pentest reports describe concrete techniques that worked, such as reusing a service account across servers, and each one is a candidate detection rule: if testers could do it unseen, so could an attacker. Red team exercises go further by checking whether existing alerts actually fired, and how long it took someone to act on them.

A useful habit after any pentest is to ask, for every successful step, “did our logs record this, and would anyone have looked?” The answers are often more valuable than the vulnerability list.

Practical points that catch people out

Authorisation must be written and specific. Testing systems without permission can be an offence under the Information Technology Act, 2000, regardless of intent. The rules of engagement should name the IP ranges, applications, dates, permitted techniques and emergency contacts. If an application is hosted by a cloud provider or a third party, check whether their policy requires notice.

Production risk is real. Even careful testing can lock accounts or crash fragile services. Agree in advance whether destructive checks are allowed and what time windows to use.

Compliance requirements differ. Payment card standards, sector regulators and customer contracts may each specify how often testing must happen and who may perform it. Read the actual requirement rather than assuming “annual VAPT” covers everything.

Retesting matters. A pentest finding is not closed when a developer says it is fixed. Ask the testers to verify, because partial fixes that block one route but leave a sibling route open are very common.

The simplest way to remember the difference: scanning tells you where the doors and windows are and which ones look unlocked. Penetration testing has someone try the handles and see which rooms they can reach. Red teaming checks whether the guards noticed.

Written by the RP Strativa Team. The company and scenarios in the examples are illustrative.

Leave a Reply

Trending