Cyber Security | VAPT Services India
VAPT Services in India: How to Choose a Provider That Finds Real Issues
The Indian market for VAPT has grown faster than its quality control. Pricing for what is nominally the same engagement can vary by a factor of eight, and the cheapest proposals frequently describe an automated scan with a logo on the cover. Buyers who cannot tell the difference end up paying for a document that satisfies an auditor and protects nothing. Choosing a cybersecurity partner well is mostly a matter of knowing which questions expose the gap.
Part of the confusion is structural. Vulnerability assessment and penetration testing are different activities that the Indian market has fused into one acronym. Assessment is broad and largely automated, enumerating known weaknesses across a defined estate. Penetration testing is narrow and manual, with a skilled tester attempting to chain weaknesses into actual compromise. A proposal that charges for both and delivers only the first is the most common failure mode, and understanding the VAPT types on offer is the first defence against it.
Table of Contents
- Start with why you are testing
- CERT-In empanelment, and what it does and does not tell you
- The questions that separate testing from scanning
- Engagement types compared
- Reading a proposal properly
- Agree the rules of engagement in writing
- Testing production without breaking it
- Timing and cadence
- Final Thoughts
- Frequently Asked Questions
Start with why you are testing
The right engagement looks very different depending on the driver, and vendors rarely ask.
Testing for an auditor
If an auditor requires it, you need defensible scope, a recognised methodology and evidence of remediation. An ISO 27001 assessor wants to see testing proportionate to the risk assessment. SOC 2 common criteria expect periodic testing with remediation tracked against a defined timeline. PCI DSS is prescriptive, requiring segmentation testing and both internal and external penetration tests at defined intervals. The security safeguards expected under DPDP are less prescriptive but will be judged after an incident against what a reasonable organisation would have done.
If a customer requires it, you usually need a summary letter you can share, which means agreeing deliverables and confidentiality terms before the work starts rather than after.
If you are testing because you genuinely want to know your exposure, scope should follow risk rather than convenience, and you should expect findings that are inconvenient to fix.
CERT-In empanelment, and what it does and does not tell you
For Indian buyers, empanelment by the Indian Computer Emergency Response Team is the most meaningful single credential. CERT-In maintains a list of empanelled information security auditing organisations, and several Indian regulators and public sector tenders require an empanelled auditor.
Empanelment confirms that an organisation met CERT-In's criteria at the time of assessment. It does not guarantee that the specific tester assigned to your engagement is strong, that the methodology suits your application, or that the report will be useful. Treat it as a necessary filter in regulated sectors and a positive signal elsewhere, then keep asking questions.
The questions that separate testing from scanning
- Who will actually perform the test, and what are their credentials? Ask for names and certifications such as OSCP, OSWE or CREST, not a company-level capability statement.
- What proportion of the engagement is manual? A credible answer for a web application test is most of it.
- Which methodology will you follow? Expect a specific reference to the OWASP Web Security Testing Guide, OWASP ASVS, OWASP MASVS for mobile, the OWASP API Security Top 10, PTES or NIST SP 800-115.
- Will testing be black box, grey box or white box? Grey box, with credentials for each user role, finds far more in the same number of days than unauthenticated testing.
- How many person-days are allocated, and to which components? Pricing by application count rather than by effort usually signals automation.
- Is business logic in scope? Authorisation flaws and broken access control between user roles are rarely found by tools and are the findings that matter most.
- Is a retest included, and for how long after the initial report?
- How are findings rated? CVSS version 3.1 or 4.0 provides a defensible baseline, adjusted for your actual context.
Engagement types compared
| Type | What it covers | Typical effort | When it fits |
|---|---|---|---|
| External network | Internet-facing infrastructure, exposed services | 3 to 5 days | Annual baseline, compliance |
| Web application | Authenticated and unauthenticated application testing | 5 to 12 days | Customer-facing SaaS, before major releases |
| API | REST and GraphQL endpoints, authorisation between roles | 4 to 8 days | API-first products, fintech |
| Mobile application | Android and iOS binary, storage, transport, backend | 5 to 10 days | Consumer apps handling personal data |
| Internal network | Post-compromise lateral movement, privilege escalation | 5 to 10 days | Mature programmes, regulated sectors |
| Cloud configuration | IAM, storage exposure, network policy, logging | 3 to 7 days | Any significant AWS, Azure or GCP estate |
| Red team | Objective-based simulated attack across all vectors | 15 to 30 days | Organisations with an established blue team |
Most Indian mid-market companies get the best value from an annual external network test combined with application or API testing scheduled around release cycles. A cloud security configuration review often surfaces more exploitable exposure than a network test, because misconfigured storage and over-permissive identity policies are far more common than unpatched perimeter services.
Reading a proposal properly
Fixed-price quotes with no stated effort should be treated with suspicion. So should proposals that promise a specific number of findings, since a tester who guarantees results before looking is describing a template. Watch for scope defined by IP address count alone, which ignores the fact that one complex application can absorb more effort than fifty static hosts.
Check what happens after the report. A provider who delivers findings and disappears leaves you with the hardest part. The valuable engagements include a walkthrough with your engineering team, clarification during remediation, and a retest that produces an updated report showing closure.
On VAPT cost, the honest framing is that price tracks tester days. A genuinely skilled tester working for eight days cannot be bought for the price of a scan, and a quote far below market usually means fewer days, less experienced testers, or tooling output passed off as analysis.
Agree the rules of engagement in writing
The engagements that go wrong in India usually go wrong on logistics rather than on technical skill. A short rules of engagement document, signed before testing starts, prevents almost all of it.
It should fix the testing window, including whether out-of-hours testing is permitted and whether weekend work is expected. It should name the authorised contacts on both sides and the escalation path if something breaks. It should state explicitly which systems are in scope by hostname, IP range and application, and which are excluded, since shared hosting and third-party SaaS frequently sit outside your authority to test.
What the document should cover
Authorisation matters more than teams expect. If your application runs on AWS, Azure or GCP, the provider's testing activity is governed by that platform's policy, and certain categories such as denial of service testing are prohibited outright. If a third party operates part of your estate, you need their written permission before anyone touches it. Testing a system you do not have authority over exposes both sides.
Finally, agree what happens if the tester finds something critical mid-engagement. A serious live exposure should be reported within hours, not held back for the final report two weeks later. Put that obligation in writing with a defined timeline.
Testing production without breaking it
Many Indian companies test staging because it feels safer, then discover the results were worthless because staging diverged from production months ago. Different data, different configuration, different integrations and often different authentication produce findings that do not reflect real exposure.
Mitigations that work
Where testing production is unavoidable, the practical mitigations are well understood. Use dedicated test accounts at each privilege level so activity is attributable and reversible. Agree rate limits so testing does not look like an attack to your own monitoring. Brief the operations team so they do not spend a day responding to an incident you commissioned. Take a verified backup first, and confirm the restore works rather than assuming it.
If you must test a non-production environment, insist it be a current clone rather than a long-lived staging instance, and have the provider state in the report which environment was tested and what that limits.
Timing and cadence
Test before an audit window opens rather than during it, since findings discovered mid-period become exceptions you have to explain. Test after significant architectural change rather than on a calendar date that no longer reflects your estate. For products shipping continuously, an annual deep test supplemented by targeted testing of major features works better than one large exercise that is stale within a quarter.
Build remediation time into the plan. A report delivered two weeks before a customer deadline leaves no room to fix anything, which defeats the purpose of commissioning it.
Final Thoughts
To scope testing that matches your regulatory obligations and your actual architecture, talk to our consulting team. We will tell you what needs testing, what does not, and what it should reasonably cost.
Frequently Asked Questions
What is the difference between vulnerability assessment and penetration testing?
Vulnerability assessment enumerates known weaknesses broadly, mostly through automated tooling. Penetration testing is manual work in which a tester attempts to exploit and chain weaknesses to demonstrate real impact. VAPT engagements should include both, with the manual component dominant.
Does my VAPT provider need to be CERT-In empanelled?
It is required for many Indian regulated sectors and public sector tenders, and it is a useful quality filter elsewhere. It is not a substitute for checking the credentials of the individual testers assigned to your engagement.
How often should we run VAPT?
At least annually for most organisations, and after any significant change to architecture, authentication or infrastructure. PCI DSS and several sectoral regulators specify their own minimum frequencies.
How long does a VAPT engagement take?
Testing usually runs five to fifteen working days depending on scope, with reporting adding roughly a week. Retests after remediation typically take two to four days.
Is a VAPT certificate a real thing?
Providers commonly issue a closure letter or certificate after a successful retest, and customers and auditors often accept it. It is a commercial artefact rather than an accredited certification, so the underlying report is what carries weight.
Can we use automated scanning instead?
Scanning has a place in continuous monitoring, but it does not find business logic flaws, broken authorisation between user roles, or chained attack paths. Those are the findings that cause real breaches.








