Engineering hubs in Dehradun & Bengaluru · Delivering across 10 countries

nitesh@redcubical.com +91 90687 14658

REDCUBICALSYSTEMS

Services / Protect

Security engineering and compliance evidence that survives an audit

We do the technical work security programmes are made of: threat modelling, secure SDLC gates, penetration testing, cloud posture hardening, vulnerability management with real service levels, and evidence preparation for SOC 2, ISO 27001, HIPAA, GDPR, the India DPDP Act 2023 and PCI DSS. We are not a certification body and not a licensed auditor, and we are explicit about where our role ends.

  • STRIDE threat models produced with your engineers, not delivered to them
  • Vulnerability SLAs by CVSS severity, measured from triage
  • Evidence collection automated so an audit is a query, not a scramble
  • Explicit boundary: we prepare and remediate, we do not certify

At a glance

Testing scopes
Web, API, cloud configuration, network, mobile, social engineering
Frameworks supported
SOC 2, ISO 27001, HIPAA, GDPR, DPDP 2023, PCI DSS
CI security tooling
Semgrep, CodeQL, Trivy, OWASP ZAP, gitleaks, Checkov
Cloud posture
AWS Security Hub, GuardDuty, Config, Azure Defender, GCP SCC
Critical remediation target
24 to 72 hours from confirmed triage
GDPR breach notification
72 hours to the supervisory authority
What we are not
A certification body, a licensed auditor, or an insurer

Design

Threat modelling with STRIDE, walked through

The cheapest security work happens before code exists. A threat model is a structured argument about how your system could be abused, and it takes half a day per bounded context.

How STRIDE works

We draw a data flow diagram with your engineers, mark trust boundaries, then walk each element against six threat categories: Spoofing, Tampering, Repudiation, Information disclosure, Denial of service and Elevation of privilege. Each identified threat gets a mitigation, an owner and a test. The output is a backlog, not a document.

  1. Draw the data flow diagram

    External entities, processes, data stores and data flows, with trust boundaries marked where data crosses from less trusted to more trusted. Drawn live with the engineers who built it, which regularly surfaces a flow nobody had documented.

    Half a day per bounded context

  2. Spoofing: can an attacker be someone else?

    Authentication of users, services and devices. Session and token handling, service-to-service identity, webhook sender verification. Typical mitigations are multi-factor authentication, short-lived tokens with rotation, mutual TLS between services, and signature verification on inbound webhooks.

    Mitigation: prove identity at every boundary

  3. Tampering: can data or code be altered?

    Integrity of data in transit and at rest, request parameters, client-supplied prices and quantities, and the build supply chain. Mitigations include server-side validation of everything, integrity checks, signed artefacts with an SBOM, and pinned dependency digests.

    Mitigation: never trust client-supplied values

  4. Repudiation: can an action be denied?

    Audit logging of security-relevant events with sufficient identity, time and detail to reconstruct what happened. Append-only storage, tamper-evident retention, and clocks synchronised across systems so a timeline actually assembles.

    Mitigation: immutable, attributable audit trail

  5. Information disclosure: what can leak?

    Excessive data in API responses, verbose error messages, identifiers that can be enumerated, personal data in logs and analytics, misconfigured storage, and cross-tenant leakage in a multi-tenant model. Mitigations are field-level authorisation, response shaping, log redaction, encryption with managed keys, and tenant isolation tested rather than assumed.

    Mitigation: minimise, redact, isolate

  6. Denial of service: what can be exhausted?

    Rate limits per identity and per endpoint, quotas, expensive query protection, request size caps, GraphQL query depth and cost limits, and back pressure. Also the neglected case: a single tenant consuming shared capacity and degrading everyone else.

    Mitigation: bound every resource

  7. Elevation of privilege: can a boundary be crossed?

    Authorisation on every request, checked server-side against the resource rather than inferred from the interface. Object-level checks to prevent one customer reading another customer records. Least privilege on service roles. This category produces the most severe findings in almost every test we run.

    Mitigation: authorise per object, per request

  8. Rank, mitigate, and convert to tests

    Each threat is scored on likelihood and impact, assigned an owner and a target date, and where possible turned into an automated test. A threat model that does not produce test cases will be out of date within a quarter.

    Output is backlog items, not a PDF

Build

Secure SDLC gates and the tooling behind them

Where security runs in the lifecycle

Security controls sit at the stage where they are cheapest. Design gets a threat model, commit gets secret scanning, pull request gets SAST and SCA, build gets image scanning and artefact signing, deploy gets infrastructure policy checks, and the running system gets DAST, posture monitoring and runtime detection.

Secure SDLC gates by stage
StageControlToolingBlocking?Typical duration
DesignSTRIDE threat model, data classification, privacy impact assessment where personal data is involvedWorkshop with a data flow diagram, threats recorded as backlog itemsGate on architecture sign-off for new contextsHalf a day per bounded context
CommitSecret detection on the diff and on historygitleaks, TruffleHog, provider-side push protectionYes, and triggers a rotation procedure because a committed secret is compromisedUnder 30 seconds
Pull requestStatic analysis on changed files, dependency and licence checks, infrastructure as code policySemgrep or CodeQL, Trivy or Dependabot, Checkov or tfsecYes on new critical and high findings with an available fix2 to 5 minutes
BuildContainer image scan, base image policy, artefact signing, SBOM generationTrivy or Grype, cosign, SyftYes on fixable criticals. Unsigned artefacts never promote1 to 3 minutes
Pre-deployInfrastructure plan policy evaluation, drift detection, privileged change reviewOPA or Checkov against the Terraform plan, with a second reviewer on stateful destroysYes on policy violationUnder 2 minutes
RuntimeCloud posture monitoring, threat detection, anomaly alerting, audit log integrityAWS Security Hub, GuardDuty and Config, Azure Defender for Cloud, Google Security Command CentreContinuous, alerting rather than blockingOngoing

Only gates that finish fast and produce actionable messages survive. A SAST run taking twenty minutes on a full codebase gets disabled within a month. We scope static analysis to changed files, tune rules to remove the noise that trains people to ignore output, and put the expensive full scans on a nightly schedule.

Security scanning tooling by category
CategoryWhat it findsTools we useBlind spots
SAST (static application security testing)Injection patterns, unsafe deserialisation, hardcoded credentials, weak cryptography, framework misuseSemgrep for fast custom rules, CodeQL for deeper data flow analysis, SonarQube where an estate already runs itCannot see runtime configuration or business logic flaws. Produces false positives that must be tuned or the output gets ignored entirely
DAST (dynamic application security testing)Reflected and stored injection, authentication and session weaknesses, missing security headers, exposed endpoints, verbose errorsOWASP ZAP in CI, Burp Suite for manual and authenticated deep testingNeeds a running, authenticated environment. Coverage is limited to what it can reach and crawl. Poor at multi-step business logic
SCA (software composition analysis)Known vulnerabilities in direct and transitive dependencies, container base image CVEs, licence policy violationsTrivy, Grype, Dependabot, Renovate for automated upgrade pull requestsReports vulnerable versions present, not whether the vulnerable code path is reachable. Expect a high volume of findings that are not exploitable in your context
Secret scanningAPI keys, tokens, private keys and credentials in code, history, configuration and container layersgitleaks, TruffleHog, platform-native push protectionCannot detect a secret that has been encoded or split. Detection is not remediation: the secret must be rotated, not merely deleted
IaC and posture scanningPublic storage buckets, permissive security groups, unencrypted volumes, missing logging, over-broad IAM policiesCheckov, tfsec, AWS Config with conformance packs, Security Hub standardsChecks configuration, not intent. A resource can be compliant with every rule and still be wrong for your threat model

Tooling is the easy half. The half that determines outcomes is triage discipline: every finding accepted, fixed or suppressed with a reason, an owner and an expiry. A dashboard with several hundred unreviewed findings gives you no security benefit and gives an auditor a documented failure of your own vulnerability management process.

Test

Penetration testing: scope, report, retest, price

What a penetration test is

A penetration test is a human attempting to compromise a defined scope within agreed rules of engagement. It finds what scanners cannot: chained business logic abuse, broken access control between tenants, privilege escalation and lateral movement. Scope, authorisation, timing and escalation contacts are agreed in writing before any traffic is sent.

Penetration testing scope types and indicative pricing
ScopeWhat is testedTypical effortIndicative price (USD)Prerequisites
Web applicationAuthentication and session management, authorisation and object-level access control, injection, business logic abuse, file handling, multi-tenancy isolation, client-side issues5 to 10 tester-days4,000 to 8,000A stable staging environment with production-like data, one account per role, and written authorisation from the system owner
APIAuthentication and token handling, object and function level authorisation, mass assignment, rate limiting, input validation, resource consumption, schema-driven fuzzing3 to 7 tester-days3,000 to 6,000OpenAPI or Postman collection, credentials per role, and a non-production environment with representative data volumes
Cloud configuration reviewIAM policy analysis and privilege escalation paths, network exposure, storage and database access, encryption and key management, logging coverage, guardrail effectiveness4 to 8 days4,000 to 9,000Read-only audit role across accounts and an architecture diagram
External networkPerimeter service discovery, exposed management interfaces, TLS and certificate configuration, patch currency, default and weak credentials3 to 6 days3,000 to 7,000IP ranges and domains in scope, written authorisation, and a maintenance window if intrusive testing is agreed
Mobile applicationLocal data storage, keychain and keystore usage, certificate pinning bypass, tamper and root detection strength, deep link handling, and the backend API behind the app5 to 9 days per platform5,000 to 10,000 per platformA test build, test accounts, and acceptance that client-side controls will be bypassed as part of the test
Social engineering and phishing simulationSimulated phishing campaign, pretext telephone calls where agreed, and physical access attempts if contracted3 to 8 days3,000 to 8,000Written authorisation from a named executive sponsor, HR and legal sign-off, and an agreed policy on how individual results are handled

Prices are indicative for scoping conversations and move with authenticated role count, endpoint count, environment complexity and reporting requirements. Regulated-scope testing with formal evidence requirements sits above these ranges. We will not quote a penetration test without scoping it, because a fixed price offered blind is almost always an automated scan with a report template on top.

What a real report contains

  • Executive summary in business language: what an attacker could achieve, what data was reachable, and the three things to fix first. Readable by a board, and no more than two pages.
  • Scope and rules of engagement stating exactly what was and was not tested, the test window, and any constraint that limited coverage.
  • Findings with reproduction steps that a developer can follow: exact request, payload, preconditions, and the observed result. A finding nobody can reproduce will not be fixed.
  • Severity with justification. CVSS vector plus a written contextual rating, because a CVSS 7.5 on an internal admin tool and on your public login page are not the same risk.
  • Specific remediation advice. Not "validate input" but the change, at the layer, in your framework.
  • Attack narrative for chained findings, showing how three medium issues combined into a critical outcome. This is what distinguishes a test from a scan.

Retest policy and cadence

  • Retest included for critical and high findings within 90 days of the original report, at no additional cost, with an updated report reflecting closure.
  • Closure requires evidence. A finding is closed when the original attack path fails on retest, not when a ticket is marked done.
  • Annual full test minimum, plus a test after any significant architectural change, new authentication mechanism, or new class of data entering the system.
  • An honest limitation. A clean report means the tester found nothing exploitable within the agreed scope, time and skill applied. It is evidence of diligence, not proof of security, and we will not let it be represented as more than that.

Operate

Cloud posture, identity, secrets and vulnerability management

Cloud security posture

Most cloud breaches are misconfiguration and over-privileged identity, not exotic exploits. We put preventative guardrails in place first so the mistake cannot be made, then detective controls for what slips through, then a vulnerability management process with service levels by severity so findings actually close.

Preventative guardrails

  • Service control policies at the organisation level that make the dangerous action impossible: no public bucket ACLs, no disabling CloudTrail, no unencrypted volume creation, no resource creation outside approved regions.
  • Block public access enforced at account level on object storage, not per bucket. Per-bucket settings get changed by someone in a hurry.
  • IMDSv2 required on every instance. Version 1 of the instance metadata service is how a server-side request forgery becomes credential theft, and requiring version 2 closes that path.
  • Least privilege by construction. Roles scoped to specific actions on specific resources, with permission boundaries and no wildcard administrator role attached to workloads.
  • No long-lived access keys. Federated human access with short-lived credentials, and OIDC-based workload identity for pipelines instead of stored keys.
  • Guardrails as code. Every control expressed in Terraform and policy-as-code, so it is reviewable, versioned and reproducible across accounts.

Identity and secrets

  • Single identity provider with SSO and enforced multi-factor authentication, phishing-resistant where the risk justifies it. One place to disable an account on the day someone leaves.
  • Just-in-time privileged access. No standing production write access for humans. Time-bound elevation with a stated reason, an approver and a full audit record.
  • Quarterly access reviews with evidence retained. This is both a genuinely useful control and one of the most commonly failed audit items.
  • Secrets in a managed store. AWS Secrets Manager, Azure Key Vault or HashiCorp Vault, retrieved at runtime by workload identity. Never in environment files in a repository, never in a container image, never in a CI variable that a build log can echo.
  • Automated rotation for database credentials and API keys, with the application able to handle a rotation without a restart.
Vulnerability management service levels by severity
SeverityCVSS rangeRemediation targetInterim control if a fix is unavailableEscalation
Critical9.0 to 10.0, or any severity with confirmed active exploitation24 to 72 hours from confirmed triageVirtual patching at the WAF, disable the affected feature, network-isolate the component, or take the service offline if data exposure is ongoingImmediate notification to the accountable executive, treated as an incident with a commander
High7.0 to 8.914 daysCompensating control documented, with monitoring specifically for exploitation attempts against that pathWeekly review until closed. Overdue items escalate to the engineering lead
Medium4.0 to 6.930 to 60 days, aligned to the release cycleRisk acceptance with a named owner and an expiry date, never an indefinite acceptanceMonthly security review
Low0.1 to 3.9Next planned release cycle, or accepted with a documented rationaleNot usually requiredQuarterly review

Two adjustments to raw CVSS. First, exposure: an internet-facing system handling personal or payment data moves up a band, an internal tool behind SSO with no sensitive data may move down. Second, exploitability: a published proof of concept or inclusion in a known-exploited catalogue overrides the base score entirely. The clock starts at confirmed triage rather than at discovery, and triage itself has a target of one business day.

Compliance

Frameworks: what each requires, and exactly what we can do

What we can and cannot do

We do the engineering and evidence work: gap assessment, control implementation, automated evidence collection, policy drafting, remediation and audit support. We are not a certification body, not a licensed CPA firm, and not your auditor. Attestation and certification come only from an accredited independent body, and that separation exists for good reason.

Compliance frameworks, requirements and timelines
FrameworkWhat it actually requiresTypical timelineWhat we doWhat we cannot do
SOC 2 Type IAn independent CPA opinion that controls are suitably designed at a point in time, across the trust services criteria you select. Security is mandatory; availability, confidentiality, processing integrity and privacy are optional3 to 5 months from a standing start to audit-readyGap assessment against the criteria, control implementation, policy set, automated evidence pipelines, readiness assessment, and support through fieldworkIssue the report. Only a licensed CPA firm can. We also cannot audit controls we built, which would be a conflict of interest
SOC 2 Type IIThe same criteria examined for operating effectiveness over an observation window, commonly 3 to 12 months. Evidence must show controls ran consistently throughout, not that they existedType I readiness plus the observation window plus 4 to 8 weeks of fieldworkEvidence automation so access reviews, change approvals, vulnerability scans and training records are collected continuously rather than reconstructed laterShorten the observation window, or produce evidence for a period during which the control was not running
ISO 27001:2022A management system, not a checklist. Scope and context, risk assessment methodology, a risk treatment plan, a Statement of Applicability across the Annex A controls, internal audit, management review, and continual improvement6 to 9 months to certification, via a Stage 1 documentation review and a Stage 2 implementation auditISMS documentation, risk assessment and treatment planning, Statement of Applicability, control implementation, internal audit preparation, and support during both stagesCertify you. That requires an accredited certification body. We also cannot act as your independent internal auditor for controls we implemented
HIPAA (US)Administrative, physical and technical safeguards under the Security Rule, Privacy Rule obligations, a documented risk analysis, workforce training, business associate agreements, and breach notification within 60 days3 to 6 months for technical safeguard readinessTechnical safeguards: encryption, access control, audit controls, integrity, transmission security. Risk analysis support, and engineering evidence for your documentationAct as your privacy officer, sign a BAA as if we were the covered entity, or provide legal interpretation of the Privacy Rule
GDPR (EU and UK)Lawful basis for processing, records of processing activities, data protection impact assessments for high-risk processing, data subject rights within one month, privacy by design and default, processor agreements, transfer mechanisms, and 72-hour breach notification2 to 4 months for technical and engineering measuresData mapping, technical measures for the rights machinery including export and erasure, retention automation, consent and preference infrastructure, transfer architecture, and breach detection and reporting capabilityServe as your Data Protection Officer, provide legal advice on lawful basis, or represent you before a supervisory authority
India DPDP Act 2023Consent and notice requirements, purpose limitation, data principal rights including correction and erasure, reasonable security safeguards, breach notification to the Data Protection Board and to affected principals, consent manager integration where applicable, and additional obligations if designated a significant data fiduciary2 to 4 months for technical readiness, subject to ongoing rulesConsent architecture and audit trail, purpose-tagged data inventory, rights fulfilment workflows, retention and deletion automation, security safeguards, and breach notification toolingAct as your Data Protection Officer or grievance officer, or give legal opinion on the Act and its rules as they evolve
PCI DSS v4.0 — SAQ ACard-not-present merchants who fully outsource payment pages to a validated provider with no card data touching their systems. Around 30 requirements, mostly policy, plus scanning of the redirect page4 to 8 weeksConfirm the integration genuinely qualifies, implement the technical requirements, and prepare the self-assessment evidenceSign the Attestation of Compliance on your behalf, or act as a Qualified Security Assessor
PCI DSS v4.0 — SAQ A-EPMerchant websites that do not receive card data but can affect the security of the payment page, typically a JavaScript-based integration. Substantially broader than SAQ A, including scanning, script integrity monitoring and change control2 to 4 monthsSegmentation and scope reduction advice, script integrity controls, quarterly scanning arrangement, and evidence preparationPerform the ASV scan as an approved scanning vendor, or issue the attestation
PCI DSS v4.0 — SAQ DMerchants or service providers that store, process or transmit cardholder data. All twelve requirement groups, roughly 250 to 350 controls depending on applicability, segmentation testing, and annual penetration testing6 to 12 months, and often best avoidedScope reduction is our first recommendation, because moving to a hosted or tokenised integration can take you from SAQ D to SAQ A. Where SAQ D is unavoidable, control implementation and evidenceAct as a QSA, perform the ASV scan, or issue the Report on Compliance

The pattern across all of these: the technical controls are the tractable part. What audits actually fail on is operational consistency, an access review skipped for a quarter, a change deployed without approval, training records missing for three joiners. We automate evidence collection precisely because manually assembled evidence is where the gaps appear.

Honest answer

Incident response, notification clocks, and why nobody can promise you will not be breached

Breach notification timelines

Notification clocks start when you become aware, not when you finish investigating. GDPR gives 72 hours to notify the supervisory authority. CERT-In directions require reporting specified incidents within six hours. HIPAA allows 60 days for individual notice. Those clocks run concurrently, so the decision tree must exist beforehand.

Breach and incident notification obligations
RegimeTriggerDeadlineNotify whomPractical implication
GDPR Article 33 and 34A personal data breach likely to result in a risk to individuals72 hours to the authority from becoming aware. Individuals without undue delay where risk is highThe lead supervisory authority, and affected data subjects where high riskA phased notification is permitted. You are expected to notify with what you know rather than wait for a complete picture
India DPDP Act 2023A personal data breachAs prescribed under the Act and its rules, without a graduated risk threshold in the primary textThe Data Protection Board of India and each affected data principalBroader than GDPR in that there is no low-risk exemption in the statute, so a defensible assessment record matters
CERT-In directions (India)Specified cyber incident types including unauthorised access, data breach and ransomware6 hours from noticing or being brought to noticeThe Indian Computer Emergency Response TeamThe tightest clock in practice. It requires a pre-agreed reporting path and a template ready to go
HIPAA Breach Notification RuleUnsecured protected health information breach60 days to individuals. Annual for breaches under 500 individuals, 60 days for largerAffected individuals, the Department of Health and Human Services, and media where 500 or more in a stateEncryption to the standard removes the notification obligation entirely, which is the strongest practical argument for it
Contractual obligationsAs defined in your customer agreementsFrequently 24 to 48 hours, and often stricter than regulationAffected customers, per contractUsually the binding constraint in practice. Enterprise customers commonly negotiate notification windows tighter than any statute

The single most useful preparation is a written decision tree answering: who declares an incident, who determines whether personal data is involved, who has authority to notify, and who speaks to customers. Without those four answers agreed in advance, the 72-hour clock is consumed by internal escalation rather than by assessment.

The response plan we implement

  1. Detect and triage. Alert reaches on-call with a runbook. Initial severity assigned within 30 minutes against a published matrix.
  2. Declare and assemble. A named incident commander, a scribe, and a communications owner who are not the people investigating. For a suspected data breach, legal and the privacy lead join immediately.
  3. Contain before eradicating. Revoke credentials, isolate the affected component, block the attack path. Preserve evidence before rebuilding anything, because a rebuilt server has destroyed the forensic record.
  4. Assess data impact in parallel. What categories of data, how many individuals, which jurisdictions. This determines the notification clocks, so it starts immediately rather than after containment.
  5. Notify on the clock. Regulator, customers, individuals, insurers and card schemes as applicable, using pre-drafted templates reviewed by legal in advance.
  6. Post-incident review within five working days. Blameless, written, with a timeline, contributing factors and dated actions with owners.
  7. Rehearse twice a year. A tabletop exercise including the notification decision, because the technical response is usually better prepared than the legal and communications response.

Security questionnaire and due diligence support

  • A maintained answer library covering the questions that recur across SIG Lite, CAIQ, vendor-specific forms and enterprise procurement templates, each answer tied to the evidence that supports it.
  • Evidence kept current. Architecture diagrams, data flow diagrams, penetration test summaries, policy documents, subprocessor lists and insurance certificates, versioned rather than rewritten each time.
  • Accurate answers, including the negative ones. Where a control is not in place we say so and attach the plan and date. Overstating a control is a contractual misrepresentation discovered at the worst possible moment.
  • Turnaround. Typically 3 to 5 working days for a standard questionnaire once the library is established, against the two to three weeks it commonly takes without one.

Answers

Cybersecurity and compliance questions

Can you certify us to SOC 2 or ISO 27001?

No. Certification and attestation can only be issued by an accredited certification body or a licensed CPA firm, and we are neither. What we do is the engineering and evidence work that determines whether an audit succeeds: gap assessment against the control set, control implementation, automated evidence collection, policy documentation, remediation of findings, and support during fieldwork. Any firm claiming it can certify you and also build your systems is describing a conflict of interest.

What does a penetration test actually cost?

Indicatively, a focused web application test runs 4 to 8 thousand US dollars for 5 to 10 tester-days, an API test 3 to 6 thousand, a cloud configuration review 4 to 9 thousand, an external network test 3 to 7 thousand, and a mobile application test 5 to 10 thousand per platform. Prices scale with authenticated role count, endpoint count and environment complexity. Anyone quoting a fixed price before scoping is either guessing or selling an automated scan with a report template.

How is a penetration test different from a vulnerability scan?

A scan is automated pattern matching against known signatures and misconfigurations. It is cheap, should run continuously, and finds known problems. A penetration test is a human attempting to chain findings into a real compromise: abusing business logic, escalating between tenants, moving laterally, and reaching data they should not. Scanners do not find broken access control between two customer accounts. Testers do. You need both, and a scan report presented as a penetration test is the most common thing we see mis-sold.

What are the vulnerability remediation SLAs you work to?

Measured from confirmed triage, not from discovery. Critical, CVSS 9.0 and above, or any actively exploited issue: 24 to 72 hours. High, 7.0 to 8.9: 14 days. Medium, 4.0 to 6.9: 30 to 60 days. Low, below 4.0: next planned release cycle. Internet-facing systems handling personal or payment data compress those windows. CVSS is a starting point and is adjusted for exploitability and your actual exposure.

How long does SOC 2 or ISO 27001 readiness take?

From a standing start, expect three to five months to be audit-ready for SOC 2 Type I, then a further three to twelve month observation window before a Type II report. ISO 27001 typically runs six to nine months to certification, including a Stage 1 documentation review and a Stage 2 implementation audit. The variable that dominates is not tooling, it is whether your organisation actually operates the controls consistently, because a Type II examines evidence over a period rather than at a point in time.

What are our breach notification obligations?

Under GDPR, notify the supervisory authority within 72 hours of becoming aware of a personal data breach where risk to individuals is likely, and notify affected individuals without undue delay where risk is high. Under the India DPDP Act 2023, notify the Data Protection Board and each affected data principal, with the Act framing notification without a graduated risk threshold. Sector rules add more: CERT-In directions require reporting specified incidents within six hours. HIPAA requires individual notice within 60 days. These run concurrently, which is why the decision tree must exist before an incident.

Can you guarantee we will not be breached?

No, and nobody honestly can. Security reduces the probability of compromise, reduces the damage when it happens, and shortens detection and recovery time. A well-run programme makes you a harder target than the alternatives and means an incident is contained rather than catastrophic. Any supplier offering a guarantee against breach is either misrepresenting what is achievable or relying on contract language you would not want to test.

We have a customer security questionnaire due. Can you help?

Yes, and this is one of the most common reasons clients come to us. We maintain your answers against the questions that recur across SIG Lite, CAIQ, vendor-specific forms and enterprise procurement templates, keep supporting evidence current, and answer the technical sections accurately. Where an answer is genuinely no, we say no and attach the remediation plan and date. Overstating a control in a questionnaire is a contractual misrepresentation, and it is discovered exactly when it matters most.

Start with a three-week security and compliance assessment

Fixed fee. You get a threat model for your primary context, a cloud posture review with prioritised findings, a secure SDLC gap analysis, a gap assessment against your target framework, and a costed remediation roadmap with severity-based timelines. The output is yours whether or not you engage us for the work.