Payout & Fund-Transfer
API Security Audit
Attackers have used legitimate bank-API credentials and trusted lender infrastructure to initiate fund transfers that bypassed checks in the loan system. Security Brigade audits the complete payout journey, from sanction and beneficiary approval to API authentication, transaction limits, monitoring and reconciliation. We test whether unauthorised transfers can be prevented, detected and investigated, and issue a signed audit report, as a CERT-In empanelled auditor, for your bank, your board and your regulator.
2 core CERT-In payout advisories · CERT-In empanelled since 2008 · 6,700+ assessments
Can a payout be initiated without a valid sanction?
- 1
Submit a payout requestloan reference with no approved sanction record - 2
Change the beneficiary or amountafter the approval stage - 3
Replay a transfer instructionpreviously authorised
Expected: rejected, recorded and alerted · evidence: request and response, approval state, alert
What changed
Six weeks, three CERT-In advisories and a reported RBI advisory
9 Oct 2026
CERT-In CIAD-2026-0046 · Critical
Threat actors exploiting payment and disbursement APIs for unauthorised fund transfers at NBFCs, digital lenders, fintechs, payment providers and wallet operators. 19 recommended actions.
15 Oct 2026
CERT-In CIAD-2026-0049 · Critical
AI agents used as attack tooling. Root causes include test APIs left reachable from the internet and payment-API keys stored in plaintext.
4 Nov 2026
CERT-In CIAD-2026-0058 · Critical
The campaign, now named CRAFTY FALCON, continued. Organisations that had already remediated were attacked again. Attacks were timed for holidays and weekends. 27 recommended actions.
Nov 2026
Reserve Bank of India (reported)
Press reports describe an RBI advisory to banks and NBFCs on API-based fund transfers, including security audits by CERT-In empanelled auditors. RBI has not published it. If your bank or supervisor has sent you a copy, we scope the audit to it.
5 Nov 2026
RBI Master Directions for non-bank payment system operators
For authorised non-bank payment system operators, paragraph 24 sets API security measures, and the system audit scope includes the security of APIs.
The attack
The payout API was called directly, from a server the bank already trusted
CERT-In’s description is specific. The attackers did not break the bank’s API. They got inside the lender and used it as intended, without the loan.
- 01
A way in
An internet-facing UAT or staging system, a deprecated API that still accepted file uploads, or a long-lived API token hard-coded in source code.
- 02
Across to production
Shared credentials, tokens and network paths between test and production.
- 03
The direct call
The disbursal API invoked from the lender’s own whitelisted infrastructure, skipping loan creation, KYC, sanction, beneficiary verification and per-transaction limits.
- 04
The money moves
A dummy loan reference, a one-rupee test transfer, then rapid high-value IMPS and NEFT transfers to mule accounts across banks.
- 05
The trail is cut
Logs deleted and endpoint protection tampered with.
Source: CERT-In CIAD-2026-0046 and CIAD-2026-0058.
What the audit covers
Seven areas we assess against the CERT-In advisories
API inventory and exposure
What CERT-In recommends
A complete API inventory; deprecated and unused APIs decommissioned; no internet exposure of internal applications.
What we do
We build the inventory from your documentation and from observed traffic, and list every API that should not be reachable.
Caller identity
What CERT-In recommends
Per-session authentication, no long-lived static tokens, MFA, and mutual TLS, signed requests or scoped keys where access relies on a shared IP allow-list.
What we do
We test whether a payout call can be made with a stolen or replayed credential, and from where.
The payment journey
What CERT-In recommends
Transaction stages bound cryptographically so the debit API cannot be called alone; every payout validated against the sanctioned record; maker-checker; beneficiary validation; transaction, daily and velocity limits.
What we do
We walk the journey end to end and attempt each bypass: a payout with no sanction, a changed beneficiary, a changed amount, a replayed instruction.
Secrets and source code
What CERT-In recommends
Secrets in a vault or HSM with rotation; secret scanning across repositories and CI/CD; branch protection and commit signing.
What we do
We review how API keys, tokens and certificates are stored, rotated and revoked, and search code and configuration for exposed ones.
Environment and application integrity
What CERT-In recommends
UAT and staging isolated from production, with no shared credentials, keys, databases or network paths; cloud management and serverless access restricted; outbound connections denied by default; client applications treated as untrusted, with every rule enforced on the server.
What we do
We test whether a foothold in a non-production system or a cloud management service reaches production, and whether a modified app can skip a server-side check.
Monitoring and reconciliation
What CERT-In recommends
Real-time transaction monitoring with automated hold; near-real-time reconciliation against bank statements; lower thresholds on holidays; logs kept 180 days and protected from tampering.
What we do
We review your rules against the pattern CERT-In describes and, where test data and system access permit, test whether they fire and how quickly a mismatch reaches a person.
Incident readiness
What CERT-In recommends
Suspend the API, revoke and rotate credentials, and verify eradication before resuming.
What we do
We review the playbook against the sequence CERT-In sets out, and who is told, when.
The applicable checks depend on the architecture, banking integrations and evidence available. Our report identifies what was tested, what was reviewed and what could not be verified.
Every CERT-In recommended action, with the evidence to collect: the payout API control checklist →
How it runs
An audit and a technical test, reported separately
The controls audit
We review the design and operating effectiveness of controls across each agreed payout integration. Findings identify the affected process, evidence, risk, responsible owner and corrective action. Closure evidence and our validation are recorded separately.
The technical security test
We test the APIs and transaction flow against the OWASP API Security Top 10 and payment-specific abuse cases, including out-of-sequence payouts, beneficiary and amount manipulation, credential misuse and replay. Testing follows an agreed scope and rules of engagement, normally using a controlled test environment.
How they fit together
The two may be commissioned together or separately. Where technical testing is necessary to establish whether a control works, the audit conclusion reflects those results. Untested or unavailable evidence is clearly identified in the report.
Who it is for
Anyone who can move money through a bank’s API
NBFCs and digital lenders
Using a bank’s payout API or a host-to-host connection for disbursement and refunds.
Payment aggregators and other payment system operators
Whose annual system audit covers API security. RBI PA-PG audit →
Lending service providers and loan-system vendors
That sit inside a lender’s payment journey.
Banks
That offer API fund transfers and need their API clients independently audited.
Deliverables
What you receive
Scope map
Every application, server, interface and third party in the payment journey, per bank integration.
Gap assessment report
Each control area with its status and the evidence behind it.
Remediation tracker in Lemon
Owner, target date, closure evidence and our validation for every gap.
Technical test report
Reproducible findings with request and response, severity and the fix.
Audit report
Signed by Security Brigade as a CERT-In empanelled auditor: findings and supporting evidence for the agreed scope, with the status of open and remediated observations. Separate reports for individual bank integrations when required.
Management summary
A short paper for the board and the bank.
Between audits
The audit is a point in time; the exposure is not
FAQ
Questions lenders are asking
Has RBI mandated an API security audit for NBFCs?
Who can conduct the audit?
We had a VAPT this year. Is that enough?
Do you test in production?
Does it cover our lending service provider and the company hosting our loan system?
How is this different from API security testing?
We are a SEBI-regulated entity. Does this apply?
Related
Adjacent services and frameworks
API Security Testing
The technical test: OWASP API Top 10, business logic and payout-specific cases.
RBI cybersecurity for NBFCs
Which chapters of the 2026 Directions bind your layer.
RBI PA-PG audit
The annual system audit for payment aggregators.
SEBI CSCRF
API security requirements for SEBI-regulated entities.
Find out whether your payout API can be called without the loan.
Talk to the team that audits payment systems for banks, NBFCs and payment companies.
Request a scoping call