Fintech and Payment Processors
PSD2 compliance, PCI DSS alignment and open banking API security.

Payment businesses carry two security burdens at once: an attacker who wants to move money, and a scheme or supervisor who wants evidence that controls operate. An assessment that addresses only the first leaves the firm exposed to the second. We scope engagements for fintech and payment clients so that findings arrive framed against both.
PCI DSS scoping. The most expensive mistake in card environments is a cardholder data environment larger than anyone intended. Systems that store, process or transmit card data are obviously in scope; systems that can affect the security of those systems are in scope too, and that second category quietly absorbs shared identity providers, monitoring agents, jump hosts and CI runners. We trace where card data actually flows against where the documented flow says it goes, identify segmentation controls that are assumed rather than enforced, and test those controls rather than accepting a network diagram. Where tokenisation is used to reduce scope, we test whether the token can be exchanged back for a primary account number by a caller who should not be able to do so.
PSD2 and strong customer authentication. SCA implementations fail at the exemption boundary far more often than at the authentication step itself. We test whether low-value, recurring and trusted-beneficiary exemptions can be forced by a caller who is not entitled to them, whether the dynamic linking of amount and payee survives the full authorisation chain, and whether an authentication downgrade path remains reachable from a legacy client. Transaction risk analysis exemptions are reviewed for whether the fraud rate thresholds are actually measured or merely asserted.
Open banking APIs. Third-party provider integrations expose consent and account access to callers whose own security posture the firm does not control. Assessment covers consent lifecycle handling — creation, scope, expiry, revocation — and whether a revoked consent immediately stops data flow across every cached path. Certificate-bound access, client authentication and token binding are tested for whether the binding is enforced by the resource server or only by the authorisation server. Account information endpoints receive the same object-level authorisation testing as any other API, because a consent identifier is just another identifier that can be swapped.
Tokenisation and key handling. Where the firm operates its own token vault or encryption service, we review key hierarchy, rotation, access separation between the vault and the applications that call it, and whether an application compromise yields bulk detokenisation or only per-transaction access. Rate limits and monitoring on detokenisation volume are treated as controls, not conveniences.
Fraud and business logic. Refund replay, partial capture manipulation, currency conversion rounding, chargeback workflow abuse and limit counters that reset on a boundary rather than a rolling window are all tested explicitly. These findings rarely appear in tooling output and are frequently the highest-value items in the report.
Reporting is structured so the same evidence answers a scheme question, an internal audit question and a supervisory question without being rewritten three times.



