#What DORA requires
| Element | Requirement |
|---|---|
| Who | Financial entities identified by competent authorities under Article 26(8) |
| How often | At least every three years; authorities can adjust frequency based on risk profile |
| Scope | Critical or important functions, on live production systems, including relevant ICT third-party providers |
| Method | Threat-led: bespoke threat intelligence drives red-team scenarios; methodology aligned with TIBER-EU |
| Detailed rules | RTS on TLPT, Commission Delegated Regulation (EU) 2025/1190 [2] |
| Testers | Article 27 sets requirements for testers; internal testers allowed under conditions |
#How a TLPT runs
- Preparation: scope critical functions, form a control team that keeps the test secret from defenders.
- Threat intelligence: a targeted threat report produces realistic scenarios for your entity.
- Red team testing: testers execute scenarios against production over an extended period.
- Closure: purple-team replay, remediation plan, and attestation.
#TLPT vs a conventional penetration test
| Conventional pentest | DORA TLPT | |
|---|---|---|
| Driver | Scope list | Threat intelligence about actors targeting you |
| Target | Systems in scope | Critical or important functions end to end |
| Environment | Often staging or production | Live production |
| Defender awareness | Usually informed | Blue team not informed (control team only) |
| Duration | Days to weeks | Weeks to months |
| Output | Vulnerability report | Scenario results, purple-team replay, remediation plan, attestation |
#Preparing between cycles
A three-year cycle invites a three-year blind spot. Entities that perform well run continuous validation of the same critical functions between TLPTs, so the formal test confirms resilience rather than discovering the obvious. The exposure-window model quantifies why.
#Sources
- [1] ECB Banking Supervision, supervisory guide (DORA applicability and TLPT)
- [2] Joint Committee of the ESAs, Final report on draft RTS on TLPT (JC 2024-29)
Talk to an operator.
Thirty minutes, no slides. We'll show you what bad AI finds in one of your systems.
Book a briefing