COFFSec is an offensive security firm based in Hong Kong. Through penetration testing, adversary emulation and specialist security reviews, we investigate how attackers could compromise critical systems and business processes. Our assessments provide technical evidence for managing strategic, operational, financial and compliance risk, with clear findings, practical remediation and explicit limitations.
Provide independent offensive security evidence that helps organisations protect critical assets, prioritise remediation and make informed risk decisions.
Vision
A security practice in which technical assurance supports business resilience, responsible investment and effective risk oversight.
Values
Professional independence, lawful authorisation and careful handling of client information guide our work. Conclusions reflect the evidence, including uncertainty and assessment limits.
Our offensive assessments establish how a technical weakness could affect a business objective. We identify the control involved, demonstrate the exposure within authorised limits and provide evidence for remediation priorities, investment decisions and the acceptance of residual risk.
Hong Kong threat context
Reported incidents in Hong Kong
HKCERT publishes monthly counts of incidents reported to the centre, including phishing, malware, unauthorised access and denial of service. The series provides regional context for threat assessment and scenario selection. Reporting volumes are influenced by public awareness, reporting practices and campaign activity. They do not establish an individual organisation’s likelihood of compromise or measure the effectiveness of its controls. Use these figures alongside your asset exposure, incident history and assessment evidence.
Threat-led penetration testing is defined by supervisory authorities rather than by the firms that supply it. Each regime below prescribes which institutions fall within its population, what intelligence must substantiate the scenario, and what evidence the institution is obliged to produce. We map an engagement onto the regime that governs it, so that the resulting assurance withstands supervisory scrutiny rather than merely satisfying an internal programme.
Select a framework
01 / 15
HK · Hong Kong
Standing
Mandatory for authorized institutions rated medium or high for inherent risk
01
Population
Authorized institutions assessed under the C-RAF; applicability follows from the inherent-risk rating assigned to the institution
Cycles
Runs in step with the framework assessment and recurs at each maturity review
Roles
An iCAST manager, a threat-intelligence specialist and testers, each qualified against CREST-mapped requirements
Governing authority
Hong Kong Monetary Authority, within the Cybersecurity Fortification Initiative (CFI 2.0)
Hong Kong situates attack simulation within a supervisory assessment of the institution as a whole, so the exercise inherits both the inherent-risk rating that triggered it and the maturity level the institution is expected to attain.
Threat intelligence is tailored to the Hong Kong banking sector, and every scenario specifies the compromise outcome sought against a named critical service — data insertion, loss of integrity or privilege escalation among them. Since CFI 2.0, detection and response are measured while the red team is still operating, and testers must hold qualifications mapped to recognised certifications. The closing report binds each scenario to a remediation plan the institution itself owns.
Test phases
01
Preparation and scoping
02
Tailored threat intelligence
03
Scenario design
04
Execution under blue-team observation
05
Closure and remediation plan
SG · Singapore
Standing
An industry guideline rather than a rule; adoption is expected throughout the financial sector
02
Population
Banks and financial institutions operating in Singapore
Scope
The people, processes and technology supporting critical systems, production environments included
Independence
Threat intelligence and red-team execution are assigned to separate teams
Governing authority
Association of Banks in Singapore, with oversight from the Monetary Authority of Singapore
AASE is drafted as an industry guideline rather than a rulebook, so the exercise is shaped by the institution’s own risk assessment and by the scenarios its supervisor would regard as credible.
The 2024 revision tightens the intelligence phase, the control group’s authority to halt activity, and the treatment of testing against live production. Scenarios are derived from threat actors assessed as pertinent to Singapore’s financial sector, and the report addresses two audiences at once: the engineers who must close the attack path, and the board that must accept the residual risk.
Test phases
01
Planning and scoping
02
Threat intelligence
03
Scenario development
04
Red-team execution
05
Reporting and remediation
UK · United Kingdom
Standing
A supervisory instrument, applied to systemically important firms and financial market infrastructure
03
Population
UK financial market infrastructure and major firms, selected by the authorities
Providers
Threat-intelligence and red-team providers accredited under the framework
Output
Findings enter the supervisory record; the test is not scored as a pass or a fail
Governing authority
Bank of England, together with the Prudential Regulation Authority and the Financial Conduct Authority
CBEST supplied the template that most later regimes reproduce: regulated threat intelligence, an accredited red team, and a test appraised on what the institution detected rather than on what it blocked.
The framework is deliberately not a compliance checklist. The authority commissions or sanctions the test, the intelligence provider works from a sector-level threat picture, and the red team must attain defined business objectives without the institution’s prior knowledge. Findings are shared with supervisors so that remediation can be tracked, and successive tests show whether the control environment has genuinely shifted.
Test phases
01
Scoping with the authority
02
Threat intelligence
03
Red-team test
04
Debrief and remediation
05
Supervisory follow-up
UK · United Kingdom
Standing
Part of the PRA and FCA supervisory toolkit
04
Population
Broader than CBEST: banks, insurers, payment firms and financial market infrastructure
Objective
Important business services and their impact tolerances rather than a named system
Providers
Teams accredited to CREST STAR and STAR-FS
Governing authority
Bank of England, the Prudential Regulation Authority and the Financial Conduct Authority, with CREST accreditation
Simulated Targeted Attack and Response for Financial Services
STAR-FS extends threat-led testing beyond the largest banks and reframes the objective: not the compromise of a system, but the disruption of an important business service.
The exercise begins with the firm’s own mapping of its important business services and the impact tolerances set for them, then tests whether an adversary could exceed those tolerances. Because the population is broader, the framework is designed for firms that hold no in-house red team: accreditation, intelligence standards and reporting expectations are fixed, so a smaller institution can commission a test of comparable rigour. Results inform operational-resilience supervision.
Test phases
01
Business-service scoping
02
Threat intelligence
03
Attack and response testing
04
Impact-tolerance assessment
05
Remediation and supervisory review
UK · United Kingdom
Standing
Applied across government departments and public bodies
05
Population
Government departments, agencies and public bodies
Distinction
Designed for mission systems whose compromise carries political and operational consequence
Providers
Suppliers operating to Cabinet Office requirements under controlled-handling rules
Intelligence-led simulated attack testing for government
GBEST carries the CBEST logic into government: the same intelligence-led attack simulation, directed at mission systems where compromise carries political and operational rather than financial consequence.
The framework is invoked where a department requires assurance that an adversary cannot reach a defined objective, and where the test itself must withstand scrutiny. Scope, handling rules and reporting are set by the sponsoring authority; because the framework is not published as a public standard, an engagement rests on accredited providers and on the department’s own control group. For suppliers to government, a GBEST-aligned exercise is often the clearest evidence that security claims have been subjected to test.
Test phases
01
Authority-led scoping
02
Threat intelligence
03
Objective-based attack
04
Response observation
05
Closed reporting
EU · European Union
Standing
A reference framework for authorities, given effect through national guides
06
Population
Financial entities across the EU and EEA, designated by their competent authority
Structure
An intelligence provider, a red team, a control team and the authority, each with defined duties
Recognition
Comparable evidence underpins mutual recognition between authorities
TIBER-EU is the reference architecture for threat-led testing in Europe: it standardises roles, the intelligence deliverable and reporting, so that a test conducted in one member state can be accepted in another.
The framework separates the threat-intelligence provider from the red team by design, so that scenarios are not authored by the team that executes them. The 2024 revision aligns it with the regulatory technical standards for threat-led penetration testing under DORA, which converts a voluntary model into a supervisory obligation for significant entities. National guides adapt the same architecture to local supervisory practice.
Test phases
01
Preparation
02
Targeted threat intelligence
03
Red-team testing
04
Closure
05
Remediation and supervisory follow-up
EU · European Union
Standing
Mandatory for entities in scope
07
Population
Significant financial entities, with authorities empowered to designate further entities
Cadence
No less than once every three years
Oversight
The authority validates scope, provider eligibility and results
Governing authority
The European Supervisory Authorities with the ECB, under Regulation (EU) 2022/2554
Threat-Led Penetration Testing under the Digital Operational Resilience Act
DORA converts threat-led penetration testing from good practice into an enforceable obligation, with scope, frequency and evidence fixed in regulation rather than in a national guide.
Entities must test live production systems where feasible, drawing on intelligence and a red team that satisfy the regulatory technical standards, and the test must cover critical or important functions. Results inform the entity’s own risk treatment and the authority’s supervisory review, and any provider engaged for testing must be independent of the entity’s control functions. TIBER-EU supplies the operating model that most authorities apply.
Test phases
01
Scope validated by the authority
02
Intelligence
03
Production testing
04
Results validation
05
Remediation and follow-up
NL · Netherlands
Standing
A supervised programme; the central bank oversees each test
08
Population
Dutch financial institutions and financial market infrastructure
Authority role
DNB selects participants, approves scope and receives the results
Extension
Advanced Red Teaming introduced in 2024 for sustained adversary emulation
Threat Intelligence-based Ethical Red Teaming, Netherlands
TIBER-NL is the original national implementation and the setting in which the model was proven; later guides, TIBER-EU among them, reproduce its separation of intelligence from attack.
The institution convenes a small control team, the blue team is left uninformed, and the report is written to serve both remediation and supervisory judgement. DNB’s 2024 introduction of Advanced Red Teaming extends to continuous, sustained adversary emulation for institutions whose exposure warrants more than a periodic test, under the same evidential standards.
Test phases
01
Preparation with the authority
02
Intelligence
03
Red-team test
04
Debrief
05
Remediation monitoring
DE · Germany
Standing
Applied within German supervision and mapped to DORA threat-led penetration testing
09
Population
Banks, insurers and financial service providers subject to German supervision
Authority role
The Bundesbank owns the technical process; BaFin carries the supervisory consequences
Providers
Certified providers able to evidence threat intelligence specific to the German sector
Threat Intelligence-based Ethical Red Teaming, Germany
Germany applies the TIBER model within its own supervisory practice, the central bank owning the technical process and the conduct authority owning the consequences.
The national guide fixes the intelligence requirements, the test phases and the reporting format, so that a TIBER-DE test can be read by both authorities without translation. Institutions outside the selected population may run a test voluntarily, and the structure now maps onto DORA’s technical standards. Threat intelligence is expected to reflect actors that target German financial infrastructure.
Test phases
01
Scoping
02
Intelligence
03
Red-team test
04
Closure
05
Supervisory review
IT · Italy
Standing
A supervised exercise programme
10
Population
Italian banks and financial intermediaries
Scope
Critical functions, together with the systems and suppliers that support them
Alignment
TIBER-EU structure, now read alongside the DORA threat-led penetration testing standards
Threat Intelligence-based Ethical Red Teaming, Italy
Banca d’Italia conducts TIBER-IT as a supervised exercise, applying the TIBER-EU structure to test critical functions against actors assessed as pertinent to the Italian market.
The central bank sets the timetable, approves the threat intelligence and reviews the final report, so the exercise carries supervisory weight rather than remaining an internal assurance project. Institutions convene a control team, keep the blue team uninformed, and must demonstrate how findings altered the control environment before the next cycle.
Test phases
01
Authority planning
02
Intelligence
03
Red-team execution
04
Results review
05
Remediation follow-up
ES · Spain
Standing
A supervised exercise programme
11
Population
Spanish banks and supervised financial entities
Emphasis
Intelligence quality and due diligence on providers
Alignment
TIBER-EU, with the DORA threat-led penetration testing requirements
Threat Intelligence-based Ethical Red Teaming, Spain
The Spanish guide states expressly that the exercise is not an audit to be passed or failed: its purpose is to deepen the entity’s own understanding of its weaknesses.
TIBER-ES sets out how the entity selects and supervises its intelligence and red-team providers, how the control team is constituted, and what the report submitted to the supervisor must contain. It further describes how results are handled after the test, including the expectation that remediation is tracked to closure, and follows the TIBER-EU principle that mutual recognition rests on comparable evidence.
Test phases
01
Preparation
02
Intelligence
03
Test execution
04
Closure
05
Remediation tracking
AU · Australia
Standing
A national industry programme with regulator participation
12
Population
Australian financial institutions and financial market infrastructure
Structure
A red-team exercise, followed by a purple replay and a gold-team crisis simulation
Providers
Recognised providers that meet the framework’s minimum requirements
Governing authority
Council of Financial Regulators, comprising APRA, ASIC and the Reserve Bank of Australia
CORIE operates as a national programme rather than as an obligation falling on each firm individually, and it is the framework that formally pairs the red-team exercise with a purple replay and a board-level crisis simulation.
The red-team exercise is followed by a purple exercise, in which the institution’s defenders replay the attack alongside the provider, and then by a gold-team tabletop that places senior executives inside the same scenario. The sequence tests whether technical findings translate into decisions taken under pressure. Reports go to the institution and, in aggregate, to the Council, so that systemic weaknesses can be identified without ranking individual firms.
Test phases
01
Programme onboarding
02
Intelligence
03
Red-team exercise
04
Purple replay
05
Gold-team crisis simulation
CA · Canada
Standing
Required of systemically important banks and internationally active insurance groups
13
Population
Every domestic systemically important bank and internationally active insurance group
Cadence
No less than once every three years
Context
Runs alongside Guideline B-13 on technology and cyber risk management
Governing authority
Office of the Superintendent of Financial Institutions
I-CRT is Canada’s counterpart to CBEST and TIBER, written for a supervisory model in which the regulator defines the critical business functions subject to test.
The framework sets out four phases and requires the test to run against the systems that support critical business functions, with intelligence drawn from actors assessed as capable of targeting the institution. OSFI oversees the exercise and may require remediation; the institution’s risk treatment plan is expected to reflect the results. Because the population is narrow, the regime concentrates on institutions whose failure would unsettle the wider Canadian financial system.
Test phases
01
Scoping of critical functions
02
Intelligence
03
Red-team testing
04
Reporting
05
Remediation under supervision
SA · Saudi Arabia
Standing
Mandatory for financial entities regulated by SAMA
14
Population
Every member organisation regulated by SAMA
Cadence
No less than once every three years
Roles
A SAMA green team, an internal white team of up to five, and a blue team kept uninformed
FEER is among the few regimes in which the central bank itself appoints the test manager, giving the exercise an explicit supervisory owner from the outset.
SAMA’s green team appoints the test manager and approves the provider; the institution forms a small white team, and the blue team operates as it normally would, so that detection is measured honestly. The red team supplies sector-specific intelligence and conducts the attack against live production systems under controlled conditions. Findings inform both institutional remediation and SAMA’s sector-level view of resilience.
Test phases
01
Green-team appointment
02
Intelligence
03
Controlled attack
04
Reporting
05
Remediation
G7 · G7 jurisdictions
Standing
A reference point for cross-border recognition
15
Population
G7 jurisdictions and the entities their authorities supervise
Purpose
Common minimum elements for national threat-led penetration testing regimes
Effect
Supports mutual recognition of test results among authorities
Governing authority
G7 Cyber Expert Group, published through finance ministries and central banks
Fundamental Elements for Threat-Led Penetration Testing
The G7 elements are not a test method but a common denominator: they define what a credible threat-led regime must contain, so that authorities can accept one another’s results.
They address the authority’s role, the independence of intelligence and red team, the use of live production systems, the treatment of third-party providers and the handling of results. For a multinational group, this makes the document the practical reference whenever one jurisdiction’s test must satisfy another’s supervisor. CBEST, TIBER-EU, CORIE and I-CRT all map onto the same elements, which is why the evidence they produce looks alike across markets.
Austria, Belgium, the Czech Republic, Denmark, Estonia, Finland, France, Greece, Iceland, Ireland, Liechtenstein, Luxembourg, Malta, Norway, Portugal, Romania, Slovakia and Sweden.
Offensive security services
Technical depth. Business relevance.
Our services cover applications, infrastructure, cloud platforms, identity, adversary emulation, AI, physical environments and human risk. Each engagement connects a defined attack scenario to the assets and business processes it could affect. Strategic risk includes exposure introduced by acquisitions, new services and platform change. Operational risk concerns disruption and recovery. Financial risk includes fraud, asset loss and response costs. Compliance risk concerns the controls and evidence required by applicable obligations. These perspectives guide reporting without replacing the technical investigation.
We establish the attacker’s starting position, test the relevant security boundaries and explain the observed business consequence. Findings distinguish confirmed effects from plausible outcomes that remain untested. Technical severity, business priority and residual risk are considered separately, with the assumptions behind each conclusion made explicit.
Defined objectives and coverage
Scoping identifies critical services, sensitive information, relevant threats and the decision the assessment will support. The plan records identities, environments, exclusions and operational safeguards before testing begins.
Evidence with technical context
Findings retain the affected version, required privileges, test sequence and observed result. Review challenges the root cause, impact and corrective guidance. Variable behaviour and gaps in evidence remain visible.
Treatment and verification
Recommendations identify controls that interrupt the attack path and dependencies that affect implementation. Retesting verifies agreed changes. Projected risk reduction is distinguished from reduction supported by completed remediation and verification; acceptance remains with the client’s authorised risk owner.
Authority chain
Threat scenario
Control validation
Business impact
Risk treatment
The assessment links technical evidence to a documented risk scenario and a verifiable corrective action.
Professional references
Recognised guidance. Explicit limitations.
Technical methods draw on OWASP, NIST and MITRE guidance appropriate to the environment. Risk reporting is informed by NIST guidance on enterprise risk management and concepts reflected in the recognized information security frameworks, including asset protection, least privilege, control effectiveness and due diligence.
Offensive testing introduces controlled activity into systems that support real operations. The engagement plan identifies delivery responsibilities, review arrangements and the client contacts authorised to approve changes or stop work. Technical judgement remains independent of commercial pressure.
Technical review
Review checks whether the evidence supports the stated impact, whether prerequisites are clear and whether remediation addresses the underlying control failure.
Written authorisation
Assets, permitted techniques, operating windows and escalation procedures are agreed before active testing. Third party permissions are confirmed where required.
Protected client information
Evidence collection is limited to the assessment purpose. Access, transfer, processing, retention and deletion follow the agreed terms. Publication or reuse requires separate written permission.
Commission an assessment
Define the risk. Validate the controls.
Identify the service, information asset or business process requiring assurance. Consultation examines relevant attack scenarios, available access and operating constraints, resulting in a proposal with defined coverage, deliverables and verification criteria.
Scope and objectives
Identify the business concern, relevant systems and evidence required. Select testing, review or investigation according to the question and the operational risk.
Authority and preparation
Confirm ownership, permissions, representative accounts, safety limits and information handling. Resolve material dependencies before scheduling active work.
Assessment and handover
Validate the agreed scenarios, communicate urgent exposure and deliver reviewed findings. Explain remediation priorities, assessment limitations and the scope of any agreed retest.
Scope, scheduling and service commitments are confirmed in the engagement agreement.
We use essential cookies to keep the site working and analytics cookies to understand how you use it, so we can improve. We do not sell your data. See our Privacy Policy for details.