Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Breach and attack simulation runs safe versions of real attacker techniques continuously against your own estate, measuring which controls block, which detect, and which do neither.

What Is Breach and Attack Simulation?
Breach and attack simulation is automated, repeatable testing of your own defences using the techniques attackers actually use. Agents deployed across the estate execute safe variants of real behaviours, the platform records what the security stack did about each one, and the result is a continuously refreshed measurement of control effectiveness rather than an annual snapshot.
A simulation runs in three parts. First, execution: the agent performs a technique such as writing a known-bad file, attempting credential access, beaconing to a controlled server, or staging data for exfiltration. Second, observation: the platform checks whether the control blocked it, whether telemetry was generated, and, through an integration with your SIEM, whether an alert was raised. Third, scoring: results are mapped to MITRE ATT&CK and tracked over time.
Safety is engineered in. Exfiltration goes to the vendor’s endpoint rather than a real adversary’s; ransomware simulations encrypt files the agent created itself; destructive steps are replaced with equivalent but harmless actions that trigger the same detection logic. That is also the main limitation, as discussed below.
| Frequency | Scope | Answers | |
|---|---|---|---|
| Vulnerability scanning | Continuous | Known weaknesses | What is exposed |
| Penetration testing | Periodic | Defined target | What can be exploited |
| Breach and attack simulation | Continuous | Technique coverage | Which controls work |
| Red teaming | Occasional | Whole organisation | Whether people and process cope |
They answer different questions and do not substitute for one another. A penetration test finds the flaw a scanner cannot reason about; simulation tells you whether your detection would catch the technique that follows exploitation. Treating the second as a replacement for the first is the most common procurement mistake in this area.
The first run is usually uncomfortable, and the findings fall into a recognisable pattern. Controls that are licensed but not enabled. Detection rules that work on the analyst’s test machine and not on the standard laptop build. Log sources that stopped forwarding weeks ago and nobody noticed. Prevention that works for the parent technique but not for the sub-technique variant. And a measurable gap between what the coverage map claims and what the estate does.
The ongoing value is regression testing. Security controls drift constantly: an exclusion added to unblock a deployment, a policy applied to the wrong group, an agent upgrade that changes default behaviour. Simulation catches the drift within a day instead of at the next assessment.
Simulated techniques are necessarily safe, which means they are not identical to the real thing, and detections that key on the harmless substitute can score green against an attack they would miss. Agent-based execution also starts from a position an attacker has to earn, so simulation says nothing about initial access through a supplier, a phishing lure that works, or a vulnerable edge device. And nothing in the platform tests whether an alert was noticed and acted on by a tired analyst at 02:00, which is what red team engagements exist to find out.
Platforms in this space include AttackIQ, SafeBreach, Picus, and Cymulate, with the open-source MITRE Caldera and Atomic Red Team covering much of the same ground for teams prepared to build the orchestration themselves.
Scope it by threat rather than by catalogue. Pick the groups credibly targeting your industry, take their technique sets, and run those first; a scorecard across 400 generic techniques is harder to act on than a failing result against the ten techniques your most likely adversary uses. Feed every failure into the detection backlog with an owner and a date, and re-run the specific test to confirm the fix. For UK organisations, the resulting evidence maps neatly onto NCSC Cyber Assessment Framework principle C1 and supports the continuous assurance expectations in FCA and PRA operational resilience rules.
No, and treating it as a replacement leaves real gaps. Simulation measures whether known techniques are blocked or detected across the estate, continuously and automatically. Penetration testing applies human creativity to find flaws in business logic, chained misconfigurations, and application behaviour that no automated technique library contains. Most organisations need both, plus periodic red teaming to test people and process.
The platforms are designed for it, using controlled infrastructure, self-created test files, and harmless substitutes for destructive actions. Risk is not zero: agents run with high privilege, tests can trigger automated containment, and a misconfigured scope can hit sensitive systems. Pilot on a representative subset, agree the scope with system owners, and tell the SOC which activity is attributable before the first full run.
Continuously for a core set covering your highest-priority techniques, typically daily or weekly, so control drift is caught quickly. Run a broader campaign monthly or quarterly, and always after a significant change such as an EDR migration, a new email gateway, or a logging pipeline rebuild. The value comes from the trend line, which only exists if the cadence is kept.
Prevention rate and detection rate per technique, the gap between telemetry generated and alerts raised, time from a failed test to a deployed fix, and the proportion of previously failing techniques that stay fixed after 90 days. That last figure is the most revealing, because it shows whether remediation is durable or whether controls drift back within a quarter.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.