Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Atomic Red Team is an open library of small, single-technique attack tests mapped to MITRE ATT&CK. Each test runs in seconds and answers one question: does our detection fire for this behaviour?

What Is Atomic Red Team?
Atomic Red Team is a library of small, self-contained attack tests maintained as an open-source project by Red Canary since 2017. Each test exercises one MITRE ATT&CK technique, usually in a handful of commands, and comes with its own prerequisites and cleanup steps. The design goal is deliberately narrow: not to simulate an adversary, but to generate one specific behaviour so you can check whether anything noticed.
The repository stores tests as YAML files under an atomics folder, one directory per technique identifier. A test declares a name, a description, supported platforms, input arguments with defaults, the executor (PowerShell, command prompt, bash, or manual), the command itself, and a cleanup command. Because the format is plain text, teams routinely fork the repository and add their own tests for behaviours specific to their estate.
The companion PowerShell module, Invoke-AtomicRedTeam, handles execution and bookkeeping:
Invoke-AtomicTest T1003.001 -ShowDetailsBrief
Invoke-AtomicTest T1003.001 -CheckPrereqs
Invoke-AtomicTest T1003.001 -TestNumbers 1
Invoke-AtomicTest T1003.001 -TestNumbers 1 -Cleanup
That sequence is the whole workflow: look at what the test does, install what it needs, run it, then put the machine back. T1003.001 in this example is credential dumping from LSASS memory, one of the most commonly tested techniques because so many detections claim to cover it.
Run a test and then look at your pipeline. Three outcomes are possible, and all three are useful. The detection fires, which validates the rule end to end including log collection, parsing, and alerting. Nothing fires but the telemetry is present, which means the data reached you and the rule is wrong. Or neither the alert nor the telemetry appears, which means the sensor is missing, misconfigured, or the log never left the host. The third case is the most common finding in a first round of testing and the one organisations are least prepared for.
Recording the results against technique identifiers produces a detection coverage map based on evidence rather than on what a vendor datasheet claims.
Atomic tests are atomic by design. They do not chain techniques, maintain access, adapt to defences, or behave like an operator with an objective. A test that triggers your detection in isolation may be missed entirely when the same technique appears inside a real intrusion with different parent processes and timing. For chained emulation you want MITRE Caldera or a commercial breach and attack simulation platform, and for an assessment of whether your people and processes cope under pressure you want an actual red team engagement.
Used for what it is, though, it is the cheapest detection assurance available. A scheduled run of 30 relevant tests each week, with the results written back into your rule repository, turns detection coverage from an assertion into a measurement, and it costs nothing but the time to set it up.
Many tests are, but not all, and the project does not guarantee it. Some modify security settings, create accounts, or drop files that antivirus will quarantine. Review the YAML for each test, run it in a lab first, and begin production testing on a small ring of machines with owner agreement. Always execute the cleanup step and verify the host afterwards rather than assuming it worked.
Atomic Red Team runs individual techniques on demand and is free and transparent, with every command visible in the repository. Commercial simulation platforms chain techniques into full scenarios, run continuously from managed agents, score coverage automatically, and include safety controls and reporting. The trade-off is cost and flexibility: atomics are better for targeted validation of a specific rule, platforms for ongoing programme-level assurance.
It depends on the question you are asking. To validate that a particular rule works, tell them, because you want a clean technical result without a wasted investigation. To test whether an alert would actually be noticed and acted on during a busy shift, do not tell the analysts, but make sure at least one person outside the test team knows, so an unplanned escalation can be stopped quickly.
Treat it as regression testing. Run the tests relevant to a rule whenever that rule changes, run a core set weekly or monthly against a sample of endpoints, and run the full relevant set after any significant platform change such as an EDR upgrade or a logging pipeline migration. Automating the run and the result capture is what makes the cadence sustainable.
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.