security 7 min read

Color Team Operations: Red, Blue, Purple, and Why You Need All Three

Red finds the holes. Blue closes them. Purple makes sure they stay closed. A security program with only one color is a security program with a blind spot — and attackers find blind spots for a living.

Tomáš Korcak (korczis) 537 words

Color teams are the oldest trick in security operations and still the most misunderstood. Red is not "people who hack." Blue is not "people who respond." Purple is not "red + blue in a meeting." Each has a distinct job, and the gap between them is where real incidents live.

Red: adversarial, goal-oriented, scoped

Red's job is to prove — through action, not opinion — that a specific outcome is reachable by someone who does not work for you. "Reachable" has to be operationally defined: "can an external attacker, starting with only public information, obtain read access to the DD case database within 5 business days?" That is a red-team question. "Is our password policy strong?" is not.

Red succeeds when it delivers an artifact: a screenshot, a captured credential, a written exploit, with a reproducible chain. The deliverable is the chain, not the finding. A finding without a chain is a suggestion. A chain is evidence.

Blue: boring, continuous, evidence-driven

Blue's job is the opposite: make the artifacts red would produce impossible or detectable. The boring parts — patch management, log retention, alert tuning, remediation tracking — are 80% of blue work and 0% of security marketing. Blue is also the team that owns threat assessment: the model of what is at risk, who would want it, and how they would get it.

Blue succeeds when it can produce detection telemetry for every stage of a red chain — or, better, when it prevents the stage entirely.

Purple: the loop that makes the other two matter

Purple is not a team in the org chart. It is a workflow. Red finds a chain. Purple walks the chain with blue, one stage at a time, and asks "what would have caught this at stage 3?" The answer is a detection, an alert, a block, or a playbook. Purple closes the gap and walks the chain again, next month, to verify the closure.

Without purple, red's findings land in a report. With purple, they land in the detection stack.

Why you need all three

  • Red only — you find holes but do not close them. Your threat posture is a list of problems.
  • Blue only — you close the holes you already know about. Your threat posture is a hope that the next attacker is not creative.
  • Red + blue, no purple — findings exist but never become detections. The same chain works twice.
  • All three — the loop closes. Each red iteration makes the next one harder.

Penetration testing is not red teaming

Pen tests are scoped to a system. Red teams are scoped to an outcome. A pen test of the DD API asks "are the endpoints correctly authenticated?" A red team asks "can I get to the DD data?" The second question finds chains the first one does not see because chains cross system boundaries.

Where to go next

Three colors. One loop. Drop any color and the loop breaks.