Red Team Cyber Security: How to Test Your Incident Response Plan
A scoped test of whether your written incident response plan works under pressure, run before the holiday freeze and reported to the board as a short list of decisions.
Your team finished awareness training and wants to test a ransomware scenario before the holiday change freeze, but nobody has agreed on what a pass looks like. For a mid-market business, a red team cyber security exercise worth running is a scoped test of whether the written incident response plan works under pressure, not a full adversary simulation. Agree the pass criteria in writing before the first inject, run the test in early December, and take the findings to the December board packet as a short list of decisions.
This week, name the one decision the exercise must prove
Start by deciding what kind of test you are buying. A tabletop is a discussion in a room: people talk through a scenario and nothing is measured. A validation exercise injects a live signal into the environment, lets staff respond as they would on a real night, and times what they actually do. A full red team engagement goes further still, with an offensive team trying to reach its objectives through whatever route works. For most mid-market teams with a written plan and one tabletop behind them, the validation exercise is the honest next step. The full engagement can come later, once there is something mature enough to test.
Before you scope anything, write down the single decision you most need to see tested. An example is whether the on-call lead can take a system offline without waiting for an executive. A narrow decision produces a finding you can act on. A broad exercise produces a long report that nobody funds. If you have not updated your incident response plan in the last year, or the people named in it have changed, fix that first. Testing a plan that names people who have left tells you nothing about your organization. The incident response plan guide covers how to write and assign the plan itself, so use it to get the document ready before you test it.
How a validation exercise differs from a tabletop
A tabletop asks, "What would we do?" A validation exercise asks, "What did we do, and how long did it take?" The difference matters for the result. In a tabletop, everyone in the room knows the scenario and can answer correctly. In a validation exercise, most of the staff do not know a drill is running, so their responses are real. Senior management alone knows the exercise is in progress. A tabletop nobody can fail is a rehearsal, not a test.
The check to run before you scope anything
Before you book a date, confirm three things. First, that the plan names an incident commander, a finance approver and a communications lead, and that each has a backup. Second, that the plan states a notification window for the board and any regulator that applies. Third, that the people in those roles can be reached on a holiday weekend. If any of these fail, the exercise will mostly measure the gap, which is useful, but you should know that is the purpose before you begin.
By Thanksgiving, get the sponsor to sign the pass criteria
Pass criteria are the most important part of the exercise and the part most often skipped. Write them down, and have the executive sponsor sign them before the first inject. If the criteria are written afterward, the exercise will pass by definition, because anyone can justify a result once they have seen it. The criteria should be specific enough that two people reading the timeline would reach the same verdict.
| Pass criterion | What is recorded | Threshold the sponsor sets |
|---|---|---|
| Time from alert to a named incident commander | Timestamp of the first alert and of the commander's acknowledgment | Set in the plan, for example within 30 minutes during business hours and within 60 outside them |
| Time to isolate affected systems | Timestamp of the isolation decision and of completed isolation | Set by the sponsor against the business's tolerance for downtime |
| Backups confirmed restorable | Restore test result for one named system, with the date restored | Restore completed and verified within the recovery time the plan states |
| Board notified within the plan's window | Timestamp of the notification and the notice sent | Within the window the plan itself states, not a window invented for the test |
| Escalation authority exercised | Who approved the decision to take a system offline, and when | A named role with authority, reached within the stated time |
Set the thresholds from your own plan and your own tolerance for disruption, not from a number you have read somewhere else. Where a threshold is missing from the plan, that gap is itself a finding, and it belongs on the list you take to the board.
Before any inject, write rules of engagement someone can stop
Rules of engagement are the document that keeps an exercise from becoming an incident. Write them before any inject, and keep them short enough that a manager can read them in five minutes. The document should cover these points:
- What is in scope. Name the systems and accounts in play, which are out of bounds, and the production systems that must not be touched.
- Who can halt it. Name who can stop the exercise, and the exact words that stop it. Write the phone numbers beside the names, not just the titles.
- Who knows about it. Keep the group that knows the exercise is running as small as you can, and write down who must not know.
- Legal and insurance review comes first. Counsel approves the scenario, and you confirm the exercise will not breach a policy condition.
- A real incident during the test. If a genuine incident starts, the exercise stops and the real response takes priority.
- Outside providers. Your managed service provider and your cloud host each have terms on testing. A test that touches their systems without notice can cause a real incident of its own.
Freeze exceptions and holiday coverage
The most common mistake is running the exercise during a change freeze without a written exception. In most mid-market organizations, changes stop in mid-December and do not resume until early January. Get the freeze exception from whoever owns change control, record it, and confirm who will be on duty before anything is scheduled. An exercise that starts during the freeze and cannot be stopped cleanly becomes a problem for change control, not a test of the plan.
Vendors, managed providers and cloud hosts
Give third parties their own line in the document. Write down who at each vendor you would call, what they need to know in advance, and whether they have consented to any activity against their systems. Your exposure often arrives through a vendor. A third-party software incident affecting Fort Worth and Crowley school districts exposed employee data without touching the districts' own systems. A scoped exercise should include at least one vendor-mediated path for that reason.
In early December, test who can decide before any tool
Run the test in early December, before the holiday absences set in, and before the change freeze takes effect. Begin with people, not technology. Test whether the incident commander, the finance approver and the communications lead can make their decisions when the scenario calls for them. Ask who can approve taking a production system offline, who can sign off a decision connected to a ransom demand, and who speaks to customers and regulators. Most plans leave those answers vague until a real event forces them. If the answer to "who can take the billing system offline at 2 a.m. on December 27" is not written down, you have found the most important gap in the plan, even if the technical controls work perfectly.
Plan the schedule around holiday absences. If the incident commander is on leave, you will be testing the absence, not the plan. Confirm the backups for each role on paper first, and run the test with the people who would actually be on duty during the holiday period.
Who can take a system offline
This is the decision most worth testing. In the exercise, an on-call lead receives an alert that a file server is encrypting data. Record how long it takes that person to decide whether to isolate the server, who they contact, and whether they wait for an executive before acting. The plan should let a named person act without waiting. If the exercise shows that the on-call lead waits thirty minutes for a reply that never comes, that finding is worth more than any tool purchase.
A physical disruption and the manual fallback
Add a physical or facilities disruption alongside the cyber event. Texas organizations already plan for heat, severe storms, flooding and occasional winter freezes, and the same continuity muscle tests a ransomware scenario. Ask how staff keep working when a site or data center is unavailable for days, and whether manual procedures exist for the critical tasks. An organization that can only recover from a cyber event is not resilient. Test the manual fallback, not just the detection. The business continuity program covers the planning that makes this possible.
While it runs, record findings against MITRE ATT&CK
The MITRE ATT&CK framework is a public, maintained knowledge base of adversary tactics and techniques. Tactics describe why an adversary acts, such as gaining access or moving through a network. Techniques describe how they do it. Use it to give your findings a shared vocabulary that your security staff already understand. Check the current version on MITRE's website when you plan the exercise, because the framework is updated over time.
Map each inject to a technique
- Write each inject as a simulated adversary action, such as a phishing email that leads to a credential being entered.
- Find the matching tactic and technique in ATT&CK, and record the identifier beside the inject.
- Note which control was supposed to detect or block that technique, and who owns it.
- After the exercise, record whether the control fired, whether a person noticed, and what they did next.
This turns a vague sense that the test went badly into a list of specific gaps. "Detection missed a credential-based movement technique" is a finding a technical owner can fix. "The team seemed slow" is not.
Keep the clock, not the impression
Assign one person to keep a timeline of every decision, detection and handoff, with timestamps taken from the systems and messages, not from memory. Do not let findings rest on how the exercise felt in the room. People remember the dramatic moments and forget the forty minutes spent waiting for approval. The timeline is the evidence the sponsor and the board will rely on, so keep it accurate.
Before the December board packet, turn findings into decisions
Rank each finding by business impact and by how likely the gap is to matter in a real event. Give each one a single owner and a single date. Separate the urgent fixes, which can be done without new money, from the items that need a budget decision. A remediation list with more than about five funded items on one page will not be read, so keep the request short.
Put the pass criteria next to each result. A board member should see that the criterion was "isolate within 30 minutes," that the result was "isolated in 48 minutes," and that the owner has a date to close the gap. A board can act on a line like that. A paragraph saying the exercise went reasonably well is only an opinion.
Urgent fixes and budget decisions
Urgent fixes are usually process changes: naming a backup for each role, writing down the authority to isolate a system, or correcting a notification window that nobody can meet. Budget decisions are things like a new monitoring capability or a restore-testing arrangement. Keep the two lists apart so the board can approve the urgent work without debating the budget, and so the budget request is not lost among small process fixes.
Fund, accept or retest
For each finding, the board chooses one of three things. It can fund the fix, accept the residual risk in writing, or ask for a retest by a set date. A finding nobody decides on will reappear in the next exercise. Schedule the retest in the same packet so the board can see when each decision will be checked. Security metrics that connect these decisions to cost and downtime help directors read the results; the board security metrics guide sets out what directors tend to ask, and the virtual CSO program covers how the board report is produced and kept current.
When a scoped exercise is enough and when it is not
Most mid-market teams do not need a full red team exercise yet. They need a tested plan with named owners, clear decision rights, and findings that are actually fixed. A full adversary simulation is a large purchase, and it pays off only when the defenders have something mature enough to test. If you are unsure which you need, the answer is usually the scoped validation exercise.
Reconsider the scope if the plan has not been updated in the last year, if the people named in it have changed, or if the organization has never run a test that involved a real decision and a real clock. In those cases, fix the plan and the people first. A virtual CISO can design the exercise, direct it, and adjudicate the results against the pass criteria, and can take the remediation list to the board. Total 360 Security is based in Fort Worth, serves Texas, California, Nevada and New Mexico, and runs incident response planning as part of its virtual CISO services. The firm does not perform live technical red teaming, so if you need an offensive engagement, that is a separate purchase. To scope a validation exercise your team can run before the year-end freeze, book a 30-minute risk discussion at total360security.com/consultation.
Frequently asked questions
What is a red team exercise for a mid-market business?
It is a goal-focused, adversarial test of how people, process and technology respond to an attack. For a mid-market team it is usually narrowed to one plan, one decision and agreed pass criteria, rather than a full adversary simulation.
How is a red team exercise different from a tabletop?
A tabletop discusses a scenario in a room. A validation exercise injects a live signal, measures and times what staff actually do, and keeps most of the staff unaware that it is a drill while the executive sponsor knows.
Do we need MITRE ATT&CK to run a validation exercise?
Not strictly. It gives the team shared vocabulary to name what the simulated adversary did and where detection missed. Check the framework's current version on MITRE's website when you plan.
Can we run the exercise during the year-end change freeze?
Only with a written freeze exception from whoever owns change control, recorded before scheduling, and with a named backup for each role confirmed on paper. Otherwise, schedule it before the freeze begins.
Who should know the exercise is running?
Keep the circle small: the executive sponsor, the exercise director and the people named in the rules of engagement, with stop words and phone numbers written down. Everyone else is tested without knowing.
How many findings should go to the board?
Keep it to a short list of no more than about five funded items on one page, each with one owner, one date and its pass criteria beside the result. Separate urgent fixes from budget decisions.
Total 360 Security provides virtual Chief Security Officer, vCISO, and Enterprise Security Risk Management programs for mid-market organizations. They design, run, and report security programs covering all 13 risk domains rather than just offering advisory decks or point solutions. The firm is Texas DPS-licensed with CPP, CISSP, and CISM credentials.
Serving Texas, California, Nevada, New Mexico.
More from the blog
10 min read
Identity Access Management Software: What to Govern First
A practical order for identity access management software: define privileged access, name an owner, inventory raw exports and review each quarter before buying a tool.
Read article7 min read
SOC 2 Audit: What Type II Requires and Who Owns Each Control
A SOC 2 audit is an independent CPA examination of your controls against the AICPA criteria. Here is what a Type II report requires, who owns each control and what evidence to keep.
Read article7 min read
Compliance Monitoring: What to Check Monthly and Quarterly
Compliance monitoring is the ongoing check that your operations follow the rules that apply to you. This guide sets a monthly evidence set, a quarterly sampling routine and clear ownership for every exception.
Read article