River
Y CombinatorBacked by Y Combinator

Free template

Security Incident Response Plan Template

An incident response plan with roles, scenario playbooks and notice templates, plus sheets that map your data to each notification deadline and its owner.

Free download · No account needed

What you get

Pack of 11 files: 6 documents, 5 spreadsheets

Security Incident Response Plan Template
  • How the Clocks WorkDocument
  • Worked Example, Stannary HealthDocument
  • Declaration and RolesDocument
  • Response Plan by ScenarioDocument
  • Notification TemplatesDocument
  • Legal and Regulatory NoteDocument
  • Data and Obligation MapSpreadsheet
  • Notification Deadline by ObligationSpreadsheet
  • Affected PopulationSpreadsheet
  • Contact and Escalation RegisterSpreadsheet
  • Tabletop Exercise LogSpreadsheet
RiverSecurity Incident Response Plan Template

Notification Deadline register · one row

One obligation, its own starting gun, and the hours you actually have

Discovery, recorded once12:40  ·  Friday, December 4, 2026

Every row in the register derives from that one field. Move it and all of them move.

Trigger eventDiscovery. Not awareness, not a materiality determination, and not the date somebody signed the contract.
Clock unitCalendar hours, which are elapsed time, so the weekend counts. Calendar days and business days do not behave like this.
Duration72
DeadlineMonday, December 7, 12:40. The same time of day the window opened.
OwnerA named person, confirmed reachable at the hour the deadline falls.

Working hours actually inside the window

Nine of seventy-two. Sixty-three hours fell outside Monday to Friday, 09:00 to 18:00. The same window opened on a Tuesday holds twenty-seven, and that is the number deciding whether a weekend gets staffed.

Sorted by expiry, never by severity

At Stannary the first two deadlines were contractual notice clauses expiring Saturday lunchtime and Sunday morning, both ahead of every regulation in the file.

The securities disclosure has no deadline on day zero, because its trigger is a materiality determination nobody has made yet. What goes in the row instead is the internal deadline for making that call, and the name of whoever makes it.

Page 1 of 1220 words

How to write a security incident response plan

A security incident response plan is the short document your team follows when someone suspects an attack or a data leak. It says who can declare an incident, who leads, what to do first for each likely incident, and which notification deadlines start. A good plan names real people with backups, works out which deadlines apply before anything happens, and has been tested in a tabletop exercise.

Write it for the worst time to use it: a Friday evening, with your usual decision-maker on a plane. Anything it leaves open gets argued over while the deadlines run.

Three engineers working side by side at laptops and monitors in an office, with a whiteboard of notes behind them
Settle the roles and deadlines before the team is at the keyboards. Photo: Alvaro Reyes on Unsplash

What goes in an incident response plan

Keep it short enough to read during an incident. It needs seven parts:

The parts of a security incident response plan

PartWhat it settles
Declaration and severityWho may declare, and severity levels based on what responders can see
RolesWho leads, investigates, communicates and tracks deadlines, each with a backup
Scenario pagesFirst steps, evidence to keep and the usual mistake for each likely incident
Obligation mapEvery law and contract attached to each data set, filled in now
Deadline registerEach notification deadline, computed at declaration from one discovery time
Notice draftsFirst notices for customers, regulators and affected people
Contacts and exercisesAfter-hours numbers with backups, and what each test found

Decide who declares and who leads

Let anyone declare an incident. Google's Site Reliability Engineering book says it is better to declare an incident early1 and close it after a simple fix than to respond hours into a growing problem.

Then script the first few minutes. Whoever declares posts what they saw, records the discovery time to the minute and leads until relieved. Computing the deadlines comes next, before triage, because they decide who you wake up.

Define severity by what responders can see, and start high when scope is unknown. An incident escalation matrix sets who gets paged at each level. Give every role a backup:

  • Incident commander: runs the response and does no technical work
  • Technical lead: contains the attack and finds out what it reached
  • Communications owner: approves everything that leaves the company
  • Deadline owner: keeps the register and calls out the next expiry
  • Counsel: joins at declaration for serious incidents

Map your data to every deadline

Most notification duties attach to data, not to the company. For each data set, record where it lives, whose it is, how many records it holds and where those people live. Note your role too (controller, processor or business associate), which can differ within one database.

Common breach notification clocks

RuleStarts whenUnitOuter limit
GDPR Article 33, controller to regulatorYou become aware of the breachCalendar hours72 hours, where feasible2
HIPAA, covered entity to individualsDiscoveryCalendar days60 days3
HIPAA, business associate to covered entityDiscoveryCalendar days60 days4
SEC Form 8-K Item 1.05, US public companiesThe company decides the incident is materialBusiness days4 business days5
US state breach lawsSet by each stateVariesCheck each state6 where affected people live
Customer and processing agreementsWhatever the signed clause saysHours or daysThe clause itself

These are ceilings. HIPAA also requires notice without unreasonable delay, and the GDPR without undue delay.

Under HIPAA, a breach counts as discovered once it was known or, by exercising reasonable diligence, would have been known3, so an unread alert may have started the clock. A covered entity is also treated as knowing of a breach once its agents do. If you are a business associate acting as a customer's agent, its 60 days may start when you find the breach.

Log the notice clause from every signed customer contract, not your standard template. US public companies and their subsidiaries add the SEC rule, and a cyber insurance policy may require prompt notice to the insurer.

Compute every deadline from one timestamp

Enter the discovery time once and derive every deadline from it, recomputing if an earlier signal turns up. Write the unit beside each number: calendar hours run through nights and weekends, calendar days count whole days, and business days skip weekends and holidays.

Sort by expiry, not by how serious each rule sounds, and count the working hours inside each window. A 72-hour window opened at 12:40 p.m. on a Friday holds 9 working hours, counting 9 a.m. to 6 p.m. on weekdays. Opened on a Tuesday, it holds 27.

One clock can't be computed on day one. The SEC's 4 business days start when a company decides an incident is material, a decision that must come without unreasonable delay after discovery5. The register holds a deadline for that decision instead, and who makes it.

ExampleA deadline register after a Friday discoveryAn illustrative example written for this guide. Names, contracts and dates are invented.

Discovery: Friday, December 4, 2026, 12:40 p.m.

Customer contract, 24 calendar hours: Saturday 12:40 p.m. (5.3 working hours).

EU processor agreement, 36 calendar hours: Sunday 12:40 a.m. (5.3 working hours).

GDPR Article 33, own EU records, 72 calendar hours: Monday 12:40 p.m. (9 working hours).

SEC Item 1.05: no date yet. Materiality decision due Monday 5 p.m., owner named.

HIPAA, notice to client clinics, 60 calendar days: Tuesday, February 2, 2027.

  • Contracts first. The two tightest deadlines are contract clauses, both expiring over the weekend.
  • Staff it on Friday. The GDPR window holds only 9 working hours, so weekend cover cannot wait.

Write a page for each likely incident

Pick four or five scenarios from your own incidents and exposure, not a generic list. Give each page the same five headings: signals, first hour, deadline changes, evidence to keep and the usual mistake.

First moves people get backwards

ScenarioDo this firstNot this
Data theftCopy logs before they rotate, and cut off the route outHolding notices until you have an exact record count
RansomwareDisconnect affected machines from the networkSwitching them off, or restoring before finding the way in
Stolen loginEnd every session, then look for new keys and sign-in methodsOnly resetting the password
InsiderPreserve logs and the device, then bring in HR and counselConfronting the person first
Vendor breachGet the vendor's discovery time in writingWaiting for their investigation to finish

The Federal Trade Commission gives the same advice for any breach: take all affected equipment offline immediately, but don't turn any machines off6 until the forensic experts arrive.

Draft the notices, then test the plan

Draft each first notice now: what you know, what you don't and when the next update comes. Give customers your discovery time, since their own deadlines may run from it. For public status updates, use an incident status page update template.

A HIPAA notice to individuals must be written in plain language3 and give, where known, the breach and discovery dates, the information involved and the steps people should take. Under the GDPR, details for the regulator may be provided in phases2, so file on time with what you know.

Then run a 90-minute tabletop exercise. Set discovery on a Friday evening, give no warning, and time how long it takes to name the first deadline and its owner. Halfway through, change the scope, since scope moves deadlines.

Get counsel's answers in writing

This guide is general information, not legal advice. Before an incident, ask a lawyer four things: which laws and contracts apply, what counts as discovery, whether you act as a customer's agent, and who decides materiality. Date each answer.

Before you call the plan ready

  • Everyone knows they may declare an incident
  • Every role has a backup with an after-hours number
  • Every signed contract's notice clause is in the map
  • Every deadline records its trigger, unit and owner
  • A tabletop ran in the last 12 months, and its actions are closed

After an incident, turn the alerts and chat into a postmortem report. Between incidents, answer the incident sections of customers' security questionnaires from the plan. The template on this page has the obligation map, deadline register and scenario pages ready, with sample entries.

How to use this template

Use the files as they are, or open them in River. River asks where your data lives and whose it is, and for your signed customer agreements and data processing agreements. It also asks whether a listed parent company or a cyber insurance policy applies, and for any plan or exercise notes you already have.

From what you send, River builds the obligation map and tells you which deadline in your file falls first. It also names the four questions your lawyer should answer in writing, since none of this is legal advice.

  1. Download it or open it in River

    Download six Word documents and five CSV sheets free, or open the whole template in River to work with AI.

  2. Map your data and contracts

    Fill in the Data and Obligation Map yourself, or send River your data list and signed agreements to build it.

  3. Write your scenarios and notices

    Adapt the scenario pages to incidents you could really face, and finish the notice drafts while nothing is urgent.

  4. Test it with a tabletop

    Run a timed exercise, log the minutes until someone names the first deadline, and close every action it raises.

What's in the template

One timestamp, seven obligations, and a weekend nobody had modeledEvery deadline below derives from a single discovery timestamp entered once. Nothing was typed in by hand.
Security Incident Response Plan Template
RiverNotification Deadline

Notification Deadline by Obligation

Stannary Health, a scheduling vendor to 41 US clinics and 6 EU clinics, with a listed parent. Discovery: Friday, December 4, 2026, 12:40.

InstrumentTriggerUnitQtyDeadlineWorking hrs inside
BAA notice clause, 27 of 41 customersDiscoveryCalendar hours24Sat Dec 5, 12:405.3
EU processor DPA, Havenbrook ClinicsDiscoveryCalendar hours36Sun Dec 6, 00:405.3
GDPR Art 33, own EU controller recordsAwarenessCalendar hours72Mon Dec 7, 12:409.0
SEC Form 8-K Item 1.05, listed parentMateriality determinationBusiness days4Fri Dec 11not keyed on discovery
HIPAA 164.410(b), notify the covered entitiesDiscoveryCalendar days60Tue Feb 2, 2027plenty
HIPAA 164.404(b), covered entity to individualsTheir discoveryCalendar days60Tue Feb 2, 2027plenty
HIPAA 164.408(b), Secretary, 500 or moreDiscoveryCalendar days60Tue Feb 2, 2027plenty

Sorted by expiry, the first two rows are contracts and the third is a regulation. The room had assumed the 72-hour clock was the tightest thing they had. Rows five and six landing on the same date is the finding that changed the plan: the outer regulatory limit gives the vendor sixty days and gives its customer none.

RiverData and Obligation Map

Data and Obligation Map

Filled in before an incident. This is the only sheet in the pack that has to exist in advance, and everything else derives from it.

Data setWhoseRecordsOur roleInstrumentTriggerUnitQty
Patient appointment records41 clinics141,200Business associateBAA notice clauseDiscoveryCal. hours24
Patient appointment records41 clinics141,200Business associateHIPAA 164.410(b)DiscoveryCal. days60
Patient intake forms41 clinics19,100Business associateHIPAA 164.410(b)DiscoveryCal. days60
EU clinic appointment records6 EU clinics9,260ProcessorProcessor DPA clauseDiscoveryCal. hours36
Stannary EU staff and inquiriesOurs2,940ControllerGDPR Art 33AwarenessCal. hours72
Any material incidentn/a0Registrant subsidiarySEC Item 1.05DeterminationBus. days4
Payment card datan/a0Out of scopeTokenized, never storednonenone0

The same data set appears twice when two instruments attach to it, because they have different triggers and different units. The out-of-scope row earns its place: without it, somebody asks the question during the incident and nobody is sure.

RiverAffected Population

Affected Population

Scope expressed in records and organizations, because that is the unit every obligation is written in. Servers are not.

SegmentOrganizationsRecordsTightest obligation
US customers with a 24h BAA clause27118,40024 calendar hours
US customers with no notice clause1441,90060 calendar days
EU clinics under a processor DPA69,26036 calendar hours
Stannary EU records, controller—2,94072 calendar hours
Total47172,500

Split by organization because notice clauses attach per agreement, and a total of 172,500 tells nobody which of 41 customers to call first. All 27 contractual notices went out fourteen hours inside the clause. The Article 33 filing used six and a half of the nine available working hours.

  • How the Clocks WorkThe three units and four triggers behind breach deadlines, and why your customer's clock may start with yours.
  • Worked Example, Stannary HealthA made-up health software vendor's Friday breach, with seven deadlines computed from one timestamp.
  • Declaration and RolesWho can declare, the first four minutes, three severity levels, five roles and the shift handover.
  • Response Plan by ScenarioFive scenarios, from data theft to a vendor breach, each under the same five headings.
  • Notification TemplatesA fill-in customer notice, what regulator and individual notices must say, and a brief for support staff.
  • Legal and Regulatory NoteThe four questions to settle with counsel before an incident, recorded as dated decisions.
  • Data and Obligation Map sheetEach data set with its system, owner, record count, your role and every law or contract that applies.
  • Notification Deadline by Obligation sheetEvery deadline from one discovery time, with its unit, the working hours inside and an owner.
  • Affected Population sheetAffected records and organizations, grouped by the tightest notice each group is owed.
  • Contact and Escalation Register sheetEach role's person, after-hours route, backup and how fast they answer, with a check date.
  • Tabletop Exercise Log sheetEach exercise with its scenario, minutes to the first named deadline, findings and actions.

Common questions

What should a security incident response plan include?

A security incident response plan should say who can declare an incident, how severity is set and who leads, with a backup for each role. Add a page for each likely incident, a map of the notification deadlines your data and contracts carry, drafted notices, an after-hours contact list and a record of tabletop tests.

Who should be on an incident response team?

Start with an incident commander, a technical lead, a communications owner, someone who tracks deadlines, and a lawyer, each with a backup. The FTC's data breach guide6 says the team may also include forensics, IT, operations, human resources, investor relations and management, depending on the size and nature of your company.

How often should you test an incident response plan?

Run a tabletop exercise at least once a year. Run another after anything that changes your deadlines: a contract with a tighter notice clause, a new country, an acquisition or a real incident. Track one number from exercise to exercise: the minutes until someone names the first deadline and who owns it.

How is an incident response plan different from a disaster recovery plan?

An incident response plan handles a security event: declaring it, containing it, finding out what was taken and meeting notification deadlines. A disaster recovery plan restores systems and data after any outage, against recovery time targets. Ransomware needs both, so cross-reference them. The disaster recovery plan template covers the restore side.

What is NIST's current incident response guidance?

NIST's current guidance is SP 800-61 Revision 37, published in April 2025. It replaced the 2012 Revision 2 and ties incident response to the rest of an organization's cybersecurity risk management under the Cybersecurity Framework 2.0. Use it to check what your plan covers, then write the plan around your own data, contracts and people.

Is this security incident response plan template free?

Yes. Download the six Word documents and five CSV sheets with no signup. The sheets come filled in for a made-up health software company, so you can see a finished deadline register before you clear it. Edit with AI opens the same template in River, which builds the map from your own records.

What does Edit with AI do?

It opens the template in a free River workspace. River asks where your data lives and whose it is, for your signed customer agreements, and whether a listed parent or insurance policy applies. It builds the obligation map, tells you which deadline falls first and lists the questions for your lawyer. Every notice is yours to review and send.

Sources

  1. Google SRE. “Incident Management: Key to Restore Operations.” sre.google (opens in a new tab)

  2. EUR-Lex. “Regulation (EU) 2016/679 (General Data Protection Regulation).” eur-lex.europa.eu (opens in a new tab)

  3. eCFR. “45 CFR 164.404 — Notification to individuals.” ecfr.gov (opens in a new tab)

  4. eCFR. “45 CFR 164.410 — Notification by a business associate.” ecfr.gov (opens in a new tab)

  5. Federal Register. “Cybersecurity Risk Management, Strategy, Governance, and Incident Disclosure.” federalregister.gov (opens in a new tab)

  6. FTC. “Data Breach Response: A Guide for Business.” ftc.gov (opens in a new tab)

  7. NIST. “SP 800-61 Rev. 3, Incident Response Recommendations and Considerations for Cybersecurity Risk Management: A CSF 2.0 Community Profile.” csrc.nist.gov (opens in a new tab)

Written by

Chandler Supple

Co-founder & CEO, River

Chandler has spent the last six years working in AI, building tools for millions of users and Fortune 100 companies. In 2025, he co-founded River to build a new kind of operating system for AI agents.

Reviewed by

Brian Joseff, Co-founder, River