What Should a Software Security Incident Response Plan Include?
A software security incident response plan should name the responders, define their containment authority, and explain how to preserve evidence, revoke access, verify recovery, and communicate. Agree these responsibilities with your build partner before an incident occurs, including who acts outside normal working hours. Use the checklist below to turn those decisions into a plan people can follow.
What should your incident response checklist include?
Ask your build partner to complete this checklist with you. Each row needs a named owner, a backup, and a written decision or procedure.
| Plan element | What to agree before an incident |
|---|---|
| Reporting | Where to report suspected incidents and how to reach a responder |
| Response leadership | Who coordinates decisions and keeps the incident record |
| Containment authority | Who may disable accounts, integrations, features, or the service |
| Evidence preservation | What to collect, where to protect it, and who may access it |
| Access revocation | Who removes compromised access and verifies the result |
| Recovery | What must pass before affected services return |
| Communication | Who prepares, approves, and sends each type of update |
| Notification review | Who assesses applicable law and contractual obligations |
| Closure | Who accepts recovery and owns follow-up work |
Keep the plan accessible through an agreed route outside the affected application. Include a fallback contact method if normal messaging or sign-in is unavailable.
Who should respond, and who makes the decisions?
Name people for the following roles, even if one person holds several roles in a small team:
- Incident lead: coordinates the response, assigns actions, and records decisions.
- Technical responder: investigates the affected software and carries out approved changes.
- Business decision-maker: accepts business disruption and approves recovery.
- Communications owner: maintains consistent updates for staff, customers, and partners.
- Legal or contractual adviser: assesses notification questions and other obligations.
Give each role a backup. Record contact methods, availability, and the escalation route if nobody answers. A shared support address alone is not a response roster.
Separate acknowledgement from resolution in your support agreement. Specify how the partner confirms that a report has reached someone authorised to act. Your software maintenance and support SLA should reflect the availability and responsibilities the response plan assumes.
What should trigger the plan?
Use observable situations rather than asking a founder to diagnose an attack. Suggested triggers include an unexplained administrator account, suspected credential exposure, unexpected access to customer records, or an unexpected change to payment instructions.
Allow a suspected incident to activate the plan without proof of a confirmed breach. The first responder should record what was reported, what remains unknown, and who is investigating.
Define escalation around the possible impact: sensitive information, privileged access, affected customers, and continuing activity. Avoid a system where everyone waits for a severity label before taking an already authorised protective action.
Who has authority to contain an incident?
Write an authority table for disruptive actions. For each action, state who can approve it, who can execute it, and what happens if the approver is unreachable.
For example, agree whether the technical responder may immediately suspend a suspected compromised account, pause an integration, or disable an affected feature. Reserve broader shutdown decisions for a named business owner, with an emergency fallback.
Record the expected business impact of each action. Pausing an integration may leave work waiting; restricting access may prevent staff from serving customers. Assign someone to reconcile interrupted work after recovery.
Do not make urgent containment depend on finding a contract signatory. Agree the decision boundaries, access arrangements, and emergency commercial terms in advance.
How should responders preserve evidence?
Make evidence preservation a specific assignment. Ask the technical responder to identify relevant activity records, access changes, alerts, messages, and system state before an incident happens.
The procedure should require responders to:
- Record what they observed and when, using a consistent time reference.
- Preserve relevant original records where practicable.
- Note who collected each item and how it was collected.
- Store evidence in a restricted location.
- Record containment and recovery actions alongside the investigation.
- Seek advice before deleting affected resources or running destructive cleanup.
Avoid an absolute rule that nothing may change until evidence is collected. Where activity is continuing, the incident lead should coordinate preservation with urgent containment and record any evidence trade-off.
Keep sensitive evidence out of general chat channels and customer updates. Agree retention and disposal with the appropriate adviser.
What access should be revoked or replaced?
Use an access inventory rather than a password-only checklist. Include user accounts, administrator permissions, active sessions, integration credentials, automated service accounts, and recovery methods.
For each affected access route, assign a responder to remove or replace it and verify the result. Ask the partner to explain which separate actions are needed for passwords, sessions, tokens, and connected applications in your product.
Review permissions and recovery settings before issuing replacement access. Specify how replacement credentials reach the right person securely, and remove temporary emergency access when it is no longer needed.
For the underlying distinctions, read authentication versus authorization. Make access revocation a checked action with recorded evidence, rather than an assumption that changing one password settles it.
What must pass before recovery is approved?
Define recovery as a set of acceptance checks. A page loading successfully should not be the only condition for returning to service.
Require the technical responder to document the suspected entry route, the corrective action, and any remaining uncertainty. Recovery checks should cover:
- Removal or restriction of affected access.
- Verification of the relevant fix or configuration change.
- Checks for unexpected accounts and permissions.
- Validation of affected records and interrupted transactions.
- Tests of critical customer journeys and access boundaries.
- Monitoring for recurrence after service resumes.
- A fallback if the recovery attempt fails.
If restoration is necessary, agree how to select and validate the recovery point. Include reconciliation of legitimate activity that may need to be recovered separately. The backup restoration checklist is a useful companion.
Name both the technical verifier and the business approver. Record any temporary restrictions customers or staff must follow.
Who communicates, and who assesses notification duties?
Give each audience an owner: responders, business leaders, support staff, affected customers, and external parties. Define who drafts an update, who checks its accuracy, and who approves publication.
Use a short update format: confirmed facts, current impact, protective action, what the recipient should do, and when the next update will arrive. Label uncertainty clearly. Avoid speculative causes and unsupported claims that everything is safe.
Have qualified counsel assess what notification duties apply under the relevant law and contracts, including customer agreements and insurance terms. Ask them to determine recipients, required content, timing, and who must submit each notice. Do not copy a universal notification deadline into the plan.
Assign someone to maintain that assessment as facts change. Keep necessary legal review moving alongside containment and recovery.
What common mistakes should founders avoid?
Naming only the build partner. Assign business decisions and customer communication to people within your organisation too.
Leaving authority vague. Replace “contact management” with a named approver, backup, and emergency rule.
Treating restoration as closure. Require access checks, data validation, and corrective work before accepting recovery.
Practising only with everyone available. Walk through a scenario where the usual technical contact cannot respond. Check whether the backup can find the plan and exercise their authority.
After the exercise or a real incident, assign owners and acceptance criteria to each improvement. Update the plan when contacts, suppliers, access arrangements, or the product change.
Frequently asked questions
Does a small product need a formal plan?
Use a short plan proportionate to the product. It should still identify responders, decision authority, evidence handling, recovery checks, and communication owners. Keep specialist procedures separate where that makes the main plan easier to follow.
Can the build partner own the whole response?
Give the partner explicit technical responsibilities, but retain named owners for business decisions, customer relationships, and notification assessment. Agree the handoff between those responsibilities before signing the support arrangement.
How do you know the plan is usable?
Run a scenario from the first report through recovery approval. Have the named responders locate contact details, make a containment decision, identify evidence, and draft an update. Record anything they cannot complete and fix it before accepting the plan.