Skip to main content

Periodic Risk Review Policy

1. Objective​

This policy defines the mandatory mechanisms for reviewing the application risk classification and the associated security controls throughout the entire lifecycle of each application.

Risk classification is not a one-off onboarding act - it is a dynamic state that must be kept up to date. Technical, business or regulatory changes may modify the exposure, the data processed or the potential impact of an incident, without this being immediately visible. Without structured review, the classification ages silently and the controls applied cease to be proportional to the real risk.

This policy establishes two complementary review mechanisms:

  • Time-based - review at a fixed periodic cadence, independent of changes
  • Event-based - review triggered by relevant technical or business events

2. Scope​

This policy applies to all applications with an active risk classification (L1, L2 or L3), regardless of their development status (under construction, in production, in maintenance or being discontinued).

Applications in the process of decommissioning retain the review obligation until their formal deactivation and the recording of their closure.


3. Time-based review - periodic cadence​

3.1 Mandatory minimum cadences​

LevelMinimum cadenceMandatory status
L1Annual (12 months)Mandatory
L2Half-yearly (6 months)Mandatory
L3Quarterly (3 months)Mandatory

The date of the next review must be recorded in the classification document at the time of each review. The absence of a scheduled review date is treated as non-compliance.

3.2 What to assess in each periodic review​

  • Reassessment of the three axes E, D and I based on the current state of the application
  • Verification of technical changes since the last review (new integrations, changes in exposure, new data types)
  • Verification of business changes with a potential impact on the risk level
  • Review of the status of the controls applied and of their adequacy to the current level
  • Review of the accepted residual risks and of their validity
  • Documented decision: keep the level / change the level
  • Scheduling of the next review

3.3 Outcome of the periodic review​

DecisionAction
Level maintainedRecord the reassessment with date, owner and justification; schedule the next review
Level raisedUpdate the classification; review the mandatory controls; communicate to the teams; obtain approval
Level loweredUpdate the classification with detailed technical justification; obtain AppSec approval
warning

Lowering the criticality level requires rigorous technical justification and approval by an AppSec Engineer. It is not acceptable to lower the level solely for reasons of operational convenience or to reduce the compliance effort.


4. Event-based review - mandatory triggers​

4.1 Triggers that require an immediate review​

The review of the classification must be triggered immediately — started within at most 5 working days and completed within 30 days — when any of the following events occurs:

TriggerDescription
New external integrationConnection to an external system, third-party API or cloud service not previously foreseen
New type of sensitive dataStart of processing of PII, regulated data, health data, financial data
Change in exposureOpening of a new public endpoint, authentication removed, new access channel
Significant architectural changeChange of deployment model, infrastructure migration, addition of a critical component
Introduction of automation or AIWhen it modifies exposure, the data processed or the impact of decisions
Relevant security incidentAny incident that reveals a misalignment between the classified level and the real risk
Applicable regulatory changeNew legal or normative obligation with an impact on the classification
Change of business modelChange in the target audience, jurisdiction or purpose of the application

4.2 Triggers that recommend a review​

The following events do not require an immediate review, but must be assessed by the team as potential triggers:

  • Significant change in the volume of data processed
  • Addition of new user profiles with different access levels
  • Change of the responsible team or of the security owner
  • Pentest results that reveal exposure not foreseen in the classification

4.3 Event-based review process​

  1. Identify the trigger - document the event that gave rise to the review
  2. Assess the impact on the E, D and I axes
  3. Determine whether the level must be changed
  4. Document the decision with clear technical justification
  5. Obtain approval from the approval authority appropriate to the resulting level
  6. Update the classification document and communicate to the teams
  7. Restart the calendar of periodic reviews based on the new level

5. Tool-assisted detection​

The organisation may use automated tools to detect events that may constitute review triggers (e.g. analysis of commits, PRs, configuration changes).

When a tool proposes reclassification or alerts to a potentially relevant change:

  • The alert must be assessed by an AppSec Engineer before any decision
  • Patch updates to dependencies do not constitute, in themselves, a review trigger
  • Minor/major changes to critical dependencies must be assessed
  • The final decision is always human - the tool signals, it does not decide

6. Mandatory documentation of each review​

Each review (periodic or event-based) must produce a record with:

  • Date of the review
  • Type of review (time-based / event-based)
  • Trigger (for event-based) or cadence (for time-based)
  • Reassessment of the E, D and I axes
  • Decision: level maintained / changed (with justification)
  • Status of controls and accepted risks verified
  • Person responsible for the review
  • Approval (if applicable)
  • Date of the next scheduled review

Main artefact: an entry in the application's classification document (risk-classification.md or equivalent), with the history of all previous reviews preserved.


7. Integration with the development lifecycle​

MomentExpected action
Sprint planning (L3)Check whether any event-based trigger occurred in the previous sprint
Release / go-liveConfirm that the classification is up to date and within its cadence
Incident retrospectiveCheck whether the periodic review would have detected the problem
Onboarding of a new teamCheck the date of the last review and start a new one if necessary

8. Responsibilities​

RoleResponsibility
AppSec EngineerConduct and approve reviews; monitor the review calendar per application
Tech Lead / DeveloperIdentify and report technical events that may constitute triggers
GRC / ComplianceMonitor compliance with the cadences; issue alerts for overdue reviews
Product ManagementCommunicate business changes with a potential impact on the classification
CISOOversee the review programme; ensure organisational coverage

9. Non-compliance​

The absence of a review within the cadence defined for the level constitutes non-compliance with this policy and must be:

  • Recorded as a non-conformity in the GRC system
  • Escalated to the AppSec Engineer and the CISO
  • Resolved through an immediate review and a record of the cause of the delay

10. Review and audit of this policy​

This policy must be reviewed annually or after any of the following events:

  • Change to the SbD-ToE classification model
  • Incident originating in an outdated classification
  • Regulatory change with an impact on the mandatory review cadences

The review history of all applications must be made available in internal and external security audits.


11. Normative and technical references​

ReferenceRelevance
SbD-ToE Ch. 01 - Application ClassificationE+D+I model, review cadences, event-based triggers
SbD-ToE Ch. 14 - Governance and ContractingContinuous cycle of review and reassessment of exceptions
ISO/IEC 27001 - Clause 6.1Risk assessment and treatment with periodic review
NIST SP 800-30Guide for Conducting Risk Assessments
NIST CSF - Identify (ID.RA)Risk Assessment with continuous review
SSDF PW.1Risk-based definition and review of security requirements