Everything the course teaches, in one place, with the sources behind it. Sources were last reviewed on 19 November 2026.
The model
A control that operates has four layers. Evidence is the record, at every layer, that it exists and actually operated.
01 · policy · in practice
Access to production must match job responsibilities.
Evidence it leaves
The approved policy, with its owner and review date.
record access-control-policy.md · v4 · owner: security · next review 2027-01
Frameworks
The examples are illustrative connections, not claims that one control satisfies a framework.
SOC 2
Attestation examination and report, not a certification
An independent, licensed CPA firm examines a service organization’s description of its system and the controls relevant to one or more Trust Services Categories: security, availability, processing integrity, confidentiality, and privacy. Security is included in every SOC 2 examination; the other categories are added when they are relevant to the service.
Applies to
No law requires it. Service organizations usually obtain a SOC 2 report because customers ask for one to assess the risk of using them as a vendor.
Assessed by
A licensed, independent CPA firm, performing the examination under AICPA attestation standards.
Remember
A Type 1 report covers the system description and whether controls are suitably designed as of a specific date. A Type 2 report also covers whether those controls operated effectively throughout a specified period. The criteria state what must be achieved; management decides how, and defines the system boundaries.
Developer action
Ship production changes only through the defined process: a reviewed pull request, passing checks, and a recorded deployment.
Evidence
For a sample of changes from the report period: approval by someone other than the author, test results, and the matching deployment record.
Status as of 19 November 2026
The current criteria are the 2017 Trust Services Criteria with revised points of focus (2022). In October 2026 the AICPA exposed proposed 2027 criteria for public comment until 15 January 2027. If adopted, they would be effective for engagements beginning on or after 31 December 2027, with early adoption permitted.
International standard for an information security management system (ISMS), certifiable
Specifies requirements for establishing, implementing, maintaining, and continually improving an ISMS. Clauses 4 to 10 are the requirements: context and scope, leadership, planning (including risk assessment and treatment), support, operation, performance evaluation, and improvement. Annex A lists 93 reference controls; the organization compares its risk treatment against them and records its decisions in a Statement of Applicability.
Applies to
Voluntary. Organizations adopt it to manage information security risk, and often certify because customers or contracts ask for it.
Assessed by
Certification bodies, through an initial certification audit and periodic surveillance audits. Certification is typically valid for three years.
Remember
It is a management system standard. It requires you to understand your risks, choose controls to treat them, measure whether they work, and correct what does not. Annex A controls are selected through risk treatment, not adopted as a fixed checklist.
Developer action
Connect a security risk you own (for example, leaked credentials) to its treatment, an owner, a periodic check, and corrective action when the check fails.
Evidence
The risk register entry, the Statement of Applicability reference, review records, and corrective action records.
Status as of 19 November 2026
ISO/IEC 27001:2022 is current. Amendment 1:2024 added climate-change considerations to clauses 4.1 and 4.2. The transition period for certificates issued against the 2013 edition ended on 31 October 2025.
International standard for an AI management system (AIMS), certifiable
Specifies requirements for establishing, implementing, maintaining, and continually improving an AI management system. It shares the high-level structure of ISO/IEC 27001 and adds AI-specific requirements, including AI risk assessment, AI risk treatment, and AI system impact assessment. Annex A lists 38 reference controls across areas such as AI policy, resources, impact assessment, the AI system life cycle, data, information for interested parties, use of AI systems, and third-party relationships.
Applies to
Voluntary. Intended for organizations of any size that develop, provide, or use AI-based products or services.
Assessed by
Certification bodies, through certification and surveillance audits.
Remember
It manages AI as an organizational system (intended use, impacts on people, data, life cycle, oversight, and monitoring), not just model accuracy. A model provider’s certification covers the provider’s own management system, not how your organization uses the model.
Developer action
Document an AI feature’s intended use, evaluation results, human oversight, and monitoring, and send model or prompt changes through change control.
Evidence
Impact assessment input, evaluation reports against acceptance criteria, the oversight design, monitoring alerts, and change records.
Status as of 19 November 2026
ISO/IEC 42001:2023 (first edition, December 2023) is current.
U.S. federal law and regulations (45 CFR Parts 160 and 164)
HIPAA’s Privacy, Security, and Breach Notification Rules apply to covered entities (health plans, health care clearinghouses, and health care providers that conduct certain transactions electronically) and to their business associates. The Security Rule requires administrative, physical, and technical safeguards for electronic protected health information (ePHI), informed by a risk analysis.
Applies to
Covered entities and business associates, including vendors that create, receive, maintain, or transmit PHI on a covered entity’s or business associate’s behalf. Handling health data does not by itself make a company subject to HIPAA; applicability depends on whom it serves and in what role.
Assessed by
Enforced by the HHS Office for Civil Rights; state attorneys general can also bring actions. There is no official HIPAA certification, and HHS does not endorse or recognize private “certifications.”
Remember
Applicability comes first. Within scope, Security Rule implementation specifications are either required or addressable. Addressable does not mean optional: the organization must assess whether the specification is reasonable and appropriate, implement it if so, and otherwise document why and implement an equivalent alternative where reasonable and appropriate.
Developer action
Keep ePHI out of systems that are not approved for it. For example, redact it before logs reach an observability vendor that has not signed a business associate agreement.
Evidence
Data-flow diagrams, logging configuration and redaction tests, business associate agreements for vendors that do receive ePHI, access controls, and audit logs.
Status as of 19 November 2026
HHS proposed significant changes to the Security Rule in January 2025 (90 FR 898). No final rule has been published; the Unified Agenda lists the action as long-term. The current Security Rule remains in effect.
Governs the processing of personal data. A controller decides why and how personal data is processed; a processor processes it on the controller’s behalf. The principles include lawfulness, fairness and transparency; purpose limitation; data minimisation; accuracy; storage limitation; integrity and confidentiality; and accountability.
Applies to
Organizations established in the EU, and organizations outside the EU that offer goods or services to people in the EU or monitor their behaviour there.
Assessed by
National supervisory authorities, coordinated through the European Data Protection Board. Fines for some infringements can reach €20 million or 4% of total worldwide annual turnover, whichever is higher.
Remember
It concerns personal data and people’s rights, not only security. Data protection by design and by default asks you to build minimisation, purpose limits, and retention into systems, and accountability means you must be able to demonstrate that you did.
Developer action
Collect only the fields a stated purpose needs, set retention for each store, and support access and erasure requests across every place the data lives.
Evidence
Records of processing activities, the data inventory, retention configuration, and logs of handled requests.
Status as of 19 November 2026
The GDPR has applied since 25 May 2018. EU institutions are negotiating amendments proposed in the Commission’s Digital Omnibus package; the Council agreed its negotiating position on 21 October 2026. None of those GDPR amendments is law as of the review date.
U.S. Department of Defense contract program (32 CFR Part 170; DFARS 252.204-7021)
CMMC verifies that defense contractors protect Federal Contract Information (FCI) and Controlled Unclassified Information (CUI). Level 1 covers the 15 basic safeguarding requirements in FAR 52.204-21. Level 2 covers the 110 security requirements in NIST SP 800-171 Revision 2. Level 3 adds selected requirements from NIST SP 800-172.
Applies to
Department of Defense contractors and subcontractors whose contracts specify a CMMC level, based on the FCI or CUI they will handle. Requirements flow down to subcontractors that handle that information.
Assessed by
Depending on the level the contract specifies: a self-assessment, an assessment by an authorized third-party organization (C3PAO), or a government assessment by DIBCAC. Results and affirmations are entered in the Supplier Performance Risk System (SPRS).
Remember
Scope follows the information. Assets that process, store, or transmit CUI, and the assets that provide security protection for them, are in the assessment scope. Keeping CUI inside a defined boundary is therefore an engineering decision with compliance consequences.
Developer action
Keep CUI inside the systems within the assessed boundary. Before adding a tool, service, or integration that would touch CUI, have it assessed and documented in the system security plan.
Evidence
The asset inventory, a network or data-flow diagram of the boundary, the system security plan, access permissions, and audit logs.
Status as of 19 November 2026
Phase 1 of CMMC implementation began on 10 November 2025. On 13 July 2026 the Department (which now also uses the name Department of War) suspended the Phase 2 transition that had been scheduled for 10 November 2026; during the suspension, contracts may require only Level 1 (Self) or Level 2 (Self) assessments. DFARS 252.204-7012 remains in effect, including NIST SP 800-171 safeguarding and 72-hour cyber incident reporting.
Individually identifiable health information transmitted or maintained in any form, with limited exclusions such as employment records. ePHI is PHI transmitted or maintained in electronic media.
Federal Contract Information: information not intended for public release that is provided by or generated for the U.S. government under a contract, excluding public and simple transactional information.
Controlled Unclassified Information: information that a law, regulation, or government-wide policy requires or permits to be protected with safeguarding or dissemination controls.