Common Criteria is the international framework, standardized as ISO/IEC 15408, for specifying the security requirements of an IT product and having an accredited laboratory verify that the product meets them. In cybersecurity it is one of the oldest security standards for product-level assurance, used mainly in government procurement of firewalls, smart cards, operating systems, network devices and cryptographic modules.
Unlike organizational frameworks such as ISO 27001 or NIST CSF, Common Criteria does not certify a company or its processes. It certifies a specific product version in a specific configuration, and the evaluation methodology is published separately as ISO/IEC 18045, the Common Methodology for Information Technology Security Evaluation (CEM).
How a Common Criteria evaluation works
A Common Criteria evaluation is a structured, evidence-based assessment in which a vendor submits a product, its documentation and its development records to an independent evaluation facility, and a national certification body confirms the result. The product under evaluation is called the Target of Evaluation (TOE), and its exact boundaries are defined before testing begins.
The central document is the Security Target (ST), written by the vendor. It describes the TOE, the threats it counters, the assumptions about its operating environment and the security functional requirements it claims to meet. An ST may claim conformance to a Protection Profile (PP), an implementation-independent set of requirements written by a user community or government agency for a whole product category.
The following table summarizes the core artifacts and roles that appear in every evaluation.
| Element | Who produces it | Purpose |
|---|---|---|
| Target of Evaluation | Vendor | Defines the exact product version and configuration evaluated |
| Security Target | Vendor | States threats, objectives and claimed requirements |
| Protection Profile | User community or agency | Sets reusable requirements for a product type |
| Evaluation technical report | Accredited laboratory | Records test results and analysis findings |
| Certification report | National certification body | Confirms the verdict and its scope |
Evaluators then work through assurance activities: they review design documentation, check guidance manuals, examine configuration management and delivery procedures, perform independent functional testing and conduct a vulnerability analysis that includes penetration testing proportional to the claimed level.
Evaluation assurance levels and assurance components
Evaluation Assurance Levels (EAL1 to EAL7) are predefined packages of assurance requirements that express how deeply a product was examined, not how secure it is. A higher EAL means more rigorous evidence and analysis, which raises confidence that the claimed security functions work as described.
| Level | Description | Typical use |
|---|---|---|
| EAL1 | Functionally tested | Low-risk products needing basic confidence |
| EAL2 | Structurally tested | Commercial products with limited developer evidence |
| EAL3 | Methodically tested and checked | Products with moderate assurance needs |
| EAL4 | Methodically designed, tested and reviewed | Highest level practical for most commercial products |
| EAL5 | Semiformally designed and tested | Smart cards and secure microcontrollers |
| EAL6 | Semiformally verified design and tested | High-risk environments with specialized design |
| EAL7 | Formally verified design and tested | Very high-risk systems with small, simple TOEs |
Levels can be augmented, written for example as EAL4+ or EAL4 augmented with ALC_FLR.2. Augmentation adds individual assurance components, such as flaw remediation (ALC_FLR) or stronger vulnerability analysis (AVA_VAN.5), without moving the whole package to a higher level. Smart card evaluations frequently claim EAL5+ or EAL6+ with AVA_VAN.5, which requires resistance to attackers with high attack potential.
The current edition, CC:2022, restructured the standard into five parts and introduced more flexible packages, including support for exact conformance to Protection Profiles and PP-Modules. Certificates based on the older version 3.1 revision 5 were subject to a defined transition period.
Mutual recognition and certification schemes
Common Criteria certificates are portable across borders through the Common Criteria Recognition Arrangement (CCRA), under which member nations accept certificates issued by other certificate-authorizing members. Recognition under the CCRA is generally limited to EAL2 or to collaborative Protection Profiles (cPPs), which were introduced to make evaluations more repeatable and comparable.
In Europe, higher levels are recognized through the SOG-IS agreement, and the European Common Criteria-based cybersecurity certification scheme (EUCC), adopted under the EU Cybersecurity Act, now provides an EU-wide scheme with the assurance levels “substantial” and “high”. In the United States, the National Information Assurance Partnership (NIAP) accepts only evaluations against approved Protection Profiles for national security system procurement.
The table below compares Common Criteria with other assurance approaches that buyers often encounter.
| Standard | What is assessed | Typical outcome |
|---|---|---|
| Common Criteria | A specific IT product version | Product certificate at an EAL or PP |
| FIPS 140-3 | A cryptographic module | Validation at security levels 1 to 4 |
| ISO 27001 | An organization’s management system | Management system certificate |
| SOC 2 | Service organization controls | Independent auditor attestation report |
| FedRAMP | A cloud service offering | Government authorization to operate |
Using Common Criteria in procurement and product development
Buyers use Common Criteria certificates as objective, third-party evidence that a product meets defined requirements, which reduces the need for each customer to test the product themselves. Government tenders often make a certificate against a named Protection Profile a mandatory eligibility condition, particularly for network devices, encryption products and identity solutions.
For vendors, an evaluation is a significant investment. Depending on the level, product complexity and scheme, a project typically takes six months to more than two years and requires sustained engineering effort to produce design documentation and respond to evaluator observations. Planning the evaluated configuration early avoids costly redesign.
| Situation | Recommended approach | Reason |
|---|---|---|
| Selling network devices to governments | Conformance to a cPP | Widely recognized under the CCRA |
| Developing a smart card or secure element | EAL5+ or EAL6+ with AVA_VAN.5 | Required by payment and identity schemes |
| Serving EU public sector buyers | Certification under EUCC | Aligns with the EU Cybersecurity Act |
| Frequent software releases | Assurance continuity process | Extends certificates to minor changes |
When a certified product changes, vendors can use assurance continuity, which allows maintenance of the certificate through an impact analysis report instead of a full re-evaluation, provided the changes do not affect the security functions.
Limitations and common misconceptions
The most frequent misconception is that a higher EAL means a more secure product. In reality, the EAL measures the depth of evaluation against the claims in the Security Target; a product with modest security functions at EAL4 may protect less than a product with strong functions at EAL2. Buyers should always read the Security Target to see what was actually claimed.
- Scope is narrow: the certificate covers only the evaluated version and configuration, so a different setting can void its relevance.
- Certificates age: new vulnerabilities appear after certification, and many schemes now limit certificate validity to about five years.
- Assumptions matter: the Security Target lists environmental assumptions, such as trusted administrators, that the operator must satisfy.
- Cost and time can delay the availability of certified versions, so the certified release may lag behind the current release.
Despite these limits, Common Criteria remains a rigorous, internationally recognized method for product assurance, and the move toward collaborative Protection Profiles and EUCC aims to make evaluations faster and more consistent.
In simple terms
Common Criteria is a shared rulebook for checking whether a security product does what its maker says it does. The maker writes down its security promises, and an independent laboratory tests those promises in detail.
Think of it like a crash test for cars. The test does not make the car safe, but it shows how the car behaves under specific conditions. The assurance level tells you how thorough the testing was, not how strong the product is. The key point is to check what was promised before relying on the label.
A practical example
A national ministry plans to replace the firewalls protecting its internal network and issues a tender. The tender requires every offered firewall to be certified against the collaborative Protection Profile for network devices and the firewall module, under a scheme recognized by the CCRA.
Three vendors respond. The procurement team checks each certification report and Security Target, and finds that one vendor’s certificate covers only an older firmware release, while the version offered has new features outside the evaluated scope. That vendor is asked to provide evidence of assurance continuity, which it cannot supply in time.
The ministry selects a firewall whose certified configuration matches the planned deployment, then configures it according to the evaluated guidance documents. The result is a documented, independently verified security baseline that auditors can trace directly to the certificate.