Explore our new Cybersecurity Terms Glossary, explained by our experts.

View the Glossary View the Glossary
PRIVIA

Cybersecurity Terms Glossary

What Is KICS?

KICS — Keeping Infrastructure as Code Secure — is Checkmarx’s open-source scanner. Its distinguishing feature is breadth: thousands of Rego queries across more than a dozen IaC formats.

Illustration for the cybersecurity term KICS
Home / Blog / Cybersecurity Terms / What Is KICS?
17 November 2026

What Is KICS?

What Is KICS?

KICS stands for Keeping Infrastructure as Code Secure. It is an open-source static analysis engine released by Checkmarx under the Apache 2.0 licence, written in Go, that scans infrastructure definitions for security misconfigurations, compliance failures, and plain mistakes. What sets it apart from similar tools is coverage: where most IaC scanners started with Terraform and added other formats later, KICS was built from the start to handle a long list of them.

What It Can Read

Supported platforms include Terraform, CloudFormation, Kubernetes manifests, Helm charts, Dockerfiles, Ansible playbooks, Azure Resource Manager templates, Google Deployment Manager, Pulumi, Crossplane, Knative, OpenAPI and Swagger definitions, Serverless Framework configurations, and GitHub Actions workflows. For an organisation whose estate accumulated through several platform decisions, that breadth removes an argument about which scanner to adopt per repository.

The query library runs to several thousand checks, each tagged with a severity, a platform, a category, and mappings to compliance frameworks. OpenAPI coverage is an unusual inclusion and a genuinely useful one, since an API definition that permits an empty security scheme or an overly broad CORS policy rarely gets reviewed by anyone.

Running It

The container image is the usual route, since it bundles the queries:

docker run -t -v "$PWD":/path checkmarx/kics:latest scan \
  -p /path -o /path --report-formats "json,sarif" \
  --fail-on "high,critical"

Useful options are few and worth knowing. -t restricts the scan to a platform type when a repository mixes several and the run is slow. -i and --exclude-queries take query identifiers to include or ignore. --exclude-paths keeps vendored modules and test fixtures out of the results. --profiling exists for the moment a monorepo scan starts taking ten minutes and somebody asks why.

Results carry the file, the line, the expected value, and the actual value, which is more helpful during review than a bare rule name. SARIF output renders findings directly on the pull request diff in GitHub and GitLab.

Queries Are Rego

KICS checks are written in Rego, the language of Open Policy Agent. A query is a directory containing query.rego, a metadata.json describing severity and category, and test fixtures showing a positive and a negative case. Custom queries are loaded with -q alongside or instead of the bundled library.

Choosing Rego has a real consequence for teams: the same language already describes your Kubernetes admission policies in Gatekeeper and your Terraform plan gates in Conftest. Skills transfer, and a rule can sometimes be expressed once and enforced at more than one point in the pipeline. The cost is that Rego takes longer to learn than the YAML conditions offered by Checkov, so the first custom query is slower to write.

Inline Exceptions and Compliance Mapping

Suppressions are comments next to the resource they apply to:

# kics-scan ignore-line
# kics-scan disable=4728cd65-a20c-1a8c-5bd4-7bd1a1b0f1a7

Keeping the exception in the code, with a comment explaining why, is the difference between an auditable decision and an unexplained gap. The alternative — a growing exclusion list in the pipeline configuration — is invisible to the people writing the infrastructure.

Query metadata maps findings to CIS Benchmarks, NIST SP 800-53, PCI DSS, and similar frameworks, which makes it straightforward to produce evidence that configuration standards are enforced rather than documented. For organisations working to ISO 27001:2022, control A.8.9 on configuration management is the clause this addresses most directly.

Choosing Between the IaC Scanners

KICS, Checkov, Terrascan, tfsec, and the misconfiguration engine inside Trivy solve the same problem with different trade-offs. If your estate is Terraform and nothing else, the Terraform-specialised tools give deeper evaluation of variables and modules. If you are scanning Ansible, OpenAPI, Helm, and CloudFormation in the same organisation, KICS covers all of it with one configuration and one results format.

Whichever you choose, run one properly rather than three badly. Overlapping scanners produce three versions of the same finding with three different identifiers, and the triage cost consumes the time the tooling was meant to save.

FAQ – Frequently Asked Questions

Is KICS Free?

Yes. KICS is open source under the Apache 2.0 licence, with no account or key required, and the full query library ships with the tool. Checkmarx maintains it and also sells a commercial application security platform that integrates the same engine with central reporting and workflow features, but the command-line scanner and its queries are fully usable on their own.

How Does KICS Compare With Checkov?

They solve the same problem and overlap heavily. KICS supports a wider range of formats, including Ansible, OpenAPI definitions, and Pulumi, and its custom queries are written in Rego. Checkov has a graph engine that reasons about relationships between Terraform resources, can scan a resolved Terraform plan, and allows simpler YAML custom policies. Pick one based on which formats you actually have.

Can KICS Scan a Terraform Plan?

It scans Terraform source files rather than the JSON output of a plan, so values that come from variables, data sources, or module inputs may not be resolvable. That is the main functional gap against Checkov, which supports plan scanning directly. If your modules push most security-relevant settings into variables, verify what KICS can actually see before relying on a clean result.

Do I Need to Know Rego to Use It?

Not to run it. The bundled query library covers common misconfigurations across every supported platform and requires no policy authoring at all. Rego becomes necessary only when you want organisation-specific rules, such as mandatory tags, approved regions, or a ban on a particular module. If your team already writes Gatekeeper or Conftest policies, that knowledge transfers directly.

Related Terms

You May Be Interested In These

What Is Synthetic Data?

Cybersecurity Terms

17 November 2026

What Is Synthetic Data?

Read More
What Is ETW Tampering?

Cybersecurity Terms

17 November 2026

What Is ETW Tampering?

Read More