Skip to content

Latest commit

 

History

History
75 lines (59 loc) · 4.61 KB

roadmap.md

File metadata and controls

75 lines (59 loc) · 4.61 KB

Kubescape project roadmap

Planning principles

Kubescape roadmap items are labeled based on where the feature is used and by their maturity.

The features serve different stages of the workflow of the users:

  • Development phase (writing Kubernetes manifests) - example: The VS Code extension is used while editing YAMLs.
  • CI phase (integrating manifests to GIT repo) - example: GitHub action validating HELM charts on PRs.
  • CD phase (deploying applications in Kubernetes) - example: running a cluster scan after a new deployment.
  • Monitoring phase (scanning application in Kubernetes) - example: Prometheus scraping the cluster security risk.

The items in the Kubescape roadmap are split into 3 major groups based on the feature planning maturity:

  • Planning - we have tickets open for these issues with a more or less clear vision of design.
  • Backlog - features that were discussed at a high level but are not ready for development.
  • Wishlist - features that we are dreaming of in 😀 and want to push them gradually forward.

Planning 👷

  • Storing scan results in cluster

    We want Kubescape scan results (both cluster and image scan) to be stored in the cluster locally as CRDs. This will enable easier integration with results by other projects as well as with scripting via kubectl. This will also make image scan based controls to avoid accessing external resources for image vulnerability scan results.

  • Vulnerability prioritization based on workload file activity

    Implementing an eBPF agent (based on Falco or Tracee) which tracks file activity in workloads to prioritize container image vulnerabilities.

  • Prioritization engine using MITRE Attack matrix based attack chains

    Create a security issue prioritization engine that scores resources based on control based attack chains. All Kubescape controls can be arranged into attack categories of the MITRE Attack matrix. The Attack matrix categories can be connected to each other based on a theoretical attack (ie. you can't have privilege escalation without initial access). Each of the Kubescape controls is to be categorized in these system and Kubescape will calculate a priority score based on the interconnections between failed controls.

  • Integration with image registries

We want to expand Kubescape to integrate with different image registries and read image vulnerability information from there. This will allow Kubescape to give contextual security information about vulnerabilities. Container registry integration.

  • Kubescape CLI control over cluster operations
    Add functionality to Kubescape CLI to trigger operations in Kubescape cluster components (example: trigger image scans, etc.)
  • Git integration for pull requests
    Create insightful GitHub actions for Kubescape.

Backlog 📅

  • JSON path for HELM charts
    Today, Kubescape can point to issues in the Kubernetes object. We want to develop this feature so Kubescape will be able to point to the misconfigured source file (HELM).
  • Create Kubescape HELM plugin
    Producing scan results in the context of HELM.
  • Kubescape based admission controller
    Implement admission controller API for Kubescape microservice to enable users to use Kubescape rules as policies.

Wishlist 💭

  • Integrate with other Kubernetes CLI tools
    Use Kubescape as a YAML validator for kubectl and others.
  • Kubernetes audit log integration
    Connect Kubescape to the audit log stream to enable it to produce more contextual security information based on how the API service is used.
  • TUI for Kubescape
    Interactive terminal based user interface which helps to analyze and fix issues.
  • Scanning images with GO for vulnerabilities
    Images scanners cannot determine which packages were used to build Go executables and we want to scan them for vulnerabilities.
  • Scanning Dockerfile-s for security best practices
    Scan image or Dockerfile to determine whether it is using security best practices (like root containers).
  • Custom controls and rules
    Enable users to define their own Rego base rules.
  • More CI/CD tool integration
    Jenkins and etc. 😀

Completed features 🎓

  • Kubelet configuration validation
  • API server configuration validation
  • Image vulnerability scanning based controls
  • Assisted remediation (telling where/what to fix)
  • Integration with Prometheus
  • Configuration of controls (customizing rules for a given environment)
  • Installation in the cluster for continuous monitoring
  • Host scanner
  • Cloud vendor API integration
  • Custom exceptions
  • Custom frameworks