terraform.gcp.security.gcp-vertexai-private-instance.gcp-vertexai-private-instance

profile photo of semgrepsemgrep
Author
unknown
Download Count*
LGPL Commons Clause
License

Ensure Vertex AI instances are private

Run Locally

Run in CI

Semgrep CI docs

Defintion

Edit RuleView Source
Open in Playground 
rules:
  - id: gcp-vertexai-private-instance
    patterns:
      - pattern: resource
      - pattern-inside: |
          resource "google_notebooks_instance" "..." {
          ...
          }
      - pattern-not-inside: |
          resource "google_notebooks_instance" "..." {
          ...
          no_public_ip = true
          ...
          }
    message: Ensure Vertex AI instances are private
    metadata:
      owasp:
        - A05:2017 - Broken Access Control
        - A01:2021 - Broken Access Control
      cwe:
        - "CWE-284: Improper Access Control"
      category: security
      technology:
        - terraform
        - gcp
      references:
        - https://owasp.org/Top10/A01_2021-Broken_Access_Control
      subcategory:
        - audit
      likelihood: LOW
      impact: LOW
      confidence: LOW
      license: Commons Clause License Condition v1.0[LGPL-2.1-only]
      vulnerability_class:
        - Improper Authorization
    languages:
      - hcl
    severity: WARNING

Examples

gcp-vertexai-private-instance.tf

# ok: gcp-vertexai-private-instance
resource "google_notebooks_instance" "pass1" {
  name = "pass1-instance"
  location = "us-west1-a"
  machine_type = "e2-medium"
  vm_image {
    project      = "deeplearning-platform-release"
    image_family = "tf-latest-cpu"
  }
  no_public_ip = true
}

# fail
# ruleid: gcp-vertexai-private-instance
resource "google_notebooks_instance" "fail1" {
  name = "fail1-instance"
  location = "us-west1-a"
  machine_type = "e2-medium"
  vm_image {
    project      = "deeplearning-platform-release"
    image_family = "tf-latest-cpu"
  }
  no_public_ip = false
}

# fail
# ruleid: gcp-vertexai-private-instance
resource "google_notebooks_instance" "fail2" {
  name = "fail2-instance"
  location = "us-west1-a"
  machine_type = "e2-medium"
  vm_image {
    project      = "deeplearning-platform-release"
    image_family = "tf-latest-cpu"
  }
}