AWS Inspector v2 for MCP Servers — Vulnerability Scanning Overview, Pricing, and Activation

Resource types and finding types

Inspector v2 scans four resource types and produces four finding types. Understanding which finding types apply to which resources is essential for scoping costs and configuring suppression rules:

Each finding carries an Inspector Score (0.0–10.0) that starts from the CVSS v3.1 base score and is adjusted by Inspector-proprietary signals: presence in CISA KEV (Known Exploited Vulnerabilities), active exploitation evidence in threat intelligence feeds, and network reachability context for EC2. A vulnerability with a CVSS base of 5.0 can become an Inspector Score 9.2 if a working public exploit exists and the affected port is reachable from the internet. Triage on Inspector Score, not raw CVSS.

Activating Inspector v2

Activation creates the service-linked role AWSServiceRoleForAmazonInspector2 and begins scanning immediately for qualifying resources in the region.

# Enable all four resource types in us-east-1
aws inspector2 enable \
  --resource-types EC2 ECR LAMBDA LAMBDA_CODE \
  --region us-east-1

# Verify activation state for each resource type
aws inspector2 batch-get-account-status \
  --account-ids "$(aws sts get-caller-identity --query Account --output text)" \
  --query 'accounts[0].resourceState'

The response includes a state per resource type: ENABLED, ENABLING (transient), DISABLED, or SUSPENDED. EC2 scanning reaches ENABLED within a few minutes for the service itself; actual instance scan coverage depends on SSM Agent connectivity and may take 30–60 minutes for large fleets. Lambda and ECR reach ENABLED within minutes.

# Check which resources are covered (and which are not)
aws inspector2 list-coverage \
  --query 'coveredResources[].{Type:resourceType,Id:resourceId,Status:scanStatus.statusCode,Reason:scanStatus.reason}' \
  --output table | head -40

statusCode: INACTIVE means the resource is not being scanned and not being billed. For EC2 this almost always means SSM Agent is not installed or not connected to SSM — see the EC2 scanning guide. Resources with statusCode: SCANNING are active and being billed after the 30-day trial.

Multi-account activation via AWS Organizations

Inspector v2 follows the delegated-administrator pattern used by GuardDuty and Security Hub. One account is designated as the Inspector delegated administrator; it sees findings from all member accounts and configures scanning policies fleet-wide.

# Step 1 — in the management account: designate delegated admin
aws inspector2 enable-delegated-admin-account \
  --delegated-admin-account-id "111122223333" \
  --region us-east-1

# Step 2 — in the delegated admin account: set auto-enable for new org members
# (lambdaCode: false unless you want CodeGuru SAST charges)
aws inspector2 update-organization-configuration \
  --auto-enable '{
    "ec2": true,
    "ecr": true,
    "lambda": true,
    "lambdaCode": false
  }'

# Step 3 — enable Inspector in existing member accounts
aws inspector2 enable \
  --resource-types EC2 ECR LAMBDA \
  --account-ids 222233334444 333344445555 444455556666

New accounts joining the organization are automatically enrolled within 30 minutes when auto-enable is set. Member accounts cannot disable Inspector if the organization configuration has any resource type set to auto-enable: true — the delegated admin must change the org configuration first.

Pricing and cost estimation

The free trial applies at the account level — all resources in the account are free for 30 days regardless of count. After the trial, the charge is per active resource. An EC2 instance running for 15 of 31 days is charged for 15/31 × $0.11. An ECR image that is pushed once and then triggers 10 re-scans over its lifetime because of new CVE publications is charged only once at push time — re-scans are included.

# Cost estimate: count active resources before trial ends
EC2_COUNT=$(aws inspector2 list-coverage \
  --filter-criteria '{"resourceType":[{"comparison":"EQUALS","value":"AWS_EC2_INSTANCE"}],"scanStatusCode":[{"comparison":"EQUALS","value":"ACTIVE"}]}' \
  --query 'length(coveredResources)')
LAMBDA_COUNT=$(aws inspector2 list-coverage \
  --filter-criteria '{"resourceType":[{"comparison":"EQUALS","value":"AWS_LAMBDA_FUNCTION"}],"scanStatusCode":[{"comparison":"EQUALS","value":"ACTIVE"}]}' \
  --query 'length(coveredResources)')
echo "EC2 monthly: \$$(echo "$EC2_COUNT * 0.11" | bc)"
echo "Lambda monthly: \$$(echo "$LAMBDA_COUNT * 0.30" | bc)"

Inspector Score severity bands and SLA targets

What Inspector v2 does not cover

Inspector scans packages and static configuration — it does not replace runtime monitoring:

• Runtime behavior anomalies: a compromised process making unexpected outbound connections is not detected by Inspector; use GuardDuty Runtime Monitoring (see the GuardDuty guide)
• Application-layer vulnerabilities: SQL injection and XSS in your MCP server request handling are not detectable via package scanning; use a dedicated SAST tool in CI or Lambda code scanning for Lambda deployments
• Secrets in environment variables: hard-coded credentials in EC2 environment files or task definition environment variables are not scanned by Inspector; use KMS with Secrets Manager or Parameter Store
• Uptime and availability: a fully patched MCP server can still go down; Inspector does not monitor endpoint availability. AliveMCP pings every registered MCP endpoint every 60 seconds and alerts when a server fails to respond

Failure modes reference