aws-marketplace-metering

Pass

Audited by Gen Agent Trust Hub on Oct 22, 2026

Risk Level: SAFE
Full Analysis
  • Least-Privilege Deployment Architecture: The skill instructs the user to set up a dedicated MarketplaceMeteringDeployer role and an awsmp-metering-boundary permissions boundary before deployment. This architecture ensures that the agent cannot escalate privileges beyond what is strictly necessary for creating the metering stack.
  • Input Validation and Sanitization: The registration logic in scripts/register.py uses a field allowlist (ALLOWED_REGISTRATION_FIELDS) and enforces maximum length constraints on all inputs. This mitigates risks associated with processing data from the unauthenticated AWS Marketplace redirect token.
  • Deployment Script Security: The scripts/deploy.sh script performs regex-based validation on all user-supplied variables (such as product codes and regions) before they are used in shell commands or CloudFormation overrides, preventing potential injection vulnerabilities.
  • Defense-in-Depth Observability: The SAM templates automatically provision CloudWatch alarms for critical failure modes, including CustomerNotSubscribed and UsageRecordRejected. This ensures that any integration or billing issues are surfaced to operators rather than failing silently.
  • Secure Data Handling: Sensitive identifiers are masked in logs (e.g., via the _mask helper in metering_core.py), and the design keeps buyer PII in-region within a specific customer-profile table, preventing unnecessary data exposure in the global subscribers table.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 22, 2026, 10:52 PM