aws-serverless

Pass

Audited by Gen Agent Trust Hub on Feb 20, 2026

Risk Level: SAFE
Full Analysis
  • [Indirect Prompt Injection] (LOW): The skill demonstrates patterns for ingesting untrusted data from external sources such as API Gateway path parameters, SQS message bodies, and HTTP request bodies.
  • Ingestion points: Ingestion occurs in sub-skills/api-gateway-integration-pattern.md (path parameters), sub-skills/event-driven-sqs-pattern.md (SQS record body), and sub-skills/lambda-handler-pattern.md (HTTP body).
  • Boundary markers: Code templates do not include delimiters or instructions to ignore embedded commands.
  • Capability inventory: Snippets demonstrate standard AWS SDK usage for DynamoDB and SQS operations. No high-risk capabilities such as arbitrary system command execution or unauthorized network access were found.
  • Sanitization: The handlers perform basic JSON parsing but lack the validation or escaping required to prevent injection when data is subsequently used in a prompt context.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 20, 2026, 05:43 AM