amplify-workflow

Pass

Audited by Gen Agent Trust Hub on Sep 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill fetches application templates from the official aws-samples GitHub organization (e.g., amplify-vite-react-template, amplify-next-template). It also references Swift packages from the official aws-amplify repository for native iOS integration.
  • [COMMAND_EXECUTION]: The skill uses the AWS CLI (aws) and the Amplify CLI (ampx) to manage cloud infrastructure, create applications, and deploy backend resources. These commands include bootstrapping environments (ampx sandbox) and configuring CI/CD pipelines (ampx pipeline-deploy).
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted project data to perform its primary function. It has a mandatory evidence chain as follows:
  • Ingestion points: The agent reads the local project structure, including the amplify/ directory, package.json, and pubspec.yaml to detect Gen2 configuration (found in references/core-web.md and references/core-mobile.md).
  • Boundary markers: None identified.
  • Capability inventory: Execution of shell commands via aws and ampx CLIs, and writing TypeScript files to define backend resources (found in references/deployment.md and references/data-backend.md).
  • Sanitization: The skill explicitly instructs users to use the secret() helper for sensitive values, ensuring credentials are stored securely in AWS Systems Manager Parameter Store rather than hardcoded in source control (found in references/functions-and-api.md).
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 30, 2026, 02:54 PM