The Agent Skills Directory

[SAFE]: The skill serves as a best-practice repository for asynchronous processing. It covers architecture, backend implementation, and frontend integration across multiple modern frameworks.
[SAFE]: Implements robust security patterns for internal communications between services. For example, in the GCP implementation, it demonstrates verifying OIDC tokens using google-auth-library to ensure that requests to worker endpoints originate exclusively from Cloud Tasks.
[SAFE]: Promotes strong data validation practices by including examples using Zod (Node.js) and Vuelidate (Vue.js) to sanitize and validate payloads before processing them in background jobs.
[SAFE]: Error handling and logging examples explicitly include sanitization logic to mask sensitive fields like password, token, secret, and api_key, preventing accidental data exposure in system logs or error tracking services.
[SAFE]: Infrastructure-as-Code (Terraform) and deployment templates follow the principle of least privilege, defining specific IAM roles for Lambda and Cloud Run with restricted access to only necessary resources like SQS or Firestore.
[SAFE]: Dockerfile examples are optimized for production security, using slim base images and creating non-root system users (rails, appuser) to execute application processes.

async-processing-best-practices