The Agent Skills Directory

Security Best Practices Implementation: The skill correctly identifies and warns against risks associated with default configurations, such as the Unauthenticated provider and broad autoentities patterns. It provides specific instructions for implementing Entra ID authentication and managed identities, which are industry-standard security practices for Azure services.
Authorization and Least Privilege: The analysis highlights the inheritance model of Data API builder (where named roles inherit from anonymous permissions). By explicitly warning users to grant the narrowest roles possible and to avoid the anonymous role during development, it helps prevent accidental data exposure. The inclusion of T-SQL commands for granting minimal database permissions (db_datareader) further supports the principle of least privilege.
Input Validation Surface (Indirect Prompt Injection Consideration): The skill describes the use of database policies and OData predicates ($filter, @claims). While these involve processing external inputs within database queries, they are standard features of the Data API builder engine designed to enforce row-level security. The skill correctly identifies that these policies should be validated and provides guidance on their proper application to restrict row access.
Secure Credential Management: The skill advises against hardcoding secrets, recommending the use of @env() or @akv() for environment variables or Azure Key Vault secrets. This is a secure approach to secret management that prevents credentials from being stored in plain text within configuration files.

dab-rest-and-graphql