The Agent Skills Directory

[SAFE]: The skill includes robust safety guidelines for handling sensitive or regulated data (PII, PHI, PCI). It explicitly instructs the agent to consult the user before accessing potentially sensitive tables and to default to a non-invasive 'profile' mode that compares statistics without transferring raw row values.
[SAFE]: Database interactions are focused on legitimate schema inspection and data parity validation. The provided SQL examples use standard system catalogs (information_schema, sys.columns) for metadata discovery.
[SAFE]: The skill documentation encourages secure authentication practices, such as Azure Active Directory and Managed Identities, rather than hardcoded credentials.
[SAFE]: No evidence of obfuscation, remote code execution, or unauthorized network activity was found. The instructions prioritize transparency by requiring a numbered plan to be presented to the user before any data operations begin.

data-parity