rds-sqlserver

Pass

Audited by Gen Agent Trust Hub on Jul 6, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [Ingestion of External Configuration]: The skill reads requirements.json files generated by other skills to tailor its advice. This creates a surface for indirect prompt injection where instructions could be embedded in the data. However, the skill incorporates mitigation strategies, such as requiring explicit user confirmation before taking action and validating input parameters like AWS regions.
  • [External Resource Downloads]: The skill provides instructions to download the Amazon RDS CA bundle, Microsoft package repository keys, and AWS SSM Session Manager plugins. These resources are retrieved from official domains managed by AWS and Microsoft, which is necessary for establishing secure, encrypted connections and managing database drivers.
  • [System-Level Configuration Commands]: To facilitate the installation of database drivers and connectivity tools, the skill includes commands that require administrative privileges (e.g., sudo yum install, sudo apt-key add). These operations are typical for setting up development environments on cloud instances or local systems.
  • [Secure Credential Management]: The skill strongly emphasizes the use of AWS Secrets Manager rather than hardcoding passwords in application code. By providing code examples that fetch credentials at runtime, the skill helps prevent the accidental exposure of sensitive information in logs or source control.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 6, 2026, 08:02 PM
Security Audit — agent-trust-hub — rds-sqlserver