azuresql-db-schema-migration

Pass

Audited by Gen Agent Trust Hub on Aug 21, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • External Container Registry Access: The skill pulls a container image from a specific Azure Container Registry (sqldbpreview-dpgaeqhmgphzd4bk.azurecr.io). This is part of a Microsoft private preview and involves a docker login step to access the official developer engine.
  • Command Execution for Migration Tooling: The skill provides instructions for the agent to execute shell commands using docker, sqlcmd, dotnet, npm, and alembic. These operations are standard for database management tasks and include safeguards like retry loops and specific connection string parameters (TrustServerCertificate=true).
  • Indirect Prompt Injection Surface: The skill is designed to process user requests related to schema migrations (e.g., "run my migrations").
  • Ingestion points: Natural language instructions and user-provided migration files or project contexts processed via the agent.
  • Boundary markers: The skill uses specific validation rules and structured recipes to guide the agent, though it does not explicitly define data delimiters for user inputs.
  • Capability inventory: Access to docker exec, local shell execution for migration tools, and network access for pulling dependencies.
  • Sanitization: The skill relies on standard tool behavior (like sqlcmd exit codes) and environment variable wiring (SQL_CONNECTION_STRING) to manage state and inputs.
  • Use of Example Credentials: The skill uses a placeholder password (YourStr0ng_Passw0rd) in its documentation and examples. While this is not a real secret, users should follow the skill's advice to manage actual credentials using environment variables.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 21, 2026, 05:53 PM
Security Audit — agent-trust-hub — azuresql-db-schema-migration