The Agent Skills Directory

Command Execution: The skill provides templates for using standard administrative tools like the Azure CLI (az) and sqlcmd. These commands are used to manage cloud infrastructure as intended by the skill's primary purpose.
Credential Management: The instructions utilize placeholders for sensitive values and recommend using environment variables during connectivity tests. The skill explicitly advises against returning connection strings containing plain-text passwords, which is a positive security practice.
Firewall Configuration: The skill describes how to set firewall rules, including the specific configuration for Azure-internal service access. It provides context to differentiate this from a general internet wildcard, helping to guide correct security posture.
Indirect Prompt Injection: The skill instructs the agent to verify deployment success by parsing the output of CLI tools and SQL queries. This represents a potential surface where external tool output influences agent behavior.
Ingestion points: The agent processes text output from az sql db show, az sql server firewall-rule list, and sqlcmd execution.
Boundary markers: No specific boundary delimiters are defined for the command outputs.
Capability inventory: The skill relies on shell command execution for both provisioning and verification steps.
Sanitization: There are no explicit instructions to sanitize or validate the command output before the agent evaluates it.

provision-azure-sql-db