The Agent Skills Directory

[Indirect Prompt Injection]: The skill facilitates the creation of a REST and GraphQL API layer over a database. Because this data is subsequently processed by an agent, it represents a potential surface for indirect prompt injection if the database contains untrusted content.
Ingestion points: Data is ingested from the Azure SQL Database container via the Data API Builder engine as described in SKILL.md.
Boundary markers: The skill recommends using entity-level permissions and roles to control access, as detailed in references/dab-config-reference.md.
Capability inventory: The skill enables shell command execution for tool installation, Docker container management, and local API interaction.
Sanitization: The implementation relies on the Data API Builder's internal schema validation and configured permissions, rather than explicit prompt-level sanitization.
[External Downloads]: The skill involves downloading the Data API Builder tool and container images from official vendor registries. These sources are considered established and are documented neutrally as part of the setup process.
[Command Execution]: The skill requires executing various shell commands to configure the environment, manage Docker containers, and interact with the API endpoints. These operations are intended for local development and follow standard usage patterns for the involved tools.
[Credentials Management]: The skill demonstrates secure credential handling by recommending the use of environment variables and the Data API Builder's env directive to avoid hardcoding sensitive passwords in configuration files. It also utilizes shell commands to prevent password exposure in terminal history.

azuresql-db-dab