The Agent Skills Directory

[PROMPT_INJECTION]: The skill processes untrusted natural language input from users to generate executable source code, creating a surface for indirect prompt injection.
Ingestion points: Business object descriptions provided by the user (as described in the 'Input' section) are used to derive field names, entities, and business logic.
Boundary markers: The skill lacks explicit delimiters or instructions for the agent to disregard potential commands embedded within the user's business description.
Capability inventory: The skill utilizes powerful tools including SAPWrite for artifact creation and SAPActivate for deployment and activation of code in the target SAP environment.
Sanitization: No validation or sanitization mechanisms are defined to filter the user's input before it is interpolated into code templates.
[REMOTE_CODE_EXECUTION]: The skill dynamically generates and deploys executable artifacts to a remote SAP system.
Source: Source code for database tables, CDS views, behavior definitions, and ABAP classes is generated at runtime based on user-provided descriptions and pre-defined templates.
Execution: The generated artifacts are deployed to the remote environment using SAPWrite and subsequently activated using SAPActivate, making them functional within the target SAP system.

generate-rap-service