The Agent Skills Directory

[SAFE]: The skill instructions and associated code templates do not contain any evidence of malicious intent, obfuscation, or unauthorized data exfiltration. All operations are aligned with the stated purpose of SAP development automation.
[COMMAND_EXECUTION]: The skill uses the SAPWrite tool to programmatically generate and activate CDS Data Definition (DDLS) objects in the SAP system. This is an intended functional requirement for building an analytical star schema.
[DATA_EXPOSURE]: Metadata from existing SAP objects (tables, classes, and views) is retrieved via SAPRead and SAPSearch to provide context for the model generation. This data is limited to technical structural definitions.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection through the ingestion of SAP metadata, though the risk is considered minimal given the structured nature of the data and the development context.
Ingestion points: Metadata retrieved from the SAP system via SAPRead and SAPSearch in Steps 1 and 2.
Boundary markers: None explicitly implemented; the skill relies on the agent to interpret technical metadata fields directly.
Capability inventory: Includes SAPWrite for object creation, SAPActivate for activation, and SAPManage for system status checks.
Sanitization: No specific sanitization or filtering of the retrieved metadata is performed prior to its use in code generation.

generate-analytics-star-schema