generate-rap-service
Warn
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: MEDIUMPROMPT_INJECTIONREMOTE_CODE_EXECUTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection. It takes untrusted natural language input from the user (business object description) and uses it to generate executable code artifacts without sufficient boundary markers or sanitization of the input intent.
- Ingestion points: User-provided description in the 'Input' section of SKILL.md.
- Boundary markers: Absent; instructions do not include delimiters or warnings to ignore embedded commands in user input.
- Capability inventory: High; includes tools like SAPWrite and SAPActivate for creating and compiling code on target systems.
- Sanitization: Minimal; relies on post-generation syntax and linting checks rather than input validation.
- [REMOTE_CODE_EXECUTION]: The skill facilitates remote code execution by generating ABAP repository objects (including classes and behavior definitions) and using the SAPActivate tool to compile and enable them on a remote SAP environment.
- [COMMAND_EXECUTION]: The skill uses tools like SAPWrite, SAPActivate, and abap_generators-generate_objects to perform batch creation and activation of complex artifact stacks. This includes generating the implementation of behavior pool classes which can contain arbitrary logic.
Audit Metadata