sap-clean-core-atc

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill implements a legitimate SAP auditing workflow. It utilizes SAP-specific tools (SAPRead, SAPContext, SAPDiagnose) to analyze code and relies on official SAP documentation for classification.
  • [EXTERNAL_DOWNLOADS]: Retrieves machine-readable API release states from the official SAP organization repository on GitHub (SAP/abap-atc-cr-cv-s4hc). This reference is used for data-driven classification and is a trusted source.
  • [PROMPT_INJECTION]: The skill analyzes external code provided by the SAP system, which constitutes an indirect prompt injection surface.
  • Ingestion points: ABAP source code and metadata retrieved from the SAP system (SKILL.md).
  • Boundary markers: Not explicitly defined in instructions to the agent.
  • Capability inventory: Performs system reads (SAPRead), searches (SAPSearch), and diagnostics (SAPDiagnose) on the SAP environment (SKILL.md).
  • Sanitization: Code is analyzed via regex and dependency tools but is not specifically sanitized against malicious instructions embedded in comments or strings.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 02:05 PM
Security Audit — agent-trust-hub — sap-clean-core-atc