sap-api-policy
Pass
Audited by Gen Agent Trust Hub on Aug 7, 2026
Risk Level: SAFE
Full Analysis
- [DATA_EXPOSURE]: The skill utilizes MCP servers that interact with SAP infrastructure via environment variables for authentication (e.g., SAP_USERNAME, SAP_PASSWORD). The skill instructions include a mandatory safety rule to never request or accept these credentials from the user.
- [COMMAND_EXECUTION]: The agent uses the ARC-1 MCP server to query live SAP systems. Security is maintained through explicit instructions to keep the connection read-only, metadata-only, and to disable data previews, freestyle SQL, and write operations.
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from various external SAP sources, which represents a potential injection surface.
- Ingestion points: Data retrieved from SAP API Hub, SAP Notes, SAP Help Portal, and system metadata via ARC-1.
- Boundary markers: The skill employs a structured evidence ledger and a report template to distinguish between external source data and the agent's internal reasoning.
- Capability inventory: Network operations to official SAP domains and metadata-only access to live SAP systems.
- Sanitization: The agent is instructed to monitor tool outputs for potential prompt injection attempts and to prefer redacted summaries over raw data.
Audit Metadata