sap-transport-overview
Pass
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: SAFEPROMPT_INJECTION
Full Analysis
- [PROMPT_INJECTION]: The skill processes untrusted data from the SAP system without explicit safety boundaries or sanitization, creating a potential surface for indirect prompt injection.
- Ingestion points: External data enters the agent context via the SAPTransport tool's 'list' and 'get' actions (specifically within transport descriptions and object names) as described in SKILL.md.
- Boundary markers: The instructions in SKILL.md lack specific delimiters or instructions for the agent to ignore potentially malicious content embedded in the SAP data.
- Capability inventory: The environment provides the SAPTransport tool with list, get, history, release, and delete actions.
- Sanitization: There is no mention in SKILL.md of sanitizing or escaping the data retrieved from the SAP system before it is processed or displayed.
- [PROMPT_INJECTION]: The instructions contain directives for the agent to apply defaults silently without asking the user, which reduces human oversight and could be leveraged to perform actions without explicit user confirmation.
- Evidence: The 'Smart Defaults' section in SKILL.md explicitly states 'apply silently, do NOT ask' for parameters like user scope and payload size.
Audit Metadata