schema-mapping
Pass
Audited by Gen Agent Trust Hub on Oct 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to perform proactive data inspection by running platform-specific queries (e.g.,
SELECT ... LIMIT 10) if environment access allows. This is an intended capability for analyzing schema data but should be used within the agent's authorized execution environment. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data including
SOURCE_SCHEMASandBUSINESS_CONTEXT, creating a surface for indirect prompt injection. - Ingestion points: External data enters the context through
SOURCE_SCHEMAS,TARGET_SCHEMAS, andBUSINESS_CONTEXTvariables. - Boundary markers: None identified to differentiate between instructions and untrusted data.
- Capability inventory: The skill can execute platform-specific database queries (e.g., SQL
SELECTstatements). - Sanitization: No specific sanitization or validation of the input schemas or business rules is described before the agent processes them.
Audit Metadata