bx-orm-observability
Pass
Audited by Gen Agent Trust Hub on Oct 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [SAFE]: The skill describes legitimate framework functionality for monitoring and debugging ORM activity. No malicious code, obfuscation, or unauthorized data exfiltration patterns were detected.
- [INDIRECT_PROMPT_INJECTION]: The skill creates a surface for indirect prompt injection by processing data from database events. 1. Ingestion points: The onORMQuery and onORMException payloads in OrmWatcher.bx. 2. Boundary markers: Not present in the provided examples. 3. Capability inventory: Uses writeLog to record event data. 4. Sanitization: Not implemented in the listener examples. While this is a standard observability pattern, it processes untrusted data from the database layer.
- [DYNAMIC_EXECUTION]: The skill uses Java interop via the BoxRuntime import to access the ORMService. This is a documented platform feature for interacting with BoxLang runtime services.
- [DATA_EXFILTRATION]: The skill describes the announceQueryParams setting which can log sensitive SQL parameters. The documentation includes an explicit safety warning against enabling this in production environments to prevent accidental data exposure.
Audit Metadata