netbox-assurance
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill manages the ingestion of network state data from potentially untrusted or external environments to compare against the intended state in NetBox.
- Ingestion points: External network data is collected from Discovery Agents (Orb), infrastructure controllers (vCenter, Catalyst Center), and the Diode SDK (SKILL.md, references/data-sources.md).
- Boundary markers: The skill explicitly requires an 'Operator reviews' phase in the deviation lifecycle where a human examines field-level changes before taking action (references/deviation-lifecycle.md).
- Capability inventory: The skill allows for the 'Apply' action which performs write operations to the NetBox database, creating or updating entities like devices, interfaces, and IP addresses (SKILL.md, references/data-sources.md).
- Sanitization: The instructions indicate that ingested data is compared against NetBox schemas and that standard NetBox API validation is enforced during the remediation process (references/deviation-lifecycle.md).
Audit Metadata