custom-object-management
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses the
hubspotCLI tool to manage CRM object schemas, including listing, creating, and deleting object definitions. - [INDIRECT_PROMPT_INJECTION]: The skill processes external JSON files containing object schemas which represent a potential attack surface for indirect instructions.
- Ingestion points: User-supplied JSON files (e.g.,
equipment-schema.json) containing schema metadata such as labels and descriptions. - Boundary markers: The instructions do not specify any delimiters or warnings to ignore embedded instructions within schema fields.
- Capability inventory: The skill can create, update, and delete CRM schemas using the
hubspotCLI. - Sanitization: No sanitization or validation of the input JSON content is mentioned prior to execution.
Audit Metadata