crm-data-quality
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes data fetched from external CRM records and pipes the results into modification commands. 1. Ingestion points: CRM record data fetched via hubspot objects search in SKILL.md. 2. Boundary markers: The skill instructions emphasize using dry-run and refers to confirmation gating patterns in a companion skill. 3. Capability inventory: Includes hubspot objects update, hubspot objects merge, and hubspot properties create for modifying CRM data (SKILL.md). 4. Sanitization: Employs jq for data transformation and reshaping before execution.
- [COMMAND_EXECUTION]: The skill utilizes the hubspot CLI and jq utility to perform operations. These are standard tools for the intended administrative purpose and are used with appropriate safety flags like --dry-run.
Audit Metadata