vertical-list-builder
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes local Python utility scripts (
scripts/build_vertical_query.pyandscripts/classify_vertical.py) to handle complex data logic, such as constructing SQL queries for the HubSpot warehouse and tiering contact roles based on predefined regex patterns. This approach is a secure alternative to generating logic directly within the LLM context. - [DATA_EXPOSURE]: The skill accesses and processes contact metadata including emails and phone numbers from CRM and prospecting platforms. This data is handled within the scope of the agent's defined workflow to populate HubSpot static lists and Google Sheets, aligning with the stated purpose of sales lead generation.
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data in the form of job titles from external prospecting sources (Apollo). The potential for indirect injection is significantly mitigated by the
classify_vertical.pyscript, which normalizes text and applies specific keyword rules rather than allowing the data to influence the agent's core instructions. - [CREDENTIALS_SAFE]: The skill references internal HubSpot owner IDs and Nooks user IDs to facilitate lead routing. These are non-sensitive application-level identifiers and do not expose API keys, passwords, or other authentication secrets.
Audit Metadata