proprietary-data-generator

Pass

Audited by Gen Agent Trust Hub on Sep 24, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The workflow involves using web_search to gather information from external sources like Reddit and general industry sites to identify research gaps. This represents an ingestion surface for untrusted data. However, the risk is negligible as the skill lacks executable code or sensitive file access capabilities, and the agent is instructed to use the data specifically for designing survey structures and research plans.
  • [EXTERNAL_DOWNLOADS]: The skill mentions external platforms and tools such as Google Forms, Typeform, and Airtable for the purpose of data collection. These are well-known services suitable for the skill's stated intent and do not represent a security risk.
  • [SAFE]: No obfuscation, persistence mechanisms, or privilege escalation patterns were detected. The instructions are transparent and aligned with the stated purpose of market research and content creation.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 24, 2026, 02:35 AM