deep-research

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [Prompt Injection] (SAFE): No instructions to bypass safety filters or ignore system prompts were found.
  • [Data Exposure & Exfiltration] (SAFE): No hardcoded credentials, sensitive file paths, or network exfiltration patterns were identified.
  • [Remote Code Execution] (SAFE): The skill does not include any scripts or package requirements that would download or execute code from external sources.
  • [Indirect Prompt Injection] (SAFE): While the skill is designed to process external research data, this is its primary intended purpose. It contains no internal capabilities, such as file-system access or command execution, that could be leveraged by an attacker.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 04:41 PM