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