counterfactual-generator

Pass

Audited by Gen Agent Trust Hub on Jun 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill provides structured instructions for scientific reasoning and counterfactual analysis. It contains no executable scripts, shell commands, or network activity indicators.
  • [PROMPT_INJECTION]: No attempts to override safety filters or bypass core instructions were identified. Indirect injection assessment: Ingestion point is the literature corpus; boundary markers are absent; no executable capabilities detected; sanitization is absent.
  • [DATA_EXFILTRATION]: There are no patterns suggesting unauthorized access to sensitive local files or transmission of data to external domains.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 16, 2026, 02:32 AM
Security Audit — agent-trust-hub — counterfactual-generator