damage-numbers

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [Prompt Injection] (SAFE): No behavioral overrides or instructions targeting the AI's core logic were found.
  • [Data Exposure & Exfiltration] (SAFE): The skill does not access sensitive local files or perform network requests to untrusted domains.
  • [Remote Code Execution] (SAFE): There are no patterns involving the download and execution of remote scripts or binary files.
  • [Indirect Prompt Injection] (SAFE): Although the UI renders variable data (damage values), there are no high-privilege tools or capabilities available for exploitation.
  • [Persistence] (SAFE): No mechanisms for establishing persistence on the host system are present.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:31 PM