yann-lecun

Pass

Audited by Gen Agent Trust Hub on Mar 16, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: No malicious patterns or security risks were detected. The skill consists of text-based persona instructions, biographical information, and educational PyTorch code examples that do not perform unsafe operations.- [SAFE]: The skill lacks external dependencies, remote code patterns, or command execution instructions. The listed tools in metadata are standard development utilities.- [SAFE]: The persona-alignment instructions are standard for role-playing skills and do not attempt to override core agent safety filters or instructions.- [SAFE]: The skill has no actionable capabilities (e.g., network access, file writing), making it resilient to indirect prompt injection risks despite processing user queries.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 16, 2026, 06:58 AM