karpathy-guidelines

Pass

Audited by Gen Agent Trust Hub on Apr 18, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is purely instructional and provides guidelines for the AI agent to follow when performing coding tasks. It contains no executable scripts, shell commands, or automation.
  • [SAFE]: No network operations or external downloads are performed. The reference to an external URL is a standard informational link to source material.
  • [SAFE]: No sensitive file access, credential usage, or data exfiltration patterns were detected.
  • [SAFE]: There are no indicators of prompt injection or attempts to bypass safety guardrails. The instructions focus on improving code quality and simplicity.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 18, 2026, 11:44 AM