unsloth-lora

Pass

Audited by Gen Agent Trust Hub on Feb 17, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE] (SAFE): No malicious patterns detected across all 10 threat categories.
  • [COMMAND_EXECUTION] (SAFE): The provided scripts (unsloth-lora_tool.py, unsloth-lora_tool.js) are limited to generating and printing configuration objects based on CLI arguments. They do not invoke shell commands, spawn subprocesses, or perform any administrative actions.
  • [DATA_EXFILTRATION] (SAFE): No network operations (curl, fetch, requests) or access to sensitive system paths (SSH keys, AWS credentials) were detected.
  • [REMOTE_CODE_EXECUTION] (SAFE): The scripts do not download or execute remote code. All operations are local and transparent.
  • [PROMPT_INJECTION] (SAFE): The instructions are focused on providing technical documentation and configuration workflows without attempting to bypass safety filters or override system prompts.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 17, 2026, 06:48 PM