sam3-api

Pass

Audited by Gen Agent Trust Hub on Jul 31, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions to download the SAM3 codebase from the official Facebook Research GitHub organization and install package dependencies from the trusted PyTorch registry.
  • [COMMAND_EXECUTION]: Utilizes standard environment configuration and setup commands including conda, pip, git, and huggingface-cli. These are routine for setting up machine learning development environments.
  • [SAFE]: No indicators of prompt injection, data exfiltration, or malicious obfuscation were detected within the documentation or code snippets.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 31, 2026, 02:20 PM
Security Audit — agent-trust-hub — sam3-api