foundation-models-on-device

Pass

Audited by Gen Agent Trust Hub on Aug 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill is purely educational, providing patterns and code snippets for using a hypothetical or upcoming Apple framework for on-device machine learning.
  • [COMMAND_EXECUTION]: All code examples are written in Swift and intended for application development within Xcode/iOS environments. There are no shell commands or subprocess calls that would affect the agent's host environment.
  • [DATA_EXFILTRATION]: The skill explicitly promotes on-device processing for privacy. No network operations or references to external data collection endpoints were found.
  • [PROMPT_INJECTION]: Instructions within the code snippets (e.g., "You are a cooking assistant") are standard examples of LLM system prompts and do not attempt to bypass safety filters or manipulate the agent's behavior.
  • [EXTERNAL_DOWNLOADS]: No external packages, remote scripts, or third-party dependencies are referenced for download.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 12, 2026, 01:00 PM
Security Audit — agent-trust-hub — foundation-models-on-device