foundation-models-on-device

Pass

Audited by Gen Agent Trust Hub on Mar 12, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill consists of benign documentation and Swift code examples. No malicious code or instructions were found.\n- [PROMPT_INJECTION]: Code examples demonstrate system instructions for defining model behavior, including safety guidelines to mitigate harmful outputs.\n- [DATA_EXFILTRATION]: The skill promotes on-device processing. No network operations, credential leakage, or data exfiltration techniques were identified.\n- [COMMAND_EXECUTION]: No usage of subprocesses, shell commands, or other risky system calls was detected.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 12, 2026, 08:39 AM