natural-language
Pass
Audited by Gen Agent Trust Hub on Mar 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill instructions and code samples follow official Apple developer patterns for on-device linguistic analysis and translation. No malicious behavior or data exfiltration attempts were found.
- [EXTERNAL_DOWNLOADS]: The code references Apple's standard mechanism for downloading linguistic assets via requestAssets in the NaturalLanguage and Translation frameworks. These downloads are performed from Apple's official infrastructure as part of the system framework's intended operation.
- [PROMPT_INJECTION]: While the skill processes natural language input which could potentially contain indirect injections, the provided code strictly performs data analysis (tagging, embedding, translating) and does not execute the input as code or commands. No vulnerability surfaces were identified that would allow an indirect injection to influence the system state.
Audit Metadata