localization

Pass

Audited by Gen Agent Trust Hub on Aug 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill instructions and accompanying C-Sharp scripts focus on standard Unity Localization and Addressables workflows. The batch processing functionality in L10nBatchProcessor.cs includes a mandatory human-confirmation step in the instructions to prevent accidental data loss.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from existing Unity scenes and UI text elements to map them to localization keys.
  • Ingestion points: Object.FindObjectsByType<Text> and the mapping dictionary in LocalizeAll (L10nBatchProcessor.cs).
  • Boundary markers: The skill explicitly requires an 'Ask before acting' prompt to confirm destructive batch operations with the user.
  • Capability inventory: The skill can modify scene files, add Unity components (LocalizeStringEvent), and trigger Addressables builds.
  • Sanitization: No explicit sanitization of text content is performed, but the operations are confined to the local Unity project environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 13, 2026, 04:49 PM
Security Audit — agent-trust-hub — localization