The Agent Skills Directory

[SAFE]: The skill consists of educational reference guides for internationalization. No malicious patterns such as prompt injection, obfuscation, or data exfiltration were found.
[DATA_EXPOSURE]: The skill includes code snippets in references/locale-detection.md that read from browser storage and cookies to detect user language preferences. This is standard industry practice for localization and does not involve sensitive data harvesting.
[COMMAND_EXECUTION]: The references/translation-management.md file provides shell commands and Node.js scripts for message extraction using the fs module and glob package. These scripts are intended for developer use during the build process to scan source code for translation keys. No arbitrary command execution vulnerability exists here.
[INDIRECT_PROMPT_INJECTION]: While internationalization libraries handle external data (translation files), the provided examples use safe interpolation methods. The skill specifically highlights the importance of escaping values in SKILL.md to prevent content injection in web frameworks.

internationalization-i18n