The Agent Skills Directory

[SAFE]: No malicious patterns or security vulnerabilities were detected in the skill instructions or the provided audit script. The functionality is consistent with its stated purpose of assisting with software localization.
[COMMAND_EXECUTION]: The skill utilizes standard tools like ripgrep and a local Python script to perform static analysis on the codebase. These commands are used to identify translation keys and do not involve the execution of untrusted external code.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from source files and locale JSON files. The risk is managed by the script's use of targeted regular expressions for key extraction rather than broad content processing. Furthermore, the skill provides clear instructions to avoid displaying raw system error messages to users, mitigating potential data exposure from ingested content.

i18n-expert