The Agent Skills Directory

[SAFE]: The skill implements standard translation workflows (quick, normal, refined) and uses clear architectural patterns for handling long documents (chunking and parallel subagents).
[SAFE]: Data handling is localized to the user's input and output files; no unexpected network operations or sensitive file access patterns were detected.
[SAFE]: The use of parallel subagents for chunking is a common performance optimization for LLM-based translation and is implemented here with appropriate context injection and consistency checks.
[SAFE]: Reference files provide legitimate guidelines for glossary management and mode selection without introducing obfuscation or malicious commands.

translate