An agent skill that edits LLM prose toward Simplified Technical English
Emerged finds new products worth a look, daily — what’s new today
An agent skill that strips AI tone from Chinese text, built on a corpus study
Its 11 rewriting rules come from measuring 629 articles rather than from folklore, and the study shows several popular tells of AI writing are simply wrong.
An MIT-licensed agent skill for removing the marks of machine writing from Chinese prose. The maker compared 2.83 million characters across 629 human and AI articles, tested 26 candidate features, and kept the 11 that separated the two. The strongest was a paragraph opening with commentary that refers back to nothing, 4.4 times more common in AI text. The study also found humans use metaphor 2.4 times more and rhetorical questions 17 times more than models. Rules, result tables and Python scripts are in the repo. Install with npx skills add.
An agent skill that edits LLM prose toward Simplified Technical English
Fixes a learner's Mandarin with the smallest change and says why
An agent skill that repairs AI-written text at the structural level
A pre-send coach in your iPhone keyboard: safer, funnier or your own voice
Nine agent skills for Chinese job seekers, from open-source work to offers
An Android overlay that reads a chat and suggests three replies, never sending