paper-polish

Polish only at the language level: fix grammar, word order, and phrasing while keeping every scientific claim identical. Any edit that could alter meaning, such as strengthening a conclusion, should be flagged for the author to confirm rather than applied silently.

Scan for typical AI signals like inflated phrases, marketing adjectives, formulaic transitions, and mechanical parallel structures, then replace them with concrete facts and plain wording. Vary sentence length and state real limitations honestly so the text reads like a human researcher wrote it.

Yes, it rewrites Chinese drafts into submission-level academic English by extracting the core meaning first rather than translating word by word. It keeps terminology consistent, restores implicit logical connectives, and fixes common Chinglish issues like missing articles and noun stacking.

No, the Skill never fabricates data, formulas, results, or references that are absent from the original text. When content is genuinely missing, it inserts a bracketed note suggesting what the author should add, leaving the decision to the author.

Rewrite when the problem is structural rather than linguistic, such as a broken logic chain in the introduction or a discussion unrelated to the key idea. Polishing cannot fix content-level problems, so redrafting the section first is the better path.