⬢github HTML · 22 ★ +15 since we first saw it · pushed 15 h ago
etodanik/deslop
Use psycholinguistics to deslop your agent's sloppy prose
deslop is a Claude skill (plus Python scripts) that rewrites drafts to strip out telltale patterns of AI-generated prose. Its rules come from corpus-linguistics and psycholinguistics research comparing human and model text across public corpora. A pass scans for 'tells', revises, then re-checks that no facts were dropped, iterating until the text falls in the human range.
Why now: It was just featured in a Show HN post, riding current interest in detecting and removing 'AI slop' from model-written text.
Who it is for: Claude users — writers, engineers, and anyone sending AI-drafted emails, docs, or posts — who want output that doesn't read like AI.
Stars over our 26 snapshots: 7 to 22, since 6 h ago.
Where people talked about it
API: https://socialmediatrends-api.osmike.com/v1/repos/etodanik/deslop