Removes AI writing patterns from drafts while preserving voice or improving reference docs.
npx skills add https://github.com/Yassimba/loom --skill stop-the-slop-yassimba
Please help me install this Agent Skill. Skill: stop-the-slop Source: https://github.com/Yassimba/loom/tree/main/skills/stop-the-slop Command: npx skills add https://github.com/Yassimba/loom --skill stop-the-slop-yassimba
High-intent search queries and answers about installing and using this skill.
Provide the draft and request an edit. The Skill picks a stance—Preserve for personal writing or Improve for reference docs—then cuts banned words, filler phrases, and structural patterns, returning the edited draft with a What changed section.
Ask for a detect, audit, or scan. The Skill names each pattern found, quotes the offending line, and gives a short fix, without rewriting the draft or guessing whether AI wrote it.
Preserve keeps the writer's voice, structure, and opinions in personal writing like blog posts, making minimal edits. Improve rewrites reference docs like READMEs and runbooks toward the clearest version, adding facts only from sources actually read.
Yes, but only when you ask for a number. It rates the draft 1-10 across five dimensions—Directness, Rhythm, Trust, Authenticity, Density—shows the table, and recommends an edit if the total falls below 35 out of 50.
In Preserve stance, yes. The edit keeps distinctive vocabulary, cadence, humor, hedges, fragments, and strong opinions, and leaves strong human sentences untouched. The register rules apply only to sentences that were slop anyway.
Avoid using it to prove authorship; the Skill names patterns as evidence but does not claim a text was AI-written, since detectors guess. It also requires an actual draft—without one it asks for the text, audience, and goal first.
babuallyCollect raw writing fragments via interview and append them to a markdown document.
Chorus-AIDLCWrite bilingual Chorus release blog posts with Chinese-first drafts and English versions.
45black-LimitedRewrite prose with plain-English style and Orwellian clarity principles.
MizorewwDetect and rewrite AI-generated text patterns to improve readability and naturalness.
