Detect ten fixed signatures of AI-generated prose in written drafts.
npx skills add https://github.com/lyndonkl/hermesworld --skill slop-detector-lyndonkl
Please help me install this Agent Skill. Skill: slop-detector Source: https://github.com/lyndonkl/hermesworld/tree/main/shared/skills/slop-detector Command: npx skills add https://github.com/lyndonkl/hermesworld --skill slop-detector-lyndonkl
High-intent search queries and answers about installing and using this skill.
Run a slop scan that tests the draft against ten fixed signatures, including meta-framing openers, prompt residue like "To summarize", and generic examples. Each signature is reported as clean or flagged with the offending span quoted.
Common signs include openers like "In this post", list-carried arguments, nominalization clusters, buzzword stuffing, hedge clusters, and paragraphs that all follow the same topic-support-transition shape. The scan checks each with a concrete rule.
Slop detection flags content problems such as generic framing and prompt residue, while readability checking measures sentence-level structure. Run the slop scan first, since fixing slop often changes the sentences a readability pass would score.
The no-first-person signature applies only to reflective essays over 800 words. How-to guides and methodology documents may legitimately lack "I" and should not be flagged for it.
That signature compares the submitted draft against the author's working notes to find removed caveats. If the notes are unavailable, the check is skipped quietly rather than guessed.
DeepJHConvert structured research synthesis into continuous Chinese literature review paragraphs.
hayatoscScore articles across seven dimensions and flag factual risks.
rediumvexAnalyze brand voice and archetype from web sources into a BRAND-VOICE.md guide.
mkolb22Enforce project-specific linting and formatting rules in TypeScript/JavaScript codebases.
