anti-ai-slop

name	anti-ai-slop
description	Removes AI-slop from any writing before it ships — the tells that make text read as machine-generated: filler openers, hedge-everything qualifiers, rule-of-three padding, "it's not just X, it's Y" antithesis, em-dash abuse, empty transitions, and self-congratulatory adjectives. Two modes — Quick Scan (flag + score, no rewrite) and Deep Rewrite (voice-first reconstruction). Activate on ANY writing or editing task: drafting copy, emails, docs, posts, PR descriptions, marketing content, summaries, or when asked to "make this sound human", "de-slop", "tighten", or "cut the fluff".
metadata	{"author":"synthex","version":"1.0","scope":"user-level","type":"writing-quality-gate"}

anti-ai-slop

A quality gate for prose. It catches the mechanical tells of AI-generated text and forces writing back toward a specific human voice: concrete, load-bearing, and unafraid to make a plain claim.

Applies to all writing tasks by default. Do not wait to be asked — if you are producing prose a human will read, run the gate.

Two modes

Quick Scan (default for short text, review requests, "does this read as AI?")

• Read the draft, flag every hit against the banned-phrases table and structural rules below.
• Return a slop score (0–100, lower is cleaner) and a bulleted list of specific hits with line references.
• Do NOT rewrite. The author decides.

Deep Rewrite (for "fix this", "make it human", "de-slop", or new drafts)

• Run the voice-first workflow below, producing a clean draft.
• Show the slop score before and after.
• Preserve the author's meaning and facts exactly — cut ornament, never substance.

Pick Deep Rewrite when the task is to produce or repair text; Quick Scan when the task is to assess it.