# Slop Cop

> **TL;DR.** Slop Cop is an automated prose linter that flags the rhetorical and structural tells of LLM-generated writing — em-dash pivots, hedge stacks, throat-clearing intros, three-item lists, filler intensifiers — and emits a JSON violation report designed for coding agents, not humans.

- **Category:** AI / Developer Tools / Code Quality
- **Stage:** validating
- **Age:** 85 days
- **Origin date:** 2026-03-10
- **First detected:** 2026-04-19
- **Canonical URL:** https://earlyterms.com/term/slop-cop
- **Sources:** 6 primary URLs

## Definition

Slop Cop is an automated prose linter that flags the rhetorical and structural tells of LLM-generated writing — em-dash pivots, hedge stacks, throat-clearing intros, three-item lists, filler intensifiers — and emits a JSON violation report designed for coding agents, not humans. It is explicitly a cliche detector, not an AI detector.

The name crystallized when [awnist](https://github.com/awnist/slop-cop) published the browser-based editor on March 10, 2026 and hit the [Hacker News front page on April 17 with 241 points](https://news.ycombinator.com/item?id=47806845). One day later, [yasyf forked it into a Go CLI](https://github.com/yasyf/slop-cop) that installs as a Claude Code / Cursor plugin and auto-revises drafts before the agent returns them.

## Example

The yasyf CLI ships 48 rules — 35 instant regex detectors plus 10 Haiku-scored sentence rules and 3 Sonnet document-tier rules — sourced from sneak's LLM_PROSE_TELLS catalog and Wikipedia's Signs of AI Writing. When a coding agent writes prose, the skill runs slop-cop, reads the JSON report, and silently rewrites flagged passages without narrating the edit.

## Analogy

A spell-checker for LLM cliches — it flags the em-dash and the 'not just X, but Y' the way a red squiggle flags 'teh'.

## Why it's emerging now

awnist's browser linter hit HN front-page on April 17, 2026 with 241 points. Within 48 hours, yasyf forked it into a Go CLI and shipped it as a Claude Code + Cursor plugin that auto-revises agent-written prose — turning a human writing tool into agent guardrails.

## Related terms

- *parent:* AI Slop
- *related:* Claude Code
- *related:* AGENTS.md
- *related:* Context Engineering
- *related:* AI Agent Identity
- *related:* Claude Agent SDK
- *competitor:* GPTZero
- *competitor:* Originality.ai
- *competitor:* Kagi SlopStop
- *related:* LLM_PROSE_TELLS
- *alias:* anti-ai-slop-writing

## Sources

1. [yasyf/slop-cop — Claude Code + Cursor plugin](https://github.com/yasyf/slop-cop)
2. [awnist/slop-cop — original browser editor](https://github.com/awnist/slop-cop)
3. [Hacker News — Slop Cop (241 points)](https://news.ycombinator.com/item?id=47806845)
4. [Show HN — Slop Cop Claude Code / Cursor Plugin](https://news.ycombinator.com/item?id=47820719)
5. [LLM_PROSE_TELLS.md — sneak's rule taxonomy](https://git.eeqj.de/sneak/prompts/src/branch/main/prompts/LLM_PROSE_TELLS.md)
6. [Wikipedia — AI slop (parent category)](https://en.wikipedia.org/wiki/AI_slop)

---
_Generated by EarlyTerms · https://earlyterms.com/term/slop-cop_