DEV Community

TengLongAI2026
TengLongAI2026

Posted on

Stop-Slop: A 5.9K★ GitHub Skill That Fixes AI Writing Garbage

#ai #writing

What Stop-Slop Actually Does

Most "AI writing improvement" tools are black boxes. They take your text, apply secret rules, and give you something slightly different.

Stop-slop is the opposite. It's transparent. Open the repo and you get three files:

stop-slop/
├── SKILL.md              # The core instruction
├── references/
│   ├── phrases.md        # Banned phrases list
│   ├── structures.md     # Structural patterns to avoid
│   └── examples.md       # Before/after comparisons
└── README.md
Enter fullscreen mode Exit fullscreen mode

It's MIT licensed. You can copy it, fork it, embed it in your app, or paste it directly into your AI's system prompt.

The Three Killers: What Stop-Slop Removes

1. Banned Phrases

The list is brutal. Every phrase that makes readers' eyes glaze over:

  • Throat-clearing openings: "In this rapidly evolving digital era..."
  • Emphasis crutches: "It's worth noting", "It's important to understand"
  • Corporate jargon: "Leverage", "Align", "Circle back", "Deep dive"
  • All adverbs: Seriously. All of them.
  • Vague declarations: "This is a game-changer", "Revolutionary approach"
  • Meta-commentary: "Let's delve deeper", "We'll explore this further"
  • Hedge words: "Pretty good", "Quite interesting", "Somewhat important"

The rule is simple: if you read a phrase and think "an AI would write that" — delete it.

2. Banned Structures

This is the more interesting part. Stop-slop flags entire structural patterns:

  • Binary contrasts: "On one hand... on the other hand"
  • Negation lists: "No more manual work... no more complex setup..."
  • Dramatic fragmentation: Short. Punchy. Sentences. Broken. Apart.
  • Rhetorical questions: "What if I told you..."
  • False subjectivity: "One might argue that..."
  • Narrator voice: "As we've seen throughout this article..."
  • Passive voice everywhere: "It was determined that..."

Every developer has seen these in AI-generated blog posts. They pattern-match instantly.

3. Sentence-Level Rules

  • No Wh- sentence starts (What/Why/How at the beginning)
  • No em-dashes for dramatic pauses
  • No staccato fragments
  • No lazy superlatives ("incredible", "amazing", "unprecedented")
  • Active voice only

Before & After

Before (AI default):

In today's rapidly evolving technological landscape, it's worth noting that artificial intelligence has fundamentally transformed the way we approach content creation. Let's dive in and explore the key benefits that this revolutionary technology brings to the table.

After (stop-slop cleaned):

AI changed how we write. The old way took hours. Now it takes minutes. But speed doesn't matter if the output reads like a press release.

Why This Matters for Dev.to Writers

Dev.to readers are developers. They've seen a thousand AI-generated tutorials that all sound the same. They can spot a ChatGPT blog post in 3 seconds.

Stop-slop is not about hiding your AI usage. It's about making the output worth reading. If the AI generates text that passes stop-slop's checks, it's good enough to publish. If it doesn't — edit it until it does.

Final Take

Stop-slop is 5.9k lines of documentation (not code) that tells you exactly what's wrong with AI writing and how to fix it. No black box. No subscription. Just a list of rules you can start using today.

github.com/hardikpandya/stop-slop

I'm tenglongai2026, writing about AI tools and developer workflows. Follow me for more practical content.

Top comments (0)