How Stop Slop Handles Extreme Qualifiers like "Every" and "Never"

Stop Slop treats sweeping words such as every, always, never, everyone, and nobody as lazy extremes that create false authority, forcing writers to replace them with concrete, quantifiable language.

The open-source hardikpandya/stop-slop repository provides a structured skill for eliminating vague, AI-generated prose. By targeting extreme qualifiers, Stop Slop ensures that claims are backed by specific data rather than blanket assertions.

Core Rule Definition

The foundation of this behavior is Rule 4 defined in SKILL.md. This rule explicitly prohibits lazy extremes that perform vague work in declarative statements.

According to lines 21-22 of SKILL.md, the rule states: "Be specific. No vague declaratives… No lazy extremes (“every,” “always,” “never”) doing vague work."

This directive transforms how language models process sweeping generalizations. Instead of accepting "every user loves this feature," the skill triggers a rewrite requiring the author to specify exact metrics or observed behaviors.

Structural Catalog in Word Patterns

Stop Slop maintains a formal registry of problematic patterns in references/structures.md. The "Lazy extremes" entry appears in the Word Patterns table, cataloging specific terms that signal vague authority.

Lines 33-34 of references/structures.md identify these qualifiers as a distinct structural category requiring elimination. The catalog includes:

• every / everyone / everybody
• always
• never / nobody

Each entry maps to a replacement strategy demanding specific, evidence-based alternatives.

Enforcement Workflow

When processing text, Stop Slop applies a checklist-based scanning mechanism. The "Quick Checks" section triggers automatic detection of extreme qualifiers during the revision phase.

If the scan detects lazy extremes, the system flags the text for immediate rewrite. Authors must replace vague terms with concrete parameters:

• Convert "every user" to "the 1,200 users who completed onboarding"
• Replace "never fails" with "has not been observed failing in the last 30 days"
• Substitute "always delivers" with "delivered successfully in 99.4% of test cases over 90 days"

This workflow removes the "vague authority" signal common in AI-generated drafts by mandating quantifiable precision.

Practical Implementation Examples

Before and After Corrections

The following example demonstrates how Stop Slop transforms prose containing extreme qualifiers:

Our product is every competitor’s nightmare. It never fails and always delivers perfect results.

After applying the skill using SKILL.md rules:

Our product outranks 8 of 10 leading competitors in benchmark tests. In 30 days of production monitoring, it showed a 0% failure rate and met all 15 performance targets consistently.

System Prompt Integration

To enforce these rules in Claude or other LLMs, reference the skill directly in the system prompt:

system: |
  You are using the "Stop Slop" skill. Follow the core rules defined in
  https://github.com/hardikpandya/stop-slop/blob/main/SKILL.md.
  When you encounter lazy extremes (every, always, never, …), rewrite the
  sentence with specific data or a concrete claim.

Batch Processing with Python

For programmatic enforcement across multiple documents, implement a wrapper that injects the Stop Slop constraints:

from pathlib import Path
import json

def apply_stop_slop(text: str) -> str:
    """Simple wrapper that sends `text` to Claude with the Stop Slop system prompt."""
    prompt = f"""You are using the Stop Slop skill (see SKILL.md). Rewrite any lazy extremes (every, always, never, ...) with specific language.\n\n{text}"""
    # send to Claude (API call omitted for brevity)

    response = claude_api.complete(prompt)   # ← placeholder

    return response.text

sample = "Our service never crashes and every user loves it."
print(apply_stop_slop(sample))

# → "Our service has a 0% crash rate in the last 90 days, and 94% of surveyed users rated it highly."

Summary

• Rule 4 in SKILL.md explicitly bans lazy extremes including every, always, and never
• references/structures.md catalogs these terms in the Word Patterns table for systematic detection
• The enforcement workflow requires replacing vague qualifiers with specific data points or time-bound observations
• Integration supports both manual editing workflows and automated batch processing via API calls

Frequently Asked Questions

What specific words does Stop Slop classify as extreme qualifiers?

Stop Slop targets every, always, never, everyone, everybody, and nobody as lazy extremes. According to references/structures.md, these words create false authority by implying universal truths without supporting evidence.

How does Stop Slop enforce the replacement of extreme qualifiers?

When the skill's checklist detects extreme qualifiers during the Quick Checks phase, it flags the text for rewrite. The author must then substitute the vague term with concrete, quantifiable language—such as replacing "never" with "has not been observed in the last 30 days" or "every user" with a specific sample size and metric.

Can Stop Slop be integrated into automated content pipelines?

Yes. You can implement Stop Slop enforcement programmatically by including the skill reference in system prompts for Claude or other LLMs. The repository provides patterns for batch processing using Python wrappers that inject the SKILL.md rules into API calls, enabling automated detection and correction of extreme qualifiers across large document sets.

Where are the extreme qualifier rules documented in the repository?

The primary rule appears in SKILL.md at lines 21-22, while the detailed structural catalog resides in references/structures.md at lines 33-34. The CHANGELOG.md tracks the evolution of these Word Patterns rules, including updates to how extreme qualifiers are identified and processed.

{
  "mcpServers": {
    "instagit": {
      "command": "npx",
      "args": ["-y", "instagit@latest"]
    }
  }
}