How Humanizer Detects and Fixes "Avoiding is and are" (Pattern 18)

Humanizer treats the "Avoiding is and are" rule as Pattern 18 — the "Avoiding is, are, and has" tell — by scanning text for specific verb constructions that replace simple copulas and applying targeted rewrites.

The humanizer repository by blader provides a lightweight, rule-based system for identifying robotic or overly complex phrasing in writing. Pattern 18 specifically targets sentences that bury simple statements in inflated verbal padding. This article explains exactly how the detection logic works and what transformation rules apply.

How Humanizer Identifies Pattern 18 Tells

The identification process relies on a "Watch for" list defined in [SKILL.md](https://github.com/blader/humanizer/blob/main/SKILL.md). When Humanizer scans input text, it flags sentences containing verbs or phrases that unnecessarily replace simple is, are, or has constructions.

Per the source documentation, the watch list includes:

• "serves as"
• "stands as"
• "functions as"
• "operates as"
• "marks"
• "represents"
• "boasts"
• "features"
• "offers"
• "maintains"
• "refers to"

These constructions appear in [SKILL.md](https://github.com/blader/humanizer/blob/main/SKILL.md) lines 64-71 as explicit detection targets. The system treats any match as a Pattern 18 violation requiring correction.

The Correction Rule for "Avoiding is and are"

Once flagged, Humanizer applies the "Problem" guidance documented in [SKILL.md](https://github.com/blader/humanizer/blob/main/SKILL.md) lines 66-71: "Simple verbs are replaced with longer phrases. Use is, are, or has."

The correction follows a three-step transformation:

1. Locate the subject — identify what entity the complex verb phrase modifies.
2. Replace the inflated construction — substitute the appropriate simple copula (is, are, or has).
3. Preserve remaining content — retain the descriptive elements that follow.

Before and After Examples

Before (Pattern 18 violation)	After (Humanizer correction)
Gallery 825 serves as LAAA's exhibition space for contemporary art.	Gallery 825 is LAAA's exhibition space for contemporary art.
The gallery features four separate spaces and boasts over 3,000 sq ft.	The gallery has four rooms totaling 3,000 sq ft.
The system operates as a cache layer for incoming requests.	The system is a cache layer for incoming requests.

These transformations demonstrate how Humanizer strips away lexical padding without altering core meaning.

Where Pattern 18 Is Defined

Two files govern this behavior:

File	Purpose	Direct Link
SKILL.md	Contains the complete Pattern 18 definition, watch list, and correction guidance	[SKILL.md](https://github.com/blader/humanizer/blob/main/SKILL.md)
README.md	Tabulated overview of all patterns for quick reference	[README.md](https://github.com/blader/humanizer/blob/main/README.md)

The [SKILL.md](https://github.com/blader/humanizer/blob/main/SKILL.md) file serves as the authoritative source. According to the blader/humanizer source code, pattern definitions are stored as structured markdown sections that the tool parses at runtime to build its detection and rewrite engines.

Why This Pattern Matters

Pattern 18 addresses a common AI-generated text artifact: the tendency to substitute simple state-of-being verbs with elaborate periphrastic constructions. Sentences like "The platform functions as a solution for..." signal synthetic origin more clearly than "The platform is a solution for...".

By enforcing direct copula usage, Humanizer produces prose that reads as naturally human-written rather than algorithmically inflated.

Summary

• Pattern 18 in Humanizer targets "Avoiding is, are, and has" tells.
• Detection relies on a predefined watch list in [SKILL.md](https://github.com/blader/humanizer/blob/main/SKILL.md) including phrases like "serves as," "boasts," and "functions as."
• Correction replaces these constructions with is, are, or has according to the guidance in [SKILL.md](https://github.com/blader/humanizer/blob/main/SKILL.md) lines 66-71.
• The blader/humanizer repository implements this as part of its broader skill-based text humanization framework.

Frequently Asked Questions

How do I add new phrases to Pattern 18's detection list?

Modify [SKILL.md](https://github.com/blader/humanizer/blob/main/SKILL.md) in your local clone of blader/humanizer. Add entries to the "Watch for" list under Pattern 18, following the same markdown format as existing items. The parser reads this file dynamically, so no recompilation is required.

Does Pattern 18 handle all forms of "to be" verbs?

No. Pattern 18 specifically targets constructions that replace simple is/are/has with longer phrases. Other patterns in the Humanizer framework address different "to be" misuses. Consult [README.md](https://github.com/blader/humanizer/blob/main/README.md) for the full pattern taxonomy.

Can Pattern 18 corrections change sentence meaning?

Rarely. The transformation is designed to be meaning-preserving — it removes stylistic inflation rather than semantic content. However, edge cases exist where "features" or "offers" carry subtle connotations (availability vs. possession) that has may not fully capture. Always review automated corrections in context-sensitive applications.

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