skills.rest
pchalasani
pchalasaniCOMMUNITY
Explore all skills by this author →
Website𝕏@pchalasaniGitHub

remove-ai-patterns

Detect and rewrite AI-writing patterns in text using a vendored pattern catalog and Node.js detector.

2.0k|131|Updated Jul 30, 2025
Search Tags:#text rewriting#ai writing detection#ai-isms#voice profiles#writing audit#Editing#nodejs detector
One-click install
npx skills add https://github.com/pchalasani/claude-code-tools --skill remove-ai-patterns
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: remove-ai-patterns
Source: https://github.com/pchalasani/claude-code-tools/tree/main/plugins/writing/skills/remove-ai-patterns
Command: npx skills add https://github.com/pchalasani/claude-code-tools --skill remove-ai-patterns

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Text drafted or edited by AI assistants often carries recognizable machine-writing patterns ("AI-isms") such as overused vocabulary, formulaic openers, and hollow intensifiers. This Skill audits text for those patterns and rewrites it to read as natural human writing, with a deterministic detector for machine-checkable results.

Core Features & Use Cases

  • Detect-only audit: Run the bundled Node.js detector against a file to get a JSON report with an overall score, document classification, and per-issue findings with severity and suggestions.
  • Edit-in-place rewriting: Remove flagged patterns directly in files while preserving meaning, verified by a bundled validator that blocks rewrites altering code blocks, frontmatter, URLs, or headings.
  • Voice profiles and convergence: Apply casual, professional, technical, warm, or blunt voice profiles, and iterate the audit-revise cycle until a scored result reports zero issues (capped at 2 passes).
  • Use Case: A user pastes a LinkedIn draft and asks to "make this sound less like AI"; the Skill flags Tier 1 vocabulary, em-dash overuse, and formulaic openers, then returns a clean rewrite in the requested voice.

Quick Start

Ask the assistant to remove AI patterns from a specific file, for example: "Run remove-ai-patterns on draft.md and rewrite it in a professional voice."

Frequently Asked Questions about remove-ai-patterns

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I detect AI writing patterns in a file?▼

Run the bundled detector with Node.js: node "$SKILL_DIR/scripts/detect.js" FILE. It prints JSON containing an overall score, a label, a document classification, and per-issue findings with type, matched text, severity, and suggestions.

How do I rewrite text to sound less like AI?▼

Invoke the skill by asking to remove AI patterns or de-AI the text, optionally naming a voice profile (casual, professional, technical, warm, blunt). The skill audits the text, rewrites flagged spans, and can iterate until a scored detector pass reports zero issues.

Does the detector work without installing npm packages?▼

Yes. The detector is a vendored snapshot of the avoid-ai-writing engine and runs with plain Node.js, no npm install required. You must invoke it via the skill's absolute path because the command runs from the user's working directory.

When should I not use remove-ai-patterns?▼

Do not use it for ordinary writing or editing tasks, clarity rules, or formal technical-prose polish; the sibling agent-style skill covers formal technical prose. It triggers only on explicit requests to remove AI patterns, and the two skills should not be interleaved on the same document.

Why does the detector report UNSCORED with no issues?▼

The detector returns document_classification UNSCORED on empty input, text under about 10 words, or text over 10,000 words. That is a refused scan, not a clean result, so split or trim long inputs until the scan produces a scored classification.

Related Skills

Explore More Content & Communication Skills→
selrai-companyselrai-company

avoid-ai-writing

Detect and rewrite AI writing patterns into plain-language alternatives.

Official
Intermediate
Chorus-AIDLCChorus-AIDLC
1.1k

blog

Write bilingual Chorus release blog posts with Chinese-first drafts and English versions.

Official
Intermediate
pydanticpydantic
8.1k

writing-style

Rewrites technical prose to read like human documentation instead of LLM output.

Official
Intermediate
skytiger6724skytiger6724
2

humanizer

Rewrite AI-generated drafts into emotionally resonant narratives with sensory details.

Community
Intermediate