ai-writing-detector

name	ai-writing-detector
description	Detect patterns of AI-generated writing in prose, creative writing, and technical documentation. Use when analyzing text for AI tells, checking if content reads like AI, reviewing drafts for machine-generated patterns, or auditing writing authenticity. Triggers on requests like "Does this read like AI?", "Check for AI patterns", "Analyze for AI writing", "Is this AI-generated?", "Detect LLM patterns".
allowed-tools	bash

AI Writing Pattern Detector

Analysis Workflow

1. Read the target file to analyze
2. Scan for patterns across these categories:
   • Vocabulary markers (overused AI words)
   • Structural patterns (em-dashes, rule of three, parallelisms)
   • Content patterns (legacy puffery, superficial analysis, hedging)
   • Style patterns (vague attribution, elegant variation)
3. Validate findings before reporting:
   • If confidence is Medium, examine context around flagged passages to rule out false positives (e.g., a single delve in an otherwise clean document may be coincidental)
   • Check whether flagged patterns cluster in one section or spread across the text — clustering reduces false-positive risk
   • For borderline cases, note ambiguity explicitly in the output rather than rounding up to High
4. Report findings with:
   • Pattern counts by category
   • Specific flagged passages with line numbers
   • Overall confidence assessment

Pattern Categories

High-Signal Vocabulary

Top indicators (see full 124-word list in references/patterns.md): delve (25x corpus increase), showcasing (9x), underscore (9x), tapestry, pivotal, meticulous, testament

Structural Tells

• Em-dash overuse: More than 1 per 200 words
• Rule of three: "X, Y, and Z" constructions in series
• Negative parallelisms: "not X, but Y", "not just...it's...", "not only...but also"
• Inline-header lists: Bullet points with bolded headers: followed by text

Content Patterns

Top indicators (see full examples in references/patterns.md):

• Legacy/symbolism puffery: "stands as a testament", "enduring legacy"
• Superficial "-ing" analyses: Sentences ending with "...reflecting its importance"

Style Markers

Top indicators (see full examples in references/patterns.md):

• Vague attribution: "Experts argue", "Observers note"
• Promotional tone: "groundbreaking", "nestled in the heart of"

Output Format

## AI Pattern Analysis: [filename]

### Summary
- **Confidence**: [Low/Medium/High] (based on pattern density)
- **Patterns found**: [count] across [categories] categories

### Vocabulary Markers ([count])
- Line X: "delve into the intricacies"
- Line Y: "pivotal role in fostering"

### Structural Patterns ([count])
- Em-dashes: [count] (rate: X per 1000 words)
- Rule of three: [count] instances
- Negative parallelisms: [count]

### Content Patterns ([count])
- Line X: Legacy puffery - "stands as a testament to..."
- Line Y: Superficial analysis - "...underscoring its significance"

### Flagged Passages
[Quoted passages with highest pattern density]

Scripts

The scripts check-ai-words and check-ai-patterns must be available on PATH; if absent, fall back to manual grep-based analysis using the pattern categories above.

Quick Check: check-ai-words

Vocabulary-only analysis with rate comparison.

check-ai-words <file> [multiplier]
# multiplier default: 3 (flag words at 3x base rate)

Full Analysis: check-ai-patterns

Comprehensive analysis including vocabulary, em-dashes, and phrase patterns.

check-ai-patterns <file> [multiplier]

Output includes:

• Em-dash rate per 1000 words (threshold: 5)
• Vocabulary violations with corpus base rates
• Common AI phrase detection
• Confidence score (0-7)

Data File: references/ai_word_rates.txt

Base rates per million words from ngrams.dev English corpus (23.6B words). This file is expected to be present in the skill bundle's references/ directory alongside references/patterns.md.

Detailed Pattern Reference

For comprehensive pattern lists with examples, see references/patterns.md. If this file is not present in the bundle, use the inline examples and vocabulary markers listed above as the primary reference.

Confidence Calibration

Pattern Density	Confidence	Notes
0-2 patterns	Low	Could be coincidence
3-5 patterns	Medium	Warrants scrutiny
6+ patterns	High	Strong AI indicators
Multiple categories	Higher	Cross-category patterns more telling

Single patterns are weak signals. Clusters of patterns, especially across categories, indicate AI generation.