AgentSkillsCN

Ai Bust

检测面向用户文本中的 AI 式表达

SKILL.md
--- frontmatter
description: Detect AI-isms in user-facing text
allowed-tools: Bash Read Glob Grep

AI Writing Buster

Args: $ARGUMENTS (file path, glob pattern, or paste text directly)

Scans user-facing text for AI writing patterns and flags violations.

Usage

code
/ai-bust content/drafts/Poi.wiki
/ai-bust content/**/*.wiki
/ai-bust "Your text to check here"

Detection Patterns

Category 1: Structural Tells

PatternExampleFix
Em dashes"for spinners -- whether"Use comma or period
Negative-to-positive flip"Not to constrain, but to free"State directly what it does
Redundant emphasis"Share globally -- communicates across distances"Delete the repetition
Perfect threes"efficient, reliable, and scalable"Break rhythm or reduce to two
"Whether you're..." endings"Whether you're beginner or expert..."Cut entirely

Category 2: Banned Openers

Flag any paragraph starting with:

  • •"In today's fast-paced world..."
  • •"In an era where..."
  • •"In the ever-evolving landscape of..."
  • •"In the realm of..."

Category 3: Blacklisted Words

Nouns: tapestry, landscape, realm, journey, nuances

Adjectives: robust, comprehensive, crucial, pivotal, seamless, cutting-edge, game-changing, next-level

Verbs: delve, leverage, harness, unlock, foster, navigate, streamline, empower

Marketing: revolutionary, effortlessly, seamlessly

Category 4: Hedging Phrases

  • •"It's worth noting that..."
  • •"It's important to remember..."
  • •"Furthermore", "Moreover", "Additionally"
  • •"In conclusion"

Category 5: Sycophantic Openers

  • •"Absolutely!"
  • •"Certainly!"
  • •"Great question!"
  • •"You're absolutely right"

Category 6: Extended Metaphor Verbs

  • •"weaving together"
  • •"painting a picture"
  • •"crafting your..."

Output Format

For each violation found:

code
FILE:LINE | PATTERN | QUOTED TEXT
  -> Suggestion: [fix or "delete this"]

Workflow

  1. •If given a file/glob: Read the file(s) and scan for patterns
  2. •If given quoted text: Scan the text directly
  3. •Report violations grouped by category
  4. •Provide suggestions for each violation
  5. •Give overall assessment: Clean / Minor issues / Needs rewrite

Severity Levels

  • •CRITICAL: Banned openers, em dashes (dead giveaways)
  • •HIGH: Blacklisted words, hedging phrases
  • •MEDIUM: Perfect threes, sycophantic openers
  • •LOW: Extended metaphor verbs (context-dependent)

What to Scan

Focus on user-facing text:

  • •Wiki articles and drafts
  • •Templates (user-visible text in infoboxes, navboxes)
  • •Main Page content
  • •Category descriptions
  • •Release notes, changelogs

Skip:

  • •Code comments (unless they're user-visible)
  • •Variable names
  • •Config files
  • •MediaWiki system messages (unless customized)

Final Question

After reporting, ask:

"Want me to fix these issues, or just use this as a reference?"