ai-humanizer
brandonwise 作
AI & LLMs概要
Humanize AI-generated text by detecting and removing patterns typical of LLM output.
説明
このスキルでできること
Detects AI-writing patterns in text and rewrites them to sound human. Checks 24 pattern types, 500+ flagged vocabulary terms, and statistical signals like burstiness and sentence length variation. Works on any prose, from blog posts to cover letters.
Combines pattern-based detection with statistical analysis (burstiness, type-token ratio, sentence length variation) rather than relying on word lists alone.
使いどころ
- Cleaning up ChatGPT-drafted blog posts before publishing
- Reviewing cover letters for AI detection risk before job applications
- Scoring academic essays before submission
- Stripping giveaway words like 'delve' and 'tapestry' from drafts
- Rewriting client deliverables that were written with AI assistance
ワークフロー例
User asks: Can you humanize this article I drafted with ChatGPT?
- 1
Scans text for all 24 AI patterns including Tier 1 vocabulary, filler phrases, and chatbot artifacts
- 2
Computes burstiness, type-token ratio, and sentence length variation against human baseline ranges
- 3
Flags specific problem sections with pattern labels and explanations
- 4
Rewrites flagged sections to vary rhythm, remove hedges, and replace vague attributions with concrete language
- 5
Returns the humanized version with a brief summary of changes made
Rewritten article with AI patterns removed, varied sentence length, and specific language replacing generic phrasing
インストール
clawhub install brandonwise/ai-humanizerCLIコマンドを実行するか、Claude Code、Cursor、Codexなどお使いのコーディングエージェントにプロンプトを貼り付けてください。