ai-writing-humanizer
hosthobbit 作
AI & LLMs概要
Automatically strip AI writing patterns and stock phrases from user-facing prose before sending.
説明
このスキルでできること
Strips common AI writing patterns from prose before it reaches readers. Covers hedging phrases, stock transitions, em dashes, rule-of-three constructs, passive voice markers, and performed authenticity cues. Operates via a Python script or importable function.
Catches specific linguistic fingerprints that general proofreaders miss, rather than relying on the author to self-edit for AI patterns.
使いどころ
- Cleaning up AI-drafted blog posts before publishing
- Preparing client-facing emails written with AI assistance
- Humanizing AI-generated product copy
- Reviewing longer AI responses before sending to customers
- Stripping AI tells from grant or proposal drafts
ワークフロー例
User asks: Humanize this 300-word AI-written support email before I send it.
- 1
Load the input text from the user
- 2
Run scripts/humanize.py --input text.txt --output cleaned.txt
- 3
Scan for hedging phrases, serial lists, em dashes, and passive voice constructs
- 4
Rewrite or remove flagged patterns
- 5
Return the cleaned text to the user
A revised version of the email with AI-characteristic phrasing removed or rewritten in a more natural voice.
インストール
clawhub install hosthobbit/ai-writing-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。