Setup & Installation
Or with the ClawHub CLI, for registry-managed skill folders outside a full OpenClaw workspace:
What This Skill Does
Rewrites AI-generated text to remove detectable machine patterns and add natural human voice. Checks against 24 pattern types and 500+ flagged vocabulary terms, computes statistical signals like burstiness and sentence variance, and can inject personality through asides, tangents, and strategic imperfection. Works on any text format: emails, articles, social posts, documentation.
Combines vocabulary blacklisting, pattern detection, and statistical analysis in one pass rather than relying on generic paraphrasing tools that swap words without addressing structural AI signals.
When to use it
- Cleaning up a ChatGPT-drafted cold email before sending
- Rewriting AI blog posts to pass editorial review
- Checking social media copy for obvious AI tells before posting
- Making product descriptions sound less like a press release
- Editing AI-generated documentation to match a team's writing voice
Example Workflow
Here's how your AI assistant might use this skill in practice.
User asks: Humanize this AI-generated paragraph about renewable energy adoption
- 1Scan the text against all 24 AI patterns, flagging matches like significance inflation and generic conclusions
- 2Check every word against Tier 1, 2, and 3 vocabulary lists and mark dead giveaways for removal
- 3Compute burstiness and type-token ratio to identify uniformly structured sentences
- 4Rewrite flagged sections, cut filler phrases, replace banned vocabulary with plain alternatives
- 5Add one or two parenthetical asides and vary sentence rhythm to inject natural voice
Rewritten paragraph with specific data replacing vague claims, chatbot artifacts removed, and a brief summary of each change made