geo-human-editor

What This Skill Does

The geo-human-editor is a sophisticated audit and rewrite utility designed by GEOly AI to strip away the robotic, sterile veneer often found in AI-generated content. Instead of simply generating new text, it identifies and remediates the specific mechanical patterns, overuse of "AI-speak," and structural artifacts—such as excessive bolding, markdown headers, and repetitive three-point lists—that signal to readers and search algorithms that content lacks human authenticity. By analyzing text across five critical dimensions (Formatting, Vocabulary, Mechanical Patterns, Hallucinations, and Human Voice), the skill transforms generic drafts into natural, authoritative prose that builds trust and meets modern content quality standards.

Installation

To integrate this tool into your OpenClaw environment, execute the following command in your terminal:

clawhub install openclaw/skills/skills/geoly-geo/geo-human-editor

Ensure you have Python installed, as the script depends on the underlying humanize_content.py engine located in the skill source repository.

Use Cases

This skill is essential for content creators, marketing teams, and technical writers who rely on LLMs for drafting but require a final polish that feels personal and vetted. Common use cases include:

• Polishing blog posts and articles to ensure they bypass AI-detection and engage readers.
• Auditing white papers and technical documentation for high-risk hallucinated claims that could damage credibility.
• Restoring a consistent brand voice in bulk-generated marketing copy.
• Improving the readability and rhythm of dense technical manuals by breaking up mechanical patterns.

Example Prompts

1. "Check this article for AI patterns. It feels a bit too robotic—please remove the bullet points, fix the repetitive sentence structure, and make it sound more like a conversation."
2. "I generated this draft about cloud infrastructure, but I'm worried it contains made-up statistics. Can you audit it for hallucination risks and rewrite it to sound more natural?"
3. "This content uses words like 'delve' and 'leverage' way too much. Rewrite it to be punchy and direct, and give me a human-voice score once you're done."

Tips & Limitations

• Verification is Key: While the tool flags high-risk hallucinations, it cannot verify facts against live databases. Always double-check URLs, specific pricing, and regulatory claims manually.
• Iterative Refining: Use the --score-only flag initially to identify weaknesses before committing to a full rewrite.
• Custom Vocabulary: You can supplement the tool by updating the blacklist in references/ai-vocabulary.md to match your specific brand style guide.
• Human-in-the-Loop: The tool works best when you provide the specific context or persona you want the output to embody, as the 'Human Voice' score is subjective.