humanize
Humanize scans text for 24 catalogued AI-writing patterns—from vocabulary and structure to statistical regularity—then rewrites to restore natural rhythm and voice. The skill preserves all facts and meaning while breaking AI-typical uniformity through targeted vocabulary replacement, structural variation, and entropy injection across five validation phases.
Humanize removes AI-generated writing patterns and rewrites text to sound naturally human while keeping all facts intact.
AI-generated summary based on this skill's SKILL.md
Install
Mathews-Tom/armory/humanize · repository language: Python
git clone https://github.com/Mathews-Tom/armory
cp -r armory/skills/humanize ~/.claude/skills/humanizenpx skillfed install Mathews-Tom/armory/humanizeFrequently asked questions
AI-generated answers based on this skill's SKILL.md and metadata
What does humanize do with AI-generated text?
Humanize scans text for 24 catalogued AI-writing patterns—from vocabulary and structure to statistical regularity—then rewrites to restore natural rhythm and voice. The skill preserves all facts and meaning while breaking AI-typical uniformity through targeted vocabulary replacement, structural variation, and entropy injection across five validation phases.
Can humanize remove AI patterns from my writing?
Yes. Humanize detects and removes AI-generated writing patterns by identifying vocabulary choices, sentence structures, and statistical regularity typical of AI output. It then rewrites affected sections with natural human prose rhythm, vocabulary variety, and authentic voice while keeping all original facts intact.
How does humanize make text sound more human?
Humanize restores natural rhythm and voice by breaking AI-typical uniformity through targeted vocabulary replacement, structural variation, and entropy injection. It replaces predictable phrasings with authentic alternatives, varies sentence length and complexity, and introduces the natural irregularities that characterize genuine human writing.
Does humanize detect if text sounds like ChatGPT?
Yes. Humanize detects and reports AI-writing tells in text with severity ratings. It identifies 24 catalogued patterns that signal AI generation, including vocabulary markers, structural uniformity, and statistical regularity, then flags these issues so you can understand what makes content sound machine-generated.
Can humanize fix specific AI-sounding paragraphs?
Yes. Humanize can rewrite targeted sections without requiring a full document rewrite. You can point it to specific paragraphs that sound robotic or formulaic, and it will apply natural language transformation to just those areas while leaving the rest of your content unchanged.
What license does humanize use?
Humanize is released under the MIT license, which permits free use, modification, and distribution for both commercial and private purposes with minimal restrictions.
SKILL.md
rendered from the published skill — quoted content, verbatim
Humanize: AI Pattern Detection and Removal
Remove AI-generated writing patterns from text. Produce natural, human-sounding output that preserves meaning.
This is not a generic rewriter. It targets specific, documented AI-writing patterns catalogued by Wikipedia's WikiProject AI Cleanup from thousands of observed instances.
Workflow
Five phases. Each phase has a clear input, transformation, and output. Do not skip phases.
Phase 1: Detection Scan
Read the input text. Load references/detection-patterns.md. Scan for two categories of signals:
A. Lexical patterns (the 24 catalogued AI-writing patterns):
| Category | Patterns
(truncated - see the full file via the links below)
Read as markdown · JSON record · Browse the source repository
File tree — 9 files
skills/humanize/SKILL.md
skills/humanize/evals/cases.yaml
skills/humanize/examples/academic.md
skills/humanize/examples/blog.md
skills/humanize/examples/professional.md
skills/humanize/references/detection-patterns.md
skills/humanize/references/statistical-signals.md
skills/humanize/references/style-guide.md
skills/humanize/references/transformation-rules.md