{
  "id": "page:docs-reference-repos-blader-humanizer-research",
  "_kind": "Page",
  "_file": "wiki/docs/reference-repos/blader/humanizer/research.md",
  "_cluster": "wiki",
  "attributes": {
    "nodeKind": "Page",
    "sourcePath": "docs/reference-repos/blader/humanizer/research.md",
    "sourceKind": "repo-docs",
    "title": "blader/humanizer",
    "displayName": "blader/humanizer",
    "slug": "docs/reference-repos/blader/humanizer/research",
    "articlePath": "wiki/docs/reference-repos/blader/humanizer/research.md",
    "article": "\n# blader/humanizer\n\n- **Archetype**: utility-with-skill\n- **Stars**: 13,495\n- **Last pushed**: 2026-04-01\n- **License**: MIT\n- **Discovered**: 2026-04-12\n- **Source**: gh-search (keyword: \"claude code skills\")\n- **Skills found**: 1\n\n## Summary\nA writing editor skill that identifies and removes signs of AI-generated text. Based on Wikipedia's \"Signs of AI writing\" guide. Detects patterns including: inflated symbolism, promotional language, superficial -ing analyses, vague attributions, em dash overuse, rule of three, AI vocabulary words, passive voice, negative parallelisms, filler phrases. Includes voice calibration from writing samples.\n\n## Assessment\nStrong domain-specific skill with a well-structured multi-step procedure: (1) identify AI patterns, (2) rewrite problematic sections, (3) preserve meaning, (4) maintain voice, (5) add soul, (6) final anti-AI pass. The voice calibration feature (analyzing a writing sample to match style) is a sophisticated technique. Extractable as a specializations/writing/humanize-text process.\n\n## Extraction Priority\nMEDIUM -- The pattern detection checklist and voice calibration procedure could be extracted as a writing quality process. Useful for any content-generation workflow.\n\n## Processes\n1. **humanize-text** -- Multi-pass AI writing detection and rewriting process with voice calibration\n   - Phase 1: Scan for AI patterns (comprehensive checklist)\n   - Phase 2: Voice calibration from sample text\n   - Phase 3: Rewrite with personality injection\n   - Phase 4: Anti-AI verification pass\n\n## Plugin Ideas\n- **writing-humanizer plugin**: Wrap the humanizer as a babysitter plugin with hooks into content generation workflows. Post-processing step after any text generation task.\n\n## Library Mapping\n\n| Extractable Process | Library Status | Action | Existing Path | Target Placement |\n|-------------------|----------------|--------|---------------|------------------|\n| Humanize Text Process | NEW | Multi-pass AI writing detection and rewriting with voice calibration | - | specializations/shared/humanize-text-process.js |\n| Voice Calibration Methodology | NEW | Voice fingerprinting from writing samples (sentence length, word choice, transitions) | - | specializations/shared/voice-calibration-methodology.js |\n| AI Writing Pattern Detection | NEW | Comprehensive taxonomy of AI writing tells with detection checklist | - | specializations/shared/ai-writing-pattern-detection.js |\n\n## Plugin Marketplace Mapping\n\n| Plugin Idea | Marketplace Status | Action | Existing Plugin | Target Placement |\n|-------------|-------------------|--------|-----------------|------------------|\n| Writing Humanizer | NEW | Post-processing hooks for content generation workflows with AI detection and rewriting | - | plugins/a5c/marketplace/blueprints/writing-humanizer/ |\n\n## Implicit Procedural Knowledge\n- The \"two-pass anti-AI\" technique: first ask \"what makes this AI-generated?\", then \"fix those tells\"\n- Voice fingerprinting from writing samples (sentence length, word choice, punctuation habits, transitions)\n- Comprehensive taxonomy of AI writing tells based on Wikipedia's AI Cleanup project\n",
    "documents": []
  },
  "outgoingEdges": [],
  "incomingEdges": [
    {
      "from": "page:docs-reference-repos",
      "to": "page:docs-reference-repos-blader-humanizer-research",
      "kind": "contains_page"
    }
  ]
}