Nsfw Detection

Detects adult and inappropriate content in images with multi-category classification, confidence scoring, and context-aware filtering using a local ONNX model.

NSFW Detection Skills

Skills for detecting adult/inappropriate content.

Essential Skills

1. NSFW Detector Pro

Skill ID: nsfw-detector-pro

Purpose: Advanced NSFW content detection

Features:

• Multi-category classification (porn, sexy, hentai, drawings)
• Confidence scoring (0-100%)
• Skin tone detection
• Context awareness (medical, art exceptions)
• Local ONNX model (no external API)

Installation:

npm install @clawhub/nsfw-detector

Configuration:

{
  "skill": "nsfw-detector-pro",
  "settings": {
    "categories": ["porn", "sexy", "hentai"],
    "threshold": 0.7,
    "contextAware": true,
    "allowArt": true,
    "allowMedical": true,
    "localModel": true
  }
}

Usage:

import { detectNSFW } from '@clawhub/nsfw-detector';

const result = await detectNSFW(imageUrl);
// {
//   isNSFW: true,
//   confidence: 0.92,
//   categories: { porn: 0.92, sexy: 0.15 },
//   action: 'block'
// }

Accuracy: 92-95% on explicit content, 75-80% on suggestive

Use Cases:

• Block pornography and explicit imagery
• Filter adult websites
• Protect children online

Troubleshooting:

• False positives on art? Enable allowArt
• Medical images blocked? Enable allowMedical
• Increase threshold for fewer false positives

Related Skills: url-reputation, content-blur

---

2. URL Reputation Checker

Skill ID: url-reputation

Purpose: Check if domains are adult/inappropriate sites

Features:

• Domain reputation database
• Adult site detection
• Phishing/malware detection
• Cached lookups

Installation:

npm install @clawhub/url-reputation

Use Cases:

• Block adult websites by domain
• Fast pre-filtering before image analysis
• Supplement visual detection

---

Installation

Via ClawHub

https://clawhub.ai/raghulpasupathi/nsfw-detection

Via npm

npm install @raghulpasupathi/nsfw-detection

Configuration Examples

Child Protection (Strict)

{
  "nsfw-detector-pro": {
    "threshold": 0.3,
    "blockAll": true,
    "allowArt": false,
    "allowMedical": false,
    "allowOverride": false
  }
}

Teen Protection (Balanced)

{
  "nsfw-detector-pro": {
    "threshold": 0.7,
    "allowArt": true,
    "allowMedical": true,
    "allowOverride": false
  }
}

Adult Filtering (Lenient)

{
  "nsfw-detector-pro": {
    "threshold": 0.9,
    "warnOnly": true,
    "allowOverride": true
  }
}

---

For violence detection, see VIOLENCE_DETECTION.md.