ai-score

name	ai-score
description	Detect AI-generated content using Genuino API. Analyzes text and returns AI probability score, writing style classification, and optional humanization guidelines. Supports text input, files, and URLs (via Jina Reader). Use when user says "AI score," "AI detection," "check if AI," "is this AI content," "genuino check," "AI probability," "humanize text," or "content authenticity." Requires GENUINO_API_KEY.
compatibility	Requires Python 3.9+, GENUINO_API_KEY, and internet access. Optional: JINA_API_KEY for URL analysis
metadata	{"author":"nodeshub","version":"0.1.0"}
allowed-tools	Bash Read Write

AI Score

Detect whether content is AI-generated or human-written using Genuino API.

Quick Start

# Analyze a text file
python3 .claude/skills/ai-score/scripts/analyze.py --file article.txt

# Analyze with humanization guidelines (+2 credits)
python3 .claude/skills/ai-score/scripts/analyze.py --file article.txt --guidelines

# Analyze with humanization prompt (+2 credits)
python3 .claude/skills/ai-score/scripts/analyze.py --file article.txt --humanize

# Both guidelines and humanization prompt (+4 credits)
python3 .claude/skills/ai-score/scripts/analyze.py --file article.txt --guidelines --humanize

# Analyze a URL (requires Jina Reader — optional JINA_API_KEY)
python3 .claude/skills/ai-score/scripts/analyze.py --url "https://example.com/article"

# Analyze inline text (min 200 words)
python3 .claude/skills/ai-score/scripts/analyze.py --text "Your long text here..."

# Raw JSON output
python3 .claude/skills/ai-score/scripts/analyze.py --file article.txt --json

Setup

Requires GENUINO_API_KEY. If not set up: run /connect-genuino.

# Verify connection
python3 -c "
import json, urllib.request
from pathlib import Path
key = json.loads(Path('.claude/settings.local.json').read_text())['env']['GENUINO_API_KEY']
req = urllib.request.Request('https://api.genuino.ai/v1/health/basic')
req.add_header('X-API-Key', key)
req.add_header('User-Agent', 'genuino-claude-skill/0.1')
resp = urllib.request.urlopen(req, timeout=10)
print(json.loads(resp.read()))
"

For URL analysis, optionally set up Jina Reader: run /connect-nodeshub (Jina key is saved in the same settings file).

What It Returns

Field	Description
Classification	AI or HUMAN
AI Probability	0-100% score
Writing Style	Detected writing pattern (with confidence 0-100%)
Style Description	Human-readable explanation of the style
Decision Threshold	Model's decision boundary
Decision Margin	How far from the boundary (higher = more confident)

Optional extras (cost extra credits)

Flag	What you get	Extra cost
--guidelines	Per-section humanization guidelines with priorities and instructions	+2 credits
--humanize	Ready-to-use prompt to rewrite text more naturally	+2 credits

Workflow

1. Get text — user provides file, URL, or inline text
2. Validate — minimum 200 words required
3. Analyze — send to Genuino /v1/analyze endpoint
4. Report — show classification, AI probability, style, and optional guidelines
5. If AI score is high — suggest using --guidelines for specific sections to rewrite

Parameters

Parameter	Description
--text	Inline text to analyze
--file	Path to text file
--url	URL to fetch via Jina Reader and analyze
--guidelines	Include humanization guidelines (+2 credits)
--humanize	Include humanization prompt (+2 credits)
--json	Output raw JSON response

API Details

• Endpoint: POST https://api.genuino.ai/v1/analyze
• Auth: X-API-Key header
• Min text length: 200 words
• Feedback: After analysis, users can submit feedback via /v1/feedback with the analysis_id

Report

After collecting data, ask the user:

"Add results to an HTML report?"

1. New report — creates a branded HTML report in reports/
2. Existing report — appends a section to a chosen report
3. Skip — no report

Use render_report_section(data) from analyze.py, then create_report() or append_section() from report.py.

Related Skills

• connect-genuino — set up Genuino API key
• nod-content-auditor — audit content against SERP reality
• nod-content-brief — generate content briefs