README.md

---
tags:
- text-classification
- recruitment
- forensics
- security
license: mit
datasets:
- dcata004/recruiter-harvesting-dataset-v1
pipeline_tag: text-classification
---

# 🐍 V.I.P.E.R. Classification Engine (v1.0)
**Maintainer:** [Cata Risk Lab](https://huggingface.co/Cata-Risk-Lab)

## 🧠 Model Overview
This repository contains the configuration and architecture definitions for the **V.I.P.E.R.** recruitment auditing system. It defines the risk thresholds and vectorization parameters used to detect "Resume Harvesting" attacks.

## 🛠️ Configuration
The model operates on a `TfidfVectorizer` pipeline optimized for short-text classification of email subjects and bodies.

- **Risk Threshold:** 0.75 (Confidence score required to flag as SPAM)
- **Labels:** `['harvesting', 'legitimate']`
- **Dataset:** Trained on forensic recruitment data (Swiss/US/UK).

## ⚖️ Sovereign AI
Designed for local inference to protect user data privacy.