# About Ratio'd

## Mission
Ratio'd is an open, deterministic scam and phishing risk detection engine designed to protect internet users and automated agents from credential harvesting, brand impersonation, urgent financial manipulation, and malicious redirection.

## Operator & Maintainer
- **Founder & Maintainer**: Vedant Baghel
- **Email**: [vedantsbaghel.2626@gmail.com](mailto:vedantsbaghel.2626@gmail.com)
- **Website**: [https://www.ratiod.online](https://www.ratiod.online)
- **Repository**: Open-source cybersecurity tooling on GitHub

## Core Architecture
Ratio'd operates through a multi-tier detection pipeline:
1. **Deterministic Rule Engine**: Instant zero-latency recognition of known threat signatures, crypto draining patterns, and spoofed security advisories.
2. **Homoglyph & Unicode Normalization**: Neutralizes visually deceptive lookalike characters (such as Cyrillic `а` vs Latin `a` in `paypаl.com`).
3. **Levenshtein Distance Scoring**: Identifies subtle typosquatting and domain permutations targeting major banking, tech, and retail entities.
4. **VegaML Hamiltonian Physics Dynamics**: Evaluates message energy states, potential traps, and momentum shifts to differentiate between legitimate marketing campaigns and genuine security threats.
5. **Safe Peek URL Unmasker**: Recursively follows HTTP `Location` headers up to safe limits to inspect the ultimate landing page without letting arbitrary scripts run.

## Agent Protocol Support
Ratio'd is fully agent accessible:
- Supports Model Context Protocol (`/mcp.json`, `/.well-known/mcp.json`).
- Supports Agent-to-Agent card (`/a2a.json`, `/.well-known/agent.json`).
- Declares full OpenAPI 3.1.0 specifications (`/openapi.json`, `/openapi.yaml`).
- Provides Markdown representations via content negotiation (`Accept: text/markdown`).
- Provides comprehensive documentation in `/llms.txt` and `/llms-full.txt`.
