Arbitration & Security
Our security system is a multi-layered defense that operates in both proactive and reactive modes. We don't just resolve disputes—we use AI and strict protocol rules to prevent them.
1. "Matching Requisites" Protocol (Main Defense Against "Triangles") This is a fundamental and unbreakable rule of the Looty platform that makes "triangular" fraud impossible.
How it works:
Requisite Registration: Both Buyer (A) and Seller (B) must register their fiat requisites in their profile beforehand (e.g., Wise: [email protected], Revolut: @user-b).
Order Creation: When Seller (B) creates an order, they explicitly state the receiving method (e.g., "Sell 100 USDC for EUR only on Wise").
Method Matching: For Buyer (A) to accept this order, they must have the same payment method (i.e., Wise) added to their profile. The platform simply will not let them start the trade if they don't have Wise.
AI Verification: When Buyer (A) uploads "Receipt 1," our AI Guardian checks that: a) The payment came from Wise (type match). b) The payment came from the account ([email protected]) belonging to Buyer (A), not a third party (owner match). c) The payment was sent to Seller (B)'s Wise account.
Result: A "triangle" becomes impossible. A scammer (A) cannot accept Seller (B)'s "Wise-to-Wise" order to make a Victim (C) pay. The platform requires the payment to come specifically from the verified Wise account of Buyer (A).
2. Proactive AI Monitoring We use a fine-tuned Gemini 2.5 Flash model as our "AI Guardian," which operates in real-time during every trade. Its tasks:
Chat Analysis: The AI monitors the conversation for suspicious phrases (e.g., "I'll send it later," "my bank has issues, let's cancel").
Document Verification: The AI instantly compares "Receipt 1" and "Receipt 2," checking them against the "Matching Requisites" rule and for signs of forgery (fonts, dates).
3. Dispute Escalation Process If a user manually opens a dispute, our two-stage arbitration process is triggered:
Stage 1: AI Analysis (Automated Report): The AI Guardian immediately gathers all trade data (chat logs, both receipts, requisite matching data) and generates an unbiased report for a human auditor.
Stage 2: Manual Audit (Final Decision): Our moderator-auditor (initially the Core Team, in the future, Qrism EAO Auditors) steps in. They review the AI report and all evidence. This auditor's decision is final.
4. "Slashing" and Compensation Mechanism This is our primary tool for punishment and compensation.
Scenario: Seller (B) was caught in fraud. They received $100 in fiat from Buyer (A) but refused to send 100 USDC.
Decision:
The Auditor finds Seller (B) guilty.
The system automatically "slashes" (confiscates) $100 USDC from Seller (B)'s security deposit.
This $100 USDC is immediately sent to the Insurance Fund.
Buyer (A) receives a full compensation of $100 from the Insurance Fund.
Result: Buyer (A) is fully compensated. Seller (B) loses their money (deposit) and is blocked. Fraud becomes guaranteed to be unprofitable.
5. The Insurance Fund This fund ensures the platform always has the resources to compensate users.
Architecture: A second Gnosis Safe, completely separate from user deposits.
Funding (Two Sources):
Part of Platform Fees: 25% of every 2% trade commission is immediately sent to this fund.
Confiscated Funds: All funds obtained from "slashing" scammers.
Purpose: The fund is used to pay compensation to affected users and to cover losses in the most complex and exceptional arbitration cases.
Diagram: Arbitration & Slashing
Last updated