How AI Screening Protects Buyers on Presello

The Screening Pipeline

Every project submitted to Presello goes through a multi-stage AI screening pipeline before it can be listed on the platform. This process is automated and runs on every new project submission, with weekly re-audits for all active projects.

The pipeline combines data from multiple independent security services to create a comprehensive risk profile. No single data source is relied upon exclusively. Instead, the system aggregates signals from contract analysis, liquidity assessment, holder distribution, and community sentiment to produce a weighted risk score.

The screening process begins the moment an issuer submits their project. The system automatically fetches the smart contract code, analyzes it against known vulnerability patterns, checks for common scam indicators, and evaluates the token's distribution and liquidity profile. Results are typically available within minutes.

Token Sniffer Analysis

Token Sniffer is one of the primary data sources in the screening pipeline. It analyzes smart contract code for known scam patterns, including honeypot mechanisms, hidden mint functions, blacklist capabilities, and suspicious fee structures.

The Token Sniffer score is factored into the overall risk rating with a significant weight. A low Token Sniffer score (indicating potential issues) will negatively impact the project's risk rating on Presello, even if other signals look positive.

Token Sniffer also checks for contract similarity to known scam contracts. If a submitted project's smart contract is structurally similar to previously identified scams, this is flagged as a high-risk indicator. The system maintains an updated database of scam patterns, so new variations are caught as they emerge.

GoPlus Security

GoPlus provides an additional layer of security analysis focused on real-time threat detection. The GoPlus API checks for a wide range of potential issues including whether the contract can be paused or self-destructed, whether the owner can modify fees after deployment, whether there are hidden transfer restrictions, and whether the token has been flagged by other security services.

GoPlus also provides malicious address detection, which helps identify whether the project's deployer wallet or associated wallets have been involved in previous scams or fraudulent activity. This cross-referencing with known bad actors adds another dimension to the risk assessment.

The combination of Token Sniffer and GoPlus means that each project is analyzed from multiple angles. Token Sniffer focuses primarily on contract code analysis, while GoPlus adds real-time security monitoring and cross-chain threat intelligence.

On-Chain Analysis and Sentiment

Beyond automated security tools, the screening pipeline includes on-chain analysis via BSCScan and Etherscan APIs. This analysis examines holder distribution (looking for excessive concentration), transaction patterns (looking for wash activity), and liquidity pool health (looking for potential rug pull setup).

The final component is AI-powered sentiment analysis using Groq's Llama model. This analyzes publicly available information about the project including official announcements, community discussions, and development activity. The sentiment score helps identify projects that may have technical green flags but concerning community signals.

All of these signals are combined using a weighted scoring algorithm to produce the final risk rating. The weights are calibrated to prioritize hard security signals (contract vulnerabilities, scam patterns) over softer signals (community sentiment) while still incorporating the full picture.

Understanding Risk Ratings

The final output of the screening pipeline is a five-level risk rating displayed on every project page:

GREEN indicates the lowest risk level. The project has passed all automated checks with high scores, has a clean contract, adequate liquidity, healthy holder distribution, and positive sentiment signals. This does not mean the project is safe or guaranteed to succeed. It means the automated screening did not find concerning signals.

YELLOW indicates minor concerns. The project may have some concentration in holder distribution, limited liquidity, or mixed sentiment. Buyers should review the specific concerns noted in the audit report.

ORANGE indicates moderate concerns. Multiple screening signals have flagged potential issues. Buyers should exercise additional caution and thoroughly review the audit details before purchasing.

RED indicates significant concerns. The project has failed one or more critical screening checks. While not automatically delisted, RED-rated projects display prominent warnings.

BLACK indicates a failed screening or confirmed malicious activity. Projects rated BLACK are suspended from the marketplace and no new purchases are processed.

Important: risk ratings are automated assessments and not guarantees. A GREEN rating does not mean a project will succeed, and even thoroughly screened projects can fail. Always do your own research before making any purchase decisions.

Weekly Re-Audits

Screening is not a one-time event. Every active project on Presello is automatically re-audited on a weekly basis. This catches changes that may occur after the initial screening, such as contract ownership transfers, liquidity removals, holder concentration changes, or new security vulnerabilities discovered in the contract code.

If a project's risk rating changes during a re-audit, all affected users (sellers with active listings and the project issuer) are immediately notified. Rating downgrades are applied automatically. An admin can review and override certain rating changes, but a BLACK rating cannot be manually overridden to GREEN.

This continuous monitoring is one of the key differentiators between Presello and informal peer-to-peer sales. When you buy through a direct deal, there is no ongoing screening or monitoring. On Presello, the AI watchdog is always watching.