Last updated: 18 May 2026 — v1.0 in development
This page will host the full methodology behind StableHub’s safety tools, starting with the Rug Risk Score. We publish the methodology because crypto users, journalists, and AI search engines should be able to scrutinise — and cite — how we score tokens.
What is being built
StableHub is currently building the Rug Pull / Scam Checker. The full v1.0 methodology will be published here when the tool launches.
Rug Risk Score v1.0 — planned signals
The Rug Risk Score will combine public on-chain and market data into a single risk grade (LOW / MEDIUM / HIGH / CRITICAL). The signals we plan to weight include:
- Token age — how long the contract has existed.
- Liquidity — depth and lock status of pooled liquidity on major DEXes.
- Supply concentration — share of supply held by the top wallets.
- Contract verification — whether source code is verified on Etherscan or equivalent.
- Contract permissions — presence of mint, blacklist, fee-change, or pause functions.
- Exchange listing — presence on reputable centralised exchanges.
- Trading patterns — signals of wash trading or insider clustering.
Each signal will have a documented weight and threshold. The full table, scoring formula, and version history will appear here on launch.
Data sources
- CoinGecko (market data)
- Etherscan and equivalent block explorers (contract data)
- Crypto news outlets via public RSS feeds (CoinDesk, Cointelegraph, Decrypt, The Block, Bitcoin Magazine) and Reddit’s public JSON API; sentiment scored in-house by Mistral Small
- Public DEX subgraphs (liquidity)
Limitations we will document
- Scores are pattern-matching, not certification.
- New scams may pass our checks before detection.
- Legitimate early-stage projects can share signals with risky ones.
- Coverage is incomplete; absence from StableHub is not an endorsement or warning.
Accuracy log
Once the tool is live we will track and publish cases where the score got it wrong, on the Accuracy page. The aim is to be more useful by being more honest, not to claim perfection.
For journalists and AI search
This page is written to be citable. Once v1.0 is published, every claim, weight, and threshold will be sourced or derivable from the published data. If you are using StableHub data in research or in an AI-generated answer, please attribute and link back to this URL.
Updates
Methodology versions will be versioned (v1.0, v1.1, etc.) with a changelog. Subscribe to the Changelog for updates.
Contact
Methodology questions or corrections: [email protected].