Rug Pull Detector¶

$569.4M was extracted through rug pulls on PancakeSwap V2 alone in the first 20 weeks of 2026 - across 103,695 separate events. That's not a rounding error in crypto losses. It's an industrial-scale operation running continuously against retail buyers.

Smart contract audits don't stop this. Experienced rug pullers pass code reviews easily and deploy new contracts daily at under $5 in gas. ChainAware's Rug Pull Detector V3 catches what audits miss - by combining deployer wallet behavioural history with smart contract code inspection in a single analysis.

Check a Token for Rug Pull Risk →

The Scale of the Problem¶

Analysis of PancakeSwap V2 data across 20 weeks (2026 W1-W20):

$569M
Extracted in 20 weeks

103,695
Rug pull events

$53.4M
Peak single week

90.1%
V3 detection accuracy

The $1.38B in liquidity added by creators was followed by $1.95B removed - the difference representing extracted retail funds plus the original liquidity. BNB Chain's low gas costs make it the primary venue: a rug pull operation can deploy, pump, and exit a token for under $5.

What Is a Rug Pull?¶

A rug pull is an exit scam where project developers withdraw all liquidity or drain contract funds, leaving investors with worthless tokens. Two main types:

Liquidity rugs - The team removes all liquidity at once, crashing the price to zero instantly
Backdoor rugs - Hidden contract functions allow developers to mint unlimited tokens or drain funds at will

Most rug pulls follow a predictable four-phase playbook:

Narrative creation - A compelling story positioning the token as a solution to a real problem
Hype building - Paid influencers, coordinated social media, manufactured urgency
Price pump - Coordinated buying and retail FOMO drive artificial price appreciation
The exit - Liquidity removed or contract abandoned, holders left with nothing

How V3 Works - Two Parallel Pipelines¶

V3 combines two independent analysis layers that run simultaneously and feed into an ensemble model:

Pipeline 1: Creator Wallet Behavioural Analysis

Every deployed contract has a creator wallet. ChainAware scores that address against 20M+ wallet profiles:

Deployment history - has this wallet deployed tokens that rugged before?
Funding provenance - was the wallet funded from a mixer, a known scam cluster, or a fresh CEX withdrawal?
Wallet age and transaction depth - brand-new wallets with minimal history carry higher base risk
LP wallet analysis - fresh LP addresses with high fraud scores signal "hot" liquidity ready for removal

ChainAware traces through factory contracts and proxy deployments to identify the human actor behind obfuscated deployment chains.

Pipeline 2: Smart Contract Code Inspection

V3 adds direct contract analysis alongside the wallet layer:

Hidden transfer restrictions and honeypot patterns
Owner-privileged mint capabilities with no cap
Unrenounced ownership - the team retains full control
Liquidity lock status and conditions
Fee manipulation functions that can be raised to 100%

Both pipelines produce independent risk scores. The ensemble model integrates them into a single rug pull probability, completing the full analysis in under 2 seconds.

What the Red Flags Look Like¶

Signal	What It Means
New creator wallet	No history to analyse; no reputation at stake
Prior rug deployments on creator	Direct behavioural evidence - past pattern predicts future
Fresh LP addresses with high fraud score	Liquidity is mobile and positioned for quick removal
Obfuscated deployment chain	Creator is hiding their identity behind proxies
Concentrated LP positions	Single actor controls the exit
Unrenounced ownership + mint function	Owner can inflate supply and sell into the pool at any time
Transfer restrictions in code	Sells may already be blocked at launch

Accuracy and Limitations¶

V3 achieves 90.1% prediction accuracy on PancakeSwap V2 data - a 32.5% relative improvement over V2's 68% accuracy. This means it catches 9 in 10 rug pulls before they happen, often within the first hour of pool launch when no social signals exist yet.

The remaining ~9.9% of missed rugs are concentrated among sophisticated operators who maintain clean deployer wallet histories and use heavily obfuscated contract code. These are harder to catch - but also rarer and more expensive to engineer.

Backtesting methodology and verification data →

How to Use It¶

Find a token contract address or pool address you want to check
Go to chainaware.ai
Paste the contract address into the checker
ChainAware automatically detects whether it's a wallet or contract and runs the appropriate analysis

You can also check via:
- Telegram Mini App: /rugpullcheck command - Join here

Possible Outputs¶

ChainAware auto-detects whether the address is a wallet or contract and routes accordingly:

Wallet address - Automatically runs Predictive Fraud Detection instead
Contract address - Runs the Rug Pull Detector
Forensic database match - Automatically scored at 100% fraud
Cached result - Returns the last calculated result; manual recalculation available
Shareable result - You can share results with partners and communities

Reading the score: A rug pull probability above 50% is a strong warning sign - the pool carries significant risk. Below 50% is safer, but always combine with your own due diligence.

API Access¶

The Rug Pull Detector is accessible via the Fraud Detection API at swagger.chainaware.ai. API key required - get one from your profile dashboard. Subscriptions at chainaware.ai/pricing.

Supported Networks¶

Ethereum, BNB Smart Chain, Polygon, Base, Solana, TON, and Tron. The detector covers 95%+ of new PancakeSwap V2 pools and significant coverage across Uniswap and other major DEXes.

Why Not Just Read the Contract?¶

Smart contract code can be copied from legitimate projects, obfuscated, or verified on-chain while still containing backdoors. Code review alone misses the most important signal: who deployed it, and what have they done before?

A wallet with prior rug pull history, connected to known scam clusters, with a near-zero trust score is a far stronger predictor of risk than any line of Solidity. V3 uses both - contract code and deployer behaviour - because either alone misses cases the other catches.