Open App

Rug Pull Prevention

Key Takeaways
- 90.1% detection accuracy before pool collapse - combining deployer wallet behavioral analysis with smart contract code inspection
- Two-pipeline approach: deployer wallet fraud scoring + static contract code analysis run in parallel
- Free for individuals via the Rug Pull Detector at chainaware.ai; available via predictive_rug_pull MCP tool for protocols
- Supports ETH, BNB, BASE, and HAQQ

The Problem: You're Investing Before the Threat Is Visible

A rug pull is simple in its mechanics and devastating in its effect. A team launches a token, attracts liquidity from investors, then withdraws all funds from the liquidity pool and disappears - leaving investors holding worthless tokens and the protocol unable to recover.

The scale is staggering. ChainAware's analysis of PancakeSwap V2 documented 103,695 rug pull events across 20 consecutive weeks in 2026, extracting $569M from retail holders - an average of $28.5M per week on a single DEX. BNB Chain's sub-$5 deployment cost makes this industrial-scale operation economically trivial to sustain.

For individual investors, the losses are real and often life-altering. For DeFi protocols that host these pools, the reputational damage is severe - and they are increasingly being held responsible for not screening what they list.

Why Standard Tools Fail

Most token security tools take one of two approaches:

1. Smart contract audits. Audit the code to check for malicious functions - hidden mint permissions, ownership backdoors, transfer restrictions. This sounds reasonable, but experienced rug pullers have adapted. They deploy contracts that pass audits cleanly, then execute the rug through LP withdrawal or coordinated selling - mechanisms that are perfectly legal in code but catastrophically fraudulent in intent.

2. Social signals. Telegram activity, Twitter followers, team doxxing. These signals are trivially manufactured. Fake community members cost pennies. KYC documents can be falsified. Social proof is the easiest thing to fake in crypto.

Neither approach looks at what actually matters: who is behind this token, what is their history of on-chain behaviour, and does the contract contain explicit sell-blocking mechanics?

How ChainAware V3 Solves It

Rug Pull Detector V3 runs two parallel analysis pipelines simultaneously, completing the full assessment in under 2 seconds:

Pipeline 1: Deployer Wallet Behavioural Analysis

  • Has this wallet deployed tokens before? If so, what happened to them?
  • Are there patterns of rapid LP withdrawal, coordinated selling, or token abandonment in the deployer's history?
  • Is this a fresh wallet with no history - itself a significant red flag?
  • Are LP wallets connected by transaction history in ways that suggest coordination?
  • Do the deployer or LP wallets connect (directly or through intermediaries) to wallets previously associated with known rug pulls?
  • Do funding sources trace back to mixers, scam clusters, or suspicious origins?

Pipeline 2: Smart Contract Code Inspection

V3 adds direct analysis of the contract code alongside the wallet layer - catching cases where the deployer has a clean history but the contract contains explicit fraud mechanics:

  • Hidden transfer restrictions that block sells
  • Owner-privileged mint functions with no cap
  • Unrenounced ownership - the team retains full control post-launch
  • Asymmetric pause logic that blocks sells but not buys
  • Fee manipulation functions settable to 100%
  • Upgradeable proxies with EOA admin

Both pipelines produce independent risk scores. An ensemble model integrates them into a single rug pull probability (0-100), delivered with a human-readable breakdown of the key factors driving the score.

90.1% detection accuracy on new pools within the first hour of launch - before any social signals exist, before influencer promotion, before the community forms. Up from 68% in V2, a 32.5% relative improvement.

Key Metrics

90.1%
V3 detection accuracy
$569M
Extracted on PancakeSwap V2 in 20 weeks
103,695
Rug pull events documented
<2s
Full analysis time

For comparison: relying on social signals (Telegram activity, Twitter mentions) typically catches 10-15% of rug pulls - mostly after the price has already collapsed and the damage is done.

Products for Rug Pull Prevention

For Individual Users: Rug Pull Detector

Available free at chainaware.ai:

  1. Paste a token contract address or liquidity pool address
  2. ChainAware runs both analysis pipelines against 20M+ behavioural profiles
  3. Receive a risk score, key risk factors, and a recommendation - in under 2 seconds

No wallet connection required. No signup. Works on BNB Smart Chain, Ethereum, Polygon, Base, and more.

For Protocols and AI Agents: Prediction MCP

The open-source Prediction MCP brings rug pull screening directly into AI agent workflows:

Tool: predictive_rug_pull
Input: { "pool_address": "0x...", "chain": "bsc" }
Output: { "risk_score": 87, "risk_level": "HIGH", "key_factors": [...] }

Integrate rug pull scoring into:
- Automated listing screening for DEXs
- AI trading agents that execute token purchases
- Telegram bots that screen links shared in your community
- DeFi protocol onboarding flows

How to Use It

Individual User Flow

  1. Find a new token you're considering investing in
  2. Copy the contract address or liquidity pool address
  3. Go to chainaware.ai and paste it into the Rug Pull Detector
  4. Review the risk score and key factors
  5. Make an informed decision - and share the result with your community

Protocol / API Flow

  1. Register for API access at chainaware.ai/pricing
  2. Integrate the rug pull endpoint: GET /v1/rugpull/{pool_address}?chain={chain}
  3. Gate listing decisions, display risk warnings, or automate rejection of high-risk pools
  4. Or use the Prediction MCP for AI-agent-native integration

API Documentation →

Frequently Asked Questions

How does ChainAware detect rug pulls before they happen?
Rug Pull Detector V3 runs two parallel pipelines: deployer wallet behavioral analysis (cross-referenced against 20M+ wallet profiles to surface prior rug history, mixer exposure, and funding source red flags) and smart contract code inspection (scanning for hidden sell restrictions, uncapped mint functions, unrenounced ownership, and fee manipulation). Both pipelines complete in under 2 seconds, before any social signals exist.

What is the detection accuracy of V3?
90.1% on new pools within the first hour of launch - up from 68% in V2. This is measured on new pools where the deployer wallet's prior history is the primary signal, before any community or trading data exists. For comparison, relying on social signals typically catches 10-15% of rug pulls, mostly after the collapse.

Which blockchains does rug pull detection support?
Rug Pull Detector V3 supports BNB Smart Chain, Ethereum, Base, and HAQQ. The free individual tool and the predictive_rug_pull MCP tool both cover these chains.

Can a protocol automate rug pull screening for token listings?
Yes. The predictive_rug_pull MCP tool integrates directly into AI agent workflows - DEX listing pipelines, Telegram bots, and automated screening systems can call it programmatically. The REST API endpoint GET /v1/rugpull/{pool_address}?chain={chain} is also available for non-agent stacks.

Further Reading

Try Rug Pull Detector Free

No signup. No wallet connection. Results in under 2 seconds.

Check a Token Now at chainaware.ai →

Last updated: