What Is a Rug Pull in Crypto? Anatomy of the Scam and How Scanners Detect It
The chart looked perfect right up until the sell button stopped working.
A rug pull is when a token's creators drain value from buyers — by removing liquidity, minting supply, blocking sells or using malicious pool hooks. The main rug types, red flags, and how automated safety scanning detects them.
A rug pull is a scam in which a token's creators extract value from buyers after launch, leaving holders with unsellable or worthless tokens. On fast launch chains like Robinhood Chain rugs are the dominant loss source for retail traders — and the reason serious trading bots build safety scanners instead of trusting charts.
The main rug types
- Liquidity removal: creators pull the LP out of the pool, collapsing price instantly. Detection: verify on-chain that LP tokens are time-locked in a recognized locker or burned — by reading the contracts, not by trusting events.
- Honeypot (sell-blocking): buys work, sells revert for everyone except insiders. On Uniswap V4 chains this is increasingly implemented in the pool hook rather than the token — which is exactly why hook scanning exists.
- Mint dump: hidden mint authority inflates supply and the creator sells into buyers. Detection: bytecode analysis of mint/freeze authorities, verified against live contract state.
- Insider/snipe concentration: deployer-linked wallets buy in the creation block and dump on the community. Detection: creation-transaction anchoring and funding-linkage analysis — see Loxley's snipe detection docs.
- Soft rug: no contract trick at all — the team simply abandons the project and lets attention (and price) decay to zero.
Red flags before you buy
- Liquidity not locked, or locked in an unknown locker contract.
- Heavy same-block insider buying at creation.
- Sell simulation fails or shows punitive output deltas.
- V4 pool hook with transfer-blocking or owner-mutable fee parameters.
- Deployer wallet funded by addresses linked to previous rugs.
How automated scanning helps
No human can run these checks in the seconds after a launch. Loxley AI's pipeline runs them automatically at detection speed: contract authorities, lock proof via direct on-chain reads, snipe concentration scoring, hook inspection and transaction simulation — with the design rule that degraded results are never cached (a failed check reports as degraded, never as safe). Unattended auto-buy hard-blocks on bad verdicts; manual buyers get explicit warnings and decide for themselves; sells are never blocked.
The honest caveat
Scanners eliminate the known-bad majority and raise scammer costs. They cannot catch novel patterns, off-chain coordination or soft rugs. Position sizing and discipline remain yours. Not financial advice.
Security guide → All technical docs
Built for operators, not slippage goblins. Talk to us →