Deployer fingerprint: how the engine recognizes serial offenders before they rug

The main predictive power of on-chain data is biography. The token is new, but the hands that launched it are almost never new. Our engine builds a registry of handwriting on this basis.

What is handwriting

Stable patterns of a deployer across launches: the share they keep for themselves; how they fund the launch (gas source); whether they hold or dump, and at what percentage of growth; contract parameters (the same taxes/limits); timing - how quickly they abandon a dead token and launch the next one. People are lazy: a scammer who has worked out a scheme repeats it with robotic precision.

How wallets are linked

A "new" deployer is often old: funding from the same addresses, reuse of pass-through wallets, identical contract bytecode, signature amounts and intervals. The transfer graph stitches identities together better than scammers can hide.

What this gives the verdict

• A deployer with a history of rugs - the verdict drops automatically, no matter how "clean" the new token looks.
• A serial deployer with no explicit rugs, but with a graveyard of abandoned tokens - a pipeline flag: the fate of the new one is statistically predictable.
• A first launch with clean behavior - neutral: there's no biography, other metrics decide.

The section with the history of specific deployers is connected to every verdict; repeat offenders are visible in the section.