AIRGAP@AirgapAI
AIRGAP Threat Engine | Wave 1 is live
we caught a serial deployer minting identical $SpaceX clones on @clanker_world, each wash-traded to ~100x the volume its liquidity could support. one of them floated to the top of our own Profitable Signals board.
that’s the moment we decided the feed itself needs to know the difference. so we built it.
every @base launch AIRGAP tracks now runs through an onchain threat layer scoring it for manipulation in real time. the methodology should be legible, so here it is.
the core is a pure scoring function: same inputs, same verdict, every time. no model in the hot path, nothing that drifts. market structure and deployer history go in, a verdict comes out with specific flags and evidence attached.
ratios, never absolutes. high volume isn’t suspicious, manufactured volume is.
› WASH measures 24h volume against the pool’s actual liquidity. a pool holding a few hundred thousand cannot honestly turn over tens of millions a day. when the ratio detaches from anything physically supportable, we say so.
› SERIAL fingerprints deployers across the whole catalog and surfaces the spam-factory signature: identical launches, minted over and over.
calibrated against reality. trip points tuned on confirmed wash tokens on one side, genuine 50x and 100x winners on the other. real momentum stays clean, wash mirages light up. ambiguous signals get amber, we never cry red on a maybe.
where you see it: per-signal badges in the launches feed and token pages, each citing the actual numbers (“24h volume is 134x the pool’s liquidity”). wash and spam tokens are pulled from Profitable Signals entirely, because a manipulated chart is not a win.
the $SpaceX clones are flagged high and removed. the deployer is fingerprinted. every future launch from that wallet lights up on arrival.
Wave 2 goes fully onchain: same-block coordinated buy detection and holder-cluster concentration. heavier forensics, same precision bar.
know what you’re trading against.
$AIRGAP
