EDGES ARE
NOT ENOUGH.

RugBurn originated in Port Harcourt, Nigeria as a Solana risk scanner. The product changed when its own casework showed that an accurate wallet read can still miss the structure formed by many wallets acting together.

The US Reserve investigation exposed the first architectural failure clearly. RugBurn already had pre-outcome observations in the broader drain cohort, and its funding resolver was returning the transaction identity needed to distinguish wallets funded together. Downstream callers discarded that identity. The evidence existed; the relationship that made it useful did not survive the product model.

KYLIE-2026-001 pushed the problem further. An 86-address funding structure and twenty parallel buyer-side structures were visible, but the first explanation — shared infrastructure serving separate customers — was wrong. Testing the endpoints against outcomes changed the interpretation to one upstream funding operation spanning the deployer side and twenty buyer-side structures.

Seeing the graph is not enough. The interpretation has to survive the evidence.

That is the product direction now: on-chain behavioral intelligence that reconstructs funding structure and related-wallet context while keeping observed facts, derived relationships, and hypotheses separate. No identity claim is stronger than the evidence behind it.