// public sandbox · real alchemy ingestion · model v0.5.0-gov-expanded · status
SYBILSHIELD

About SybilShield

Open-methodology Sybil detection for token distributions.

Why we built this

Every airdrop in 2023-2025 was extracted by farmers running thousands of wallets. Projects either overspend on detection ($150K consulting) or underspend and lose 20-40% of their distribution. The tooling gap is real: Trusta is Ethereum-focused and black-box, Nansen is too expensive and not airdrop-specific, custom Dune analyses take weeks.

We're building the "credit-scoring layer for token distributions" — an evidence-based, auditable, open-methodology service that any project can plug into their TGE flow.

A neutral second opinion

When a project's chosen filter — anyone's, not just ours — flags a real contributor and the community pushes back, there's rarely a neutral party either side can point to. The filter vendor made the call; disputing it usually means disputing the same vendor.

SybilShield is structurally suited to be that second opinion: MIT-licensed detection code anyone can read line by line, an honest-holdout accuracy record published with its own caveats rather than a marketing number, and no commercial stake in any one project's outcome — the whole thing is free. Not a replacement for whichever filter a team already runs; an independent read anyone can check, including the people it flags.

Team Beta

Solo founder + open-source contributors. Funded contributors named here as they join.

Founder · Eng

Background in on-chain analytics and ML. DM on Telegram or open a GitHub issue to chat.

Open seat · Data

Coming after first grant. Curating labelled corpus + adversarial red-team.

How we sustain

Legal posture

An open, grant-funded public-good project — free to use, MIT-licensed, no monetization. Anyone using SybilShield scores in a public filter list is expected to provide an appeal flow so scored parties can dispute a result. That appeal infrastructure (public endpoint, 48h SLA, immutable audit log) is usable standalone, alongside any filter — not only SybilShield's own scoring.