Published by TOSScoin Research

Sybil Defense for Crypto: Human Passport & Verification Tools

Key Findings

Source: https://docs.passport.xyz/building-with-passport/models/available-models
Findings:
- Supports multiple models (aggregate, ethereum_activity, nft, arbitrum, base, optimism, polygon, zksync).
- Scores are -1 (insufficient data) or 0–100.
- Published threshold tables show human-qualification vs Sybil-penetration tradeoffs, with aggregate model recommendation around 50+.
Why it matters: Lets campaigns choose strict vs growth-friendly gating with quantifiable false-positive / false-negative tradeoffs.

Source: https://docs.passport.xyz/building-with-passport/models/api-reference
Findings:
- Endpoint: GET https://api.passport.xyz/v2/models/score/{address}?model=...
- Requires API key and returns model score payloads.
- Documented limits: Tier1 15 req / 15 min; Tier2 350 / 15 min; Tier3 2000 / 15 min.
Why it matters: Supports real-time wallet scoring pipelines before allowing quests/referral rewards.

Source: https://docs.holonym.id/for-developers/custom-sybil-resistance
Findings:
- Projects can define custom actionID values to scope uniqueness proofs per campaign/action.
- Docs explicitly discuss bribery risk differences between default/global uniqueness and per-action uniqueness.
Why it matters: Useful for campaign-specific anti-Sybil control, but must be designed carefully to avoid exploitable low-stakes bribery vectors.

Use Human Passport model score checks for first-pass wallet filtering.
Apply stricter thresholds on high-value referral rewards than on low-value awareness quests.
Use Holonym action IDs only when campaign context truly requires scoped uniqueness.
Keep a manual review lane for borderline wallets near threshold cutoffs.