ARCADIAN
RISK ORACLE
RISK ORACLE
VERIFIABLEON-CHAINDEFI SIGNAL

Scoring the very best — and worst — DeFi protocols. A Telegraph Miner that wraps live DefiLlama & CoinGecko data into a single verifiable 0–100 risk score any agent can act on.

06
COMPONENTS
04
DATA SOURCES
589
HACKS INDEXED
0–100
SCORE RANGE
LIVE RISK QUERY

No protocol slug needed — type any asset and Arcadian finds every protocol offering it on Ethereum, then ranks them by risk.

SCORING METHODOLOGY · 100 PTS
010–20
APY CREDIBILITY

Sigma (30d std-dev), outlier flag, reward ratio. Source: DefiLlama Pools.

020–20
LIQUIDITY DEPTH

TVL tiers + 30d TVL drawdown. Source: DefiLlama History.

030–20
EXPLOIT HISTORY

589 real hacks matched by protocol, recency-weighted. Source: DefiLlama Hacks.

040–20
PROTOCOL MATURITY

Age since listing, audit count, fork status. Source: DefiLlama Protocols.

050–10
CONCENTRATION / IL

LP vs single-asset, stablecoin vs volatile IL exposure.

060–10
REWARD TOKEN

Market-cap rank + 30d price volatility. Source: CoinGecko.

All data sourced live from DefiLlama (pools · hacks · protocols) & CoinGecko. Lower score = safer. Telegraph on-chain integers: risk_score · apy_bps · tvl_k · confidence.

API REFERENCE
POST /api/telegraph/risk
Content-Type: application/json
// DISCOVERY MODE — no protocol slug needed
{
  "chain": "Base",
  "asset": "USDC"
}
// RESPONSE
mode ranked
safest protocol slug with the lowest score
ranked[] every protocol, safest first
risk_score int 0–100 // lower = safer
verdict LOW | MEDIUM | HIGH | VERY_HIGH
apy_bps · tvl_k · confidence on-chain integers
components 6-component breakdown
explanation AI narrative (Qwen)
// DIRECT MODE — pin one protocol
{ "protocol": "aave-v3", "chain": "Base" }
GET /api/telegraph/universe?chain=Base
// live chain + asset lists that back the pickers
DATA SOURCES:
DEFILLAMA
COINGECKO
TELEGRAPH
BASE