Risk Scoring

How the screener computes token safety ratings

Risk Scoring

The Base Screener actively evaluates smart contracts against multiple safety parameters to shield trading agents from malicious deployments.

Score range

Tokens are scored from 0 (safe) to 100 (extremely risky). A higher score indicates more danger.

Evaluation criteria

Check	Risk points	How it works
Honeypot	+50	Simulates buy/sell constraints through the Honeypot.is engine to detect if tokens can actually be sold after buying
Mintable	+20	Parses contract bytecode for `mint(address,uint256)` function selector (`0x40c10f19`), indicating the owner can inflate supply
Unverified	+15	Checks Basescan API for verified source code — unverified contracts score higher risk
Liquidity unlocked	+10	Checks if LP tokens are locked in a locker contract (Unicrypt, Team.finance)
Concentrated holders	up to +20	Calculates top 10 wallet concentration: >80% = +20, >50% = +10

Risk score examples

Token profile	Score	Meaning
Verified, liquid, distributed	0–10	Low risk — safe for agents
Unverified, moderate concentration	25–35	Medium risk — proceed with caution
Honeypot detected	50+	High risk — avoid
Honeypot + mintable + concentrated	90+	Critical risk — do not trade

Update frequency

Risk scores are computed by the risk-worker cron job running every 5 minutes. New tokens without a risk score are prioritized — up to 20 unscored tokens are processed per cycle.

Using risk data in queries

graphql
{
  newTokens(maxAgeSeconds: 3600, limit: 20) {
    tokens {
      symbol
      risk {
        score
        isHoneypot
        isMintable
        isVerified
        liquidityLocked
        top10HolderPct
      }
    }
  }
}

Agent-side filtering

typescript
const safe = tokens.filter(t =>
  t.risk.score < 40 &&
  !t.risk.isHoneypot &&
  !t.risk.isMintable
)