Lintmus runs a multi-source safety analysis on any EVM or Solana token and returns a 0–100 score with a full breakdown.
Identifies tokens you can buy but not sell, using GoPlus Security's on-chain simulation engine.
Surfaces hidden buy and sell taxes. Anything above 10% is flagged — above 30% is penalised heavily.
Detects mint functions, hidden owners, the ability to reclaim ownership, and pausable transfers.
Checks current liquidity depth, 24h volume, LP lock percentage, and how long the pair has existed.
Confirms whether source code is publicly verified. Unverified contracts score lower by default.
Looks up the wallet that deployed the contract. Known rug deployers are penalised per incident.
Starts at 100. Penalties are applied for red flags, bonuses for trust signals. Capped at 0–100. Missing data always scores neutral — never inflated.
| Signal | Impact |
|---|---|
| Honeypot confirmed | −80 |
| Sell tax > 30% | −40 |
| Sell tax 15–30% | −25 |
| Hidden owner | −20 |
| Can reclaim ownership | −15 |
| Top 10 holders > 80% | −15 |
| Known rug deployer | −15 per rug |
| Transfer pausable | −10 |
| Mintable supply | −10 |
| Creator holds > 20% | −10 |
| Liquidity > $100K | +10 |
| LP locked > 80% | +5 |
| Contract verified | +5 |
| Token age > 1 month | +5 |
| Token age > 1 week | +3 |
Works with any agent that supports the SKILL.md format — Claude Code, OpenClaw, Cursor.
install the lintmus skill from https://github.com/lintmus/lintmusThen ask your agent:
analyze token 0x532f27101965dd16442e59d40670faf5ebb142e4 on base
is this safe to buy? 0xADDRESS on ethereum
run lintmus before I ape into thisPayment of 0.03 USDC is handled automatically via the x402 protocol. Reports are cached for 60 minutes.
POST https://x402.bankr.bot/0xD3aDb4D4B787eE631A1B2618464e21B229873075/analyze
{
"address": "0x532f27101965dd16442e59d40670faf5ebb142e4",
"chain": "base"
}Every token analyzed by Lintmus is logged. Outcomes — rug, alive, or abandoned — are recorded as they happen and published openly so anyone can verify the scoring methodology and track historical accuracy.
This dataset grows with every report and forms the foundation for future ML-based scoring improvements. The data is yours to use.