Initialising agent
Home The problem Live sample Method Signals Coverage $HYPERSHIP Our coin
What it checks

Twenty-six signals, in four families.

Grouped the way losses actually happen: the contract traps you, the token lies about what it is, the market cannot absorb your exit, or the counterparty behind it has done this before.

Contract integrity

8 signals
Mint authority renouncedWhether any key can still inflate supply after you are holding.
No clawback or burn-fromOwner-gated burns that take balances straight out of a holder’s wallet.
Pause & freeze authorityA single key able to halt transfers and strand you mid-position.
Implementation immutableReads EIP-1967 slots — clean code means nothing if it can be swapped tomorrow.
Proxy admin distributionWhether upgrade power sits behind a multisig and a timelock, or one hot wallet.
Custom transfer hooksLogic on the transfer path that can claw back, burn, or redirect after purchase.
BlocklistsAn allow/deny list that can quietly stop you, specifically, from selling.
Honeypot sell simulationSimulated buy → sell against forked state. If the exit reverts, it is named outright.

Token authenticity

6 signals
Issuer registry matchDoes the deployer belong to the authorised issuer set for this asset, or is it unrelated?
Symbol & name collisionContracts imitating real Stock Token tickers, look-alike unicode, and near-miss names.
Metadata consistencyName, symbol and decimals agreeing with each other and with the issuer’s record.
Multiplier sanityChecks uiMultiplier() handling so a double-applied ERC-8056 multiplier cannot misprice what you see.
Oracle freshnessWhen the feed last updated, and how far it has drifted from spot.
Oracle pause stateWhether oraclePaused is set — a flag that is advisory, not enforced on-chain.

Market structure

6 signals
Liquidity depthReal Uniswap depth measured against your intended size, not against one token.
LP lock or burnWhether liquidity can be removed, by whom, and when the lock expires.
Holder concentrationMaps top wallets and clusters ones that move together — a coordinated exit is one wallet in effect.
Same-block insider bundlesLaunch-block accumulation that quietly positions you as the exit liquidity.
Exit slippage modelWhat leaving actually costs at the size you were going to take.
Wash-trading detectionWhether the volume drawing you in comes from distinct wallets or circular flow.

Counterparty history

6 signals
Deployer trailEvery prior deployment by the same wallet, and what became of each one.
Funding pathWhere the deployer’s gas came from: a long-lived wallet, a fresh bridge, or a mixer.
Drainer bytecode matchThe contract checked against known drainer sets and unbounded-approval solicitation.
Contract ageHow long this code has existed, weighted against everything else on the list.
Flagged wallet proximityEdges — direct or one hop — to addresses already tied to losses.
Bridge & withdrawal exposureL2 exit assumptions: the fraud-proof window and the single-sequencer dependency.

Verify an address before you sign it.

26 signals, one score, the reasoning attached. Read-only — the agent never signs and never asks for an approval.

Run a verification