The check that runs before value moves

1. 01Read the source before you trust it

   Scan

   Inspect verified Base contracts and source for security patterns that matter before launch. Rule-based detectors, rug-surface mapping and famous-hack patterns all run without a wallet or an account.

   Run a Base scan→
2. 02The problems surface before money moves

   Guard

   Give an app or agent a machine-readable preflight signal before it interacts. One HTTP call returns a bounded decision you can gate on. These are the shapes it looks for:

   • —Spend permissions with overly broad allowance or no expiry
     Signals unbounded approvals before an app grants them.
   • —EIP-7702 delegation and replay-risk shapes
     Reviews delegated-account authority and replay exposure.
   • —Permit2 amount and expiry hazards
     Flags amount and expiry mismatches for review.
   • —No-min-out swaps and weak oracle or TWAP use
     Signals missing slippage bounds and single-source pricing.

   Open Guard→
3. 03A decision should leave a trail

   Receipts

   A paid scan can add an EAS attestation on Base, so the decision carries an evidence trail anyone can verify independently rather than a claim you have to trust.

   • —Shows the finding category and severity behind a signal.
   • —Distinguishes static-only scans from EAS-attested receipts.
   • —Links to the public receipt when one exists.
   • —Makes a new scan obvious when no public preflight exists yet.

   View all receipts→
4. 04Put a check before the call

   Developers

   One HTTP call returns a machine-readable preflight signal for any Base contract. Wire it into an agent, a hook, or CI.

   curl -sS <APP_URL>/api/guard/base/<ADDRESS>
   
   # → { decision, score, grade, proofStatus,
   #     easAttested, riskSurface, reasons }
   # decision ∈ low_risk_signal | review
   #            | block_for_review | needs_scan

   Read developer docs→