Automated pre-audit for DeFi teams. Slither and Echidna for breadth, Claude for the reasoning detectors can't do, and a runnable Foundry PoC behind every HIGH. Not warnings. Evidence.
Four stages, roughly six hours for a 1,000-line protocol.
We compile with the exact solc version, resolve dependencies, then run Slither and Aderyn's full detector suite over the compiled AST. A typical 1,000-line protocol yields 150–250 raw findings.
Echidna and Medusa try to falsify derived protocol invariants over millions of calls and transaction sequences.
Triage: discard detector noise against real code semantics — typical discard rate 80–95%. Semantic hunt: read the source for what pattern matchers cannot express: missing authorization on admin setters, broken accounting, wrong rounding direction. A permissionless setter produces no detector output; it produces a finding here.
Each HIGH and CRITICAL ships with a Foundry test that reproduces it. We run it — no passing test, no finding. Quick tier and above.
No — it's a pre-audit layer. It catches most exploitable issues before you spend $50K on a firm and tells you what to fix first. For serious TVL, still get the human audit; use us so it isn't your only defense.
You get a report saying so, with the full tool output. That's not a guarantee of safety — it means we couldn't break it with the methods we ran, and we state exactly what those were.
Two passes, different jobs. Triage discards false positives against your compiled code. The semantic pass reads for bug classes detectors can't express. Both carry confidence scores and a "not verified" field. On paid tiers every HIGH must also survive a PoC.
Run the free scan first. If it surfaces something you didn't know about, you'll know what the paid tier is worth.
Scan a contract — free →