What to Expect During a Smart Contract Audit, From Kickoff to Final Report
Security Wiz team4 min read
If you've never been through a smart contract audit, it can feel like dropping your car at a garage: you hand it over and something comes back later. It shouldn't feel that way. Here's what happens, in order, and where you're involved.
1. Quote
You send us your code and a short description of what your product does. We return a fixed quote within 24 hours, based on the size and complexity of the code and not on an hourly meter. A single contract can take days. A full product is typically measured in weeks. You get a firm timeline with the quote.
2. Intake
Work starts when your code is frozen at one exact version and we have the essentials: the code and the ID of the version to review, a list of what's included and what isn't, code that builds on a fresh computer, your tests, and your launch scripts. We also need a plain-English description of how the product works, the rules you believe must never break, who holds which keys, and what you assume about the outside services you rely on. Our pre-audit checklist covers all of it.
We also agree ground rules for testing anything that's already live, and collect your written consent for how your code is handled.
3. Test plan and sign-off
Before testing, we work out what kind of product yours is (lending, an exchange, a vault, a bridge between blockchains, and so on) and build a test plan for it. The plan has a base checklist that always applies, plus sections specific to your kind of product. We also write a one-page threat model: the money or assets that could be stolen or frozen, the people involved, and what you're trusting each of them to do.
You review and sign off on the plan, the version of the code, and what's included. This is the agreement on what the audit will and won't cover, and it becomes the table of contents for your report.
4. Testing
This is the part you see least of, and it has several layers:
- Merging the code into one continuous document, so a reviewer can follow each action from start to finish.
- An automated scan, which produces leads. A scanner warning is never a finding on its own.
- A manual review by a senior researcher, who confirms the leads and looks for what no tool can find: rules with holes in them, risky assumptions, and roles with too much power.
- Fork testing, on the Fork tier, where we carry out working attacks against a private copy of the live blockchain. See Core vs Fork.
Every finding is confirmed before it reaches you, either with a test or by following the exact route through the code. A confirmed Critical doesn't wait for the final report. We send it as soon as it's confirmed, so you can start on the fix.
5. Findings
You receive each confirmed finding with its severity, a description of how an attacker would use it, and a concrete fix. Severity for problems in the contract code is scored with OVSS, our open scoring system, so you can see exactly how we reached it. Expect to answer some questions during this stage. A technical contact who can reply within a day keeps things moving.
You can dispute a finding and give your reasoning, and we'll look at it again against the code. Findings you decide to live with are recorded as acknowledged.
6. Your fixes
You fix the problems and send us the changes. The quality of a fix matters as much as the original finding. New code is new risk, and fixes are where fresh bugs often creep in.
7. Re-review
We review your fixes and confirm each problem is closed. One round of fix re-review is included in every audit, so the final grade reflects the code as it will actually go live.
8. Final report
You get a report with a plain-language summary, exactly what was reviewed and at which version, the threat model, a table showing how each OWASP Top 10 category was tested, every finding with its proof, a retest section, and a grade. It's written to be publishable. Whether it goes public is your call, and we can hold back anything you want kept private until it's fixed. Read more about what a good report contains.
What it asks of you
Mostly three things: freeze the code, answer questions quickly, and fix things promptly. The teams that get the most from an audit treat it as a collaboration.
Ready to start? Request an audit, or browse the FAQ.
Shipping something onchain?
Send us your repo and we’ll return a quote within 24 hours.