Smart Contract Audit Checklist: What Auditors Check
A serious audit tests contract code against the attacks that have caused real losses. This checklist explains what auditors check, how to prepare, and how to read the report.
Key Takeaways
- A smart contract audit checklist tests code against known attack patterns, from access control to oracle safety, before launch.
- An audit reduces risk but never guarantees safety, since it reviews one code version at one point in time.
- Access control led the OWASP 2026 ranking, so privilege checks deserve the closest attention.
- Read every audit report by scope, severity, and fix status, not by the auditor name on the cover.
A serious smart contract audit checklist is a structured review that tests whether contract code does exactly what its design promises, and nothing else. It covers access control, business logic, oracle safety, arithmetic, reentrancy, upgradeability, and emergency controls, and it ends with a written report that grades every finding by severity. A smart contract audit is a line-by-line security review of contract code by independent specialists, using automated tools and manual reading, carried out before the contract handles real user funds. The checklist matters because a deployed contract is far harder to change than an ordinary web application, so flaws found after launch can be exploited immediately. This article explains what a serious audit checks, how auditors work through the list, how a team should prepare, how to read the report, and how to choose an auditor.
What Is a Smart Contract Audit?
A smart contract audit is an independent security review of blockchain contract code before deployment. Auditors combine automated scanning with manual line-by-line reading and targeted exploit testing, then deliver a report listing each finding, its severity, and a recommended fix. An audit is not a guarantee, an endorsement, or legal clearance. It is evidence about one version of the code at one moment in time.
A smart contract is a program stored on a blockchain that moves funds or records agreements automatically when its conditions are met. Because the code runs exactly as written, a single logic error can be triggered by anyone who understands it. According to Chainalysis, more than $3.4 billion in cryptocurrency was stolen between January and early December 2025, with the February compromise of Bybit alone accounting for around $1.5 billion of that total. Auditors therefore fix the scope in writing first: which contracts are reviewed, which commit hash identifies the code, and which parts are out of scope.
What Does a Smart Contract Audit Checklist Cover?
A smart contract audit checklist covers the vulnerability classes that have caused real losses, organised so nothing important is left to memory. According to OWASP, the Smart Contract Top 10 for 2026 was built on 122 deduplicated smart contract incidents from 2025, with total losses of $905.4 million. According to OWASP data, access control vulnerabilities alone accounted for $220.0 million across 30 incidents, which is why it sits at the top of the ranking.
| OWASP 2026 category | What auditors test |
|---|---|
| Access Control | Who can call privileged functions, and whether roles are set correctly |
| Business Logic | Whether the economic rules match the written design |
| Oracle Manipulation | Whether market data feeds can be distorted to misprice assets |
| Flash Loans | Whether a single transaction loan can distort balances |
| Input Validation | Whether untrusted inputs are checked before they change state |
| Unsafe External Calls | Whether calls to other contracts can fail silently |
| Arithmetic Errors | Whether rounding errors can drain value over many transactions |
| Reentrancy | Whether an external call can re-enter a function before its state is settled |
| Integer Overflow | Whether numbers can wrap past their limits and corrupt balances |
| Proxy and Upgradeability | Whether upgrade paths and storage layout are safe |
Access control gets the closest attention because it causes the largest losses. Auditors list every state-changing function and ask who can call it, under what conditions, and what happens if that caller loses control of their key. They check role assignments, ownership transfer, and timelocks, and test that accounts without a role are rejected. An uninitialised proxy or a public function meant to be internal is a critical finding.
Reentrancy is an attack in which a contract calls an outside contract before updating its own state, and the outside contract calls back into the first function while the balances are still stale. Auditors trace every external call, check that state is updated before funds move, and look for cross-function paths where a guard on one function does not protect another. Arithmetic and rounding receive the same care, because small errors compound.
Oracle safety and flash loan resistance are tested together, because a large temporary loan inside one transaction is a standard way to distort the market data a contract reads. Auditors ask what happens when a data feed is stale, delayed, or manipulated, and whether a single source can misprice a liquidation. Upgradeability and proxy safety form the newest category in the OWASP 2026 ranking: auditors verify that proxies are initialised exactly once, that storage layout cannot collide between versions, and that only an authorised admin can trigger an upgrade. When a team plans a new contract, the design choices made during development decide how much of this checklist the audit will pass, which is why experienced blockchain and Web3 development services treat the checklist as a design guide rather than a final hurdle.
How Does a Smart Contract Audit Checklist Work in Practice?
A smart contract audit checklist works in practice as a staged review: scope is fixed, automated tools flag known patterns, human auditors read the code and test the logic, findings are discussed, and fixes are retested before launch. A practical introduction is covered in the guide to smart contracts for business in 2026.
- Scope and freeze. The team and the auditors agree the contracts, the commit hash, and the intended behaviour in writing, then the code is frozen for the review.
- Automated scanning. Static analysis tools such as Slither and Mythril flag reentrancy patterns, unsafe calls, and risky operations.
- Manual review. Senior auditors read the code line by line, trace call paths, and test whether the business logic matches the specification.
- Exploit testing. Auditors build proofs of concept for suspicious findings, using fuzzing and targeted tests to show whether an attack path is real.
- Findings and discussion. The team receives a draft list of issues with severity ratings, and both sides discuss which findings are genuine.
- Report and retest. The final report records every finding and its status, and the auditors retest each fix to confirm it resolves the issue.
Manual review is where most high-impact logic flaws are found, because tools can flag patterns but cannot judge whether a protocol behaves as intended. Auditors write down the invariants that must always hold, such as total shares matching total deposits, and then try to break each one.
Why Does an Audit Not Guarantee Safety?
An audit does not guarantee safety because it is a time-limited review of specific code by specific people, and it can only report what those people found in the time they had. Auditors can miss issues, especially in novel business logic with no known pattern to match, and new attack techniques appear after the review. OWASP states that its Smart Contract Top 10 is a risk prioritisation and awareness tool, not a certification, a security guarantee, or a substitute for a full audit.
Scope limits are the most common reason a guarantee would be false. If a later upgrade changes the logic, or an unaudited integration is added afterwards, the original report says nothing about the new risk. Off-chain parts such as websites, admin panels, and servers are also outside a contract audit, so they need their own security review.
How Should a Team Prepare Code Before an Audit?
A team should prepare code before an audit by freezing a clean, documented, and fully tested version of the contracts, so auditors spend their time on security, not on guessing intent. Preparation is the single biggest influence on audit quality. Teams still shaping the business case can ground the design in real blockchain use cases for business in 2026.
- Freeze the code at a recorded commit hash, with build instructions, compiler version and settings, and locked dependencies.
- Write a short specification stating the intended behaviour, the invariants that must always hold, and the roles for each privileged function.
- Provide a full test suite that passes, including edge cases, failed transactions, and access control from the wrong account.
- Run static analysis tools first and fix or document every warning, so auditors start from a clean baseline.
- Remove dead code, debug functions, and commented out logic that could confuse a reviewer.
- Document deployment scripts, ownership handover, upgrade procedures, and key management.
- List known issues and accepted trade-offs openly, with the reason each was accepted, so auditors can focus on unknown risks.
Good preparation also means naming a single point of contact who can answer auditor questions quickly. A team that arrives with clear docs, passing tests, and an honest list of known issues will receive a deeper review.
How Do You Read a Smart Contract Audit Report?
You read a smart contract audit report by checking the scope and code version first, then the status of every critical and high finding, and only then the summary. The scope tells you what was reviewed, the version pins the exact code, and the finding statuses tell you what risk remains in the code deployed today.
Findings are usually graded on a five level scale. A critical finding means an attacker could realistically steal funds directly or freeze the protocol, so an unresolved critical issue in deployed code is a clear warning sign. A high finding means serious loss is possible under realistic conditions and must be fixed and retested. A medium finding needs more specific conditions, but a cluster of them can signal weak engineering discipline. Low findings cover edge cases, while informational findings cover code quality notes.
The status of each finding matters as much as its severity. Fixed means the team changed the code and the auditor checked the change. Acknowledged means the team knows about the issue and chose not to fix it, so every acknowledged item deserves a reading of the reason. Partially fixed or mitigated means the risk was reduced but not removed. Before trusting any report, confirm it on the auditor own website or repository, and check that the commit hash matches the deployed code.
How Do You Choose a Smart Contract Auditor?
You choose a smart contract auditor by weighing public evidence of past work, depth of manual review, and clarity of process, rather than by reputation alone. Start with the auditor public reports: read two or three full reports in your contract type and judge whether the findings show genuine understanding of the business logic.
Ask how the review will be staffed and what it will include. A serious engagement names the auditors, explains the mix of automated scanning and manual review, includes exploit testing for high severity findings, and includes a retest of fixes. Ask for a sample report, check experience with contracts similar to yours, and confirm independence, since an auditor with a financial interest cannot review it impartially.
Frequently Asked Questions
What does a smart contract audit check first?
A smart contract audit checks scope and access control first. Auditors confirm which contracts and code version are in scope, then map every privileged function to the role allowed to call it. A missing role check can hand full control to an attacker.
How long does a smart contract audit take?
The duration of a smart contract audit depends on the size and complexity of the codebase, the contracts in scope, and how well the team has prepared. A small and well documented contract can be reviewed quickly, while a large protocol with lending, vaults, and governance takes far longer.
Can a smart contract be audited after deployment?
A smart contract can be audited after deployment, and a post-deployment review is better than no review at all. The audit still identifies flaws and grades them by severity, but fixing them is harder once the contract is live and holding funds. Upgradeable contracts can be patched, while immutable contracts may need redeployment.
What is the difference between an audit and a bug bounty?
An audit is a structured, time-limited review by named specialists before launch, while a bug bounty is an ongoing open invitation for researchers to test a live system. An audit gives a deep review of a fixed code version, and a bug bounty gives continuous coverage after launch. Serious teams use both, in that order.
Should every smart contract be audited before launch?
Every smart contract that holds user funds, controls permissions, or moves value should be audited before launch, because deployed code is hard to change and flaws can be exploited repeatedly. A simple contract with no funds at risk and no privileged functions may need only careful internal review and thorough testing.
A serious audit tests a contract against the attacks that have caused real losses, and this checklist is the standard a team should hold any review to. Skills Technology Solutions, a software agency based in Rahim Yar Khan, Pakistan, serving clients in Pakistan and the UK, can help you put this into practice. Contact our team to discuss your project.
Related Articles
Keep reading — more guides you may find useful.
Related Services
Services connected to this guide.
