“Audited” is treated in crypto marketing as close to a safety seal. The report behind that word is usually a long, specific, technical document with a defined scope, a list of findings graded by severity, and the auditor’s own commentary on what was and wasn’t checked. Reading the actual report, instead of the badge, is where most of the useful information lives.
Finding the actual report
A credible audit firm publishes its reports, either on its own site or in a public repository, tied to a specific commit hash or version of the code. If a project claims to be “audited” but only links to a logo or a one-line summary with no downloadable report, that is a gap worth noting before trusting the claim at all. The report itself should be traceable to a specific, named firm – checking that the firm is a real, identifiable entity with a track record of other public reports is a reasonable first step before reading further.
What to read inside the report
- Scope. Every audit report should state exactly which files, contracts, or commit hash were reviewed. A report that audited an earlier or different version of the code than what’s actually deployed is reviewing something other than what you’d be interacting with – this is worth checking directly against the verified source on the current contract address.
- Methodology. Reports typically describe whether the review was manual, automated (static analysis tooling), or both, and over what time frame. A short engagement window on a large, complex codebase is a signal about depth, not just a date.
- Findings and severity. Findings are usually graded (critical, high, medium, low, informational, or similar scales that vary by firm). Read the actual description of each finding, not just the count – a report with several low-severity, informational findings and zero critical ones tells a very different story than the number of findings alone suggests.
- Resolution status. This is one of the most frequently skipped sections. Findings are typically marked resolved, partially resolved, acknowledged (the team is aware but chose not to fix it), or disputed. A critical finding marked “acknowledged, not fixed” is materially different from one marked “resolved,” and a summary claim of “audited” tends to erase that distinction entirely.
- Disclaimers. Nearly every audit report includes language stating that the audit does not guarantee the absence of all vulnerabilities, does not constitute investment advice, and reflects the state of the code at the time of review only. These disclaimers are the firm’s own description of the limits of its work and are worth reading in full rather than skipping to the conclusion.
Step by step: checking an audit claim
- Locate the full report from the named firm, not a summary or a badge image.
- Confirm the scope section lists the actual contract(s) in question, and note the commit hash or version reviewed.
- Compare that commit hash or version, where stated, against the currently deployed and verified contract source – code can be redeployed or upgraded after an audit without a new audit necessarily following.
- Read the findings list and their resolution status, especially anything above low severity.
- Check whether more than one firm has reviewed the code; multiple independent reports covering the same scope add more weight than a single report alone.
What an audit report doesn’t prove
An audit is a time-boxed, scoped review by a specific firm using specific methods; it is not a guarantee that no vulnerabilities exist, and it is not an endorsement of the project’s business model, team, or tokenomics. Audit firms routinely miss issues that are later found by others, and a clean report from one firm does not rule out an issue a different reviewer might catch. Most importantly, an audit report describes a specific version of the code; if the contract is upgradeable (see how to check if contract ownership is renounced, which covers proxies) or has simply been redeployed since the audit, the code actually running may no longer match what was reviewed at all.
Reading severity ratings without over-trusting the labels
Severity labels are useful but are ultimately the auditing firm’s own judgment call, applied using that firm’s own internal criteria, which can differ somewhat between firms. A “medium” finding from one firm and a “medium” finding from another aren’t guaranteed to represent exactly the same level of real-world risk, which is one more reason to read the actual description of a finding rather than relying on the severity word alone. A well-written finding describes the specific condition that triggers the issue, the potential impact if exploited, and the firm’s suggested remediation; a thin description with just a label and no real explanation is a weaker piece of work regardless of what severity it’s assigned.
It’s also worth reading the executive summary separately from the detailed findings, since marketing-oriented summaries of “this audit found no critical issues” sometimes gloss over a cluster of high-severity findings that, taken together, could matter a great deal even though none individually carried the firm’s highest label. The detailed findings section, not the summary paragraph, is where the substance actually lives.
Not financial advice
This guide explains how to read a smart contract audit report. It is general education, not financial, legal, or investment advice. An audit, even a thorough one with no unresolved findings, does not make a protocol or token a safe place to put funds.
A quick check on the auditing firm itself
Before weighing a report’s conclusions heavily, it’s worth spending a few minutes confirming the auditing firm is a real, identifiable organization with a visible track record: a firm’s own site listing other, verifiable client reports, a public history of its work being referenced by other credible projects, or coverage of its work from outside its own marketing. A report attributed to a firm with no other discoverable track record, or a firm name that can’t be found anywhere outside the one report in question, is a much weaker piece of evidence than one from an established firm whose other work is independently checkable.
FAQ
If a project has an audit badge, does that mean the code is safe?
Not by itself. The badge only has meaning if you can trace it to a real, downloadable report, check that its scope matches the current deployed code, and confirm any significant findings were actually resolved rather than just acknowledged.
What does it mean if a finding is marked “acknowledged” rather than “resolved”?
It means the team was told about the issue and chose not to change the code to fix it, for whatever reason they gave. This is a materially weaker outcome than a finding the team actually fixed, and it is worth reading the stated reason rather than assuming it’s minor.
Can code change after an audit without a new audit happening?
Yes, especially for upgradeable contracts controlled by a proxy. Always check the commit hash or version the report covers against what is actually deployed right now.
Does having multiple audits mean a project is safer than one with a single audit?
It can be a positive signal, especially if the audits are independent and from different firms covering overlapping scope, but it still isn’t a guarantee. Read what each report actually found rather than counting badges.
Not financial advice. Human-edited; AI-assisted drafting is disclosed in our editorial guidelines. Spotted an error? See how corrections work.