In this guide: How to scan your website for malware with free tools: signs of infection, online scanners, server-side scanning, reading scan results, and next steps if…
Malware doesn’t always announce itself. Many infections are designed to hide from site owners while redirecting visitors or injecting spam into search results. A malware scan can reveal certain indicators, but detection depends on coverage, configuration and the techniques used by the threat. No clean report guarantees that a site is uncompromised.
Scan immediately if you notice:
site:yourdomain.com in Google./wp-content/uploads/.Learn how infections happen in What Is Malware?.
Online (remote) scanners check your public pages without accessing your server.
Limitation: remote scanners only see what’s publicly visible. Backdoors and hidden files on your server won’t show up. Use them as a first check, not the only check.
Server-side scanners can inspect files and other components within their configured access and feature scope. Verify database coverage rather than assuming every product includes it.
WordPress security plugins
Hosting scanners
Manual checks via SSH (advanced)
# Recently modified PHP files (last 7 days)
find . -name "*.php" -mtime -7
# PHP files in uploads (should normally be none)
find wp-content/uploads -name "*.php"
# Common suspicious functions (review results carefully)
grep -rl "eval(base64_decode" .
Also check:
<script or unfamiliar domains)..htaccess files for unexpected redirect rules.wp core verify-checksums.Scanners report different severity levels:
Tips:
Full walkthrough: How to Fix a Hacked WordPress Site Step by Step. If you’re not comfortable doing it yourself, professional cleanup services and many hosts can help.
Further reading: WordPress Advanced Administration Handbook
Choose the question before opening a tool. Are you checking a public warning, investigating suspicious files or reviewing an alert after a deployment? These questions need different evidence. A public URL reputation result cannot establish the integrity of every server file. A local file comparison cannot show exactly what a visitor was served from a compromised third-party script. Write the purpose and the systems you are authorized to examine.
A useful malware scan record includes the hostname, relevant paths, tool, scan time, access level and limitations. Mark components outside the scope clearly. This prevents a later reader from interpreting a narrow test as a complete security assessment. Do not scan systems you do not own or have permission to examine, and do not treat a vendor’s marketing description as evidence that its default configuration covers your entire application.
A remote service requests pages from outside the server. It may identify visible injections, warnings or suspicious redirects, but it cannot automatically read private source files and database tables. Conditional behavior can differ by device, location, referrer or time. A clean result for the homepage does not show that every route, account or response variant is safe. Record which public URL was actually checked.
Sucuri SiteCheck states that its remote access and results have limitations. Use it as one observation rather than a certificate. If the tool could not reach the site, record an incomplete test instead of reporting no malware. A firewall challenge, outage or timeout can prevent examination. Repeatedly submitting the same URL without resolving access limitations does not improve the underlying evidence.
Public analysis services are not a safe default destination for private backups, customer exports or password-reset URLs. A URL can contain a token that grants access even when its page title looks harmless. Check the service’s submission and sharing terms before sending anything. Use a public, non-sensitive page only when that is appropriate to the task and you are authorized to submit it.
The VirusTotal explanation of its service describes sharing with examining partners. Do not assume that an ordinary submission is confidential. If analysis requires private material, consult a qualified specialist and use an appropriate agreed process. Never upload the full contents of a potentially compromised hosting account merely because a free malware scan form accepts files.
Before relying on a plugin or host scanner, review the enabled checks, excluded directories, file-size limits and completion status. Confirm that the account has the access required for the intended review. A job may finish with skipped components, or it may fail before reaching important paths. Keep warnings with the report instead of summarizing a partial result as an all-clear.
Choose a schedule that fits the site’s activity and hosting resources. Running several overlapping scanners can create load or conflicting notifications without guaranteeing better coverage. If a malware scan repeatedly stops, use the product’s documented troubleshooting and involve the host. Do not disable unrelated security controls globally simply to obtain a green completion message. Any temporary exception should be scoped, authorized and removed after investigation.
An outdated component or a published vulnerability indicates risk, not proof that exploitation occurred. A modified file indicates a difference, not automatically malicious intent. A known malicious signature is a stronger reason for urgent investigation, but its context and scope still matter. Keep these categories separate in your notes so the team can prioritize correctly without either dismissing a real incident or exaggerating an unconfirmed one.
A malware scan finding should have an owner, affected component and next action. The action may be comparison with a legitimate vendor copy, controlled replacement, further investigation or incident containment. Do not mark every warning ignored simply because the website still loads. Equally, avoid deleting all flagged files before understanding whether they include legitimate custom code or information needed to investigate an incident.
Compare distributed files with a legitimate copy of the relevant version. An integrity check is meaningful only when the reference and scope are clear. WordPress core checksums do not validate every premium plugin, uploaded file, configuration entry or database record. A modified file may require a developer’s review to distinguish an intentional change from an unauthorized one. Keep that decision documented.
Read the WordPress hardening guidance alongside integrity checks. Use official distribution sources rather than downloading a supposedly clean replacement from an unknown mirror. For custom work, consult the maintained repository and deployment record if available. A malware scan cannot compensate for having no reliable record of what code belongs on the site.
The example commands above list candidates for review; they do not diagnose an infection. Recently modified files may be part of a normal update, and suspicious-looking functions may occur in legitimate software. Conversely, malicious code can avoid those exact strings. A search with no matches therefore does not establish that the system is clean. Only an authorized administrator should inspect the relevant server paths.
Avoid running unknown repair scripts copied from a forum. Keep initial inspection separate from destructive changes, preserve appropriate evidence and understand a command before using it. Do not publish credentials or private database content when asking for help. If you cannot explain why a file should be changed, escalate the finding instead of treating a generic search expression as permission to delete it.
Review relevant account changes, scheduled tasks, access logs and configuration alongside a finding. A suspicious public redirect may originate in stored content, a server rule, an external script or a compromised account. Different explanations require different fixes. Record observations without inventing a single cause from one log line. The first visible symptom is not necessarily the first malicious activity.
Use the WordPress user-role guide to assess expected access and the security review record to organize gaps. Keep investigations within the agreed scope. An unknown user should be checked against legitimate ownership records before it is labeled an attacker, while confirmed unauthorized access should be handled promptly through the incident response process.
Protect visitors through a trustworthy containment method and contact the host or responsible security professional. Preserve a restricted copy of relevant evidence where practical before changes. Choose an appropriate cleanup, rebuild or recovery route based on the incident scope. A normal maintenance page may not stop malicious scripts or jobs elsewhere in the account. Do not keep serving harmful content simply to finish a routine checklist.
Follow the hacked WordPress recovery workflow rather than assuming that deleting one file completes the incident. Review credentials, sessions and integration tokens according to the suspected exposure. Close the relevant weakness and validate important business journeys. Use verified recovery points carefully, accounting for newer orders and other data that a rollback might overwrite.
Assign responsibility for reading reports and acting on failed jobs. An unattended weekly notification does not provide a response capability. Set a review frequency appropriate to the site, and revisit it after a migration, major software change or incident. Store results with a consistent time zone and enough context to compare like-for-like scans. Preserve coverage changes so an apparent improvement is not merely a newly excluded directory.
After remediation, repeat relevant checks and monitor for recurrence. A malware scan with no current detection is a useful observation within its scope, not proof of permanent safety. Keep software maintenance, access control and backup restoration exercises active. The WAF guide explains why filtering incoming requests is a separate layer and does not remove an infection already stored on the server.
For each finding, record the component, reported category, scan timestamp, evidence reference, owner and next review date. Add a field for scope limitations and another for the verified outcome. This can be a simple restricted document; it does not require publishing raw logs. Mark an issue resolved only when the responsible person has checked the relevant change. These are suggested fields, not findings from a test of your site.
If two tools disagree, compare what each examined and when. A reputation entry may be cached, while a local scan may use a different signature set or exclude the relevant path. Do not average their severity numbers as though they were one measurement. Investigate the underlying observation and contact the reporting vendor through its supported correction process when a false positive has been established.
Promotional recommendation: review KAWEDMA’s SiteLock offering if you need a website security service to evaluate. Confirm current scanning scope, supported applications, cleanup terms and response responsibilities directly. This recommendation does not establish that every plan includes manual incident response or that Hostreta measured its detection performance.
Affiliate disclosure: Hostreta may earn a commission if you purchase through this link. When a separate hosting review supports changing infrastructure, compare hosting.com’s managed WordPress service. Confirm the exact plan and division of maintenance responsibilities. A new host should not be presented as a guaranteed cure for compromised files, unsafe credentials or unresolved application weaknesses.
Imagine that a file-integrity alert appears shortly after a developer publishes a custom update. Start by matching the file path and time with the authorized deployment record. Ask the developer for the relevant change and compare it with the maintained source. Do not dismiss the alert solely because a release happened that day, and do not delete the file solely because it differs from a generic reference.
If the change is legitimate, document why and follow the scanner’s supported method for managing the finding. Keep the exception narrow enough that future unrelated modifications remain visible. If ownership cannot explain the change, preserve the evidence and escalate. An exception without a reason, owner and review date can quietly conceal the next problem.
After the review, rerun the relevant malware scan and confirm that it completed with the expected coverage. Check the affected business function as well as the report. Record the decision in the maintenance history so the next administrator does not repeat the same uncertain investigation or inherit an unexplained ignored-result list. This example illustrates a review process; it is not a claim that a particular warning on your website is harmless.
We may earn a commission on purchases made through these links, at no extra cost to you.
Don’t miss the chance to participate in the biggest giveaway of this year.