Help for owners
I got a pentest or security scan report and I do not know what to fix first
Short answer
Do not fix the list top to bottom. Sort findings by what an attacker could actually reach, group the ones that share a cause, and ask the tester which are confirmed and which are only scanner guesses. Fix in small changes with a backup, then ask for a retest so the report can be closed.
Does this sound like your situation?
- A customer or insurer asked for a security test and I now hold a long PDF.
- The report has dozens of findings labelled critical, high and medium and none of it is explained.
- My developer says some of it is not a real problem, and I cannot judge.
- The same finding appears many times on different pages.
- I need to show the customer the issues are fixed by a date.
Why does this happen?
Scanners and humans produce different reports
An automated scan matches your site against known patterns and can flag things that are not exploitable in your setup. A human tester tries to chain issues together. Both reports are called findings, and the first needs checking before anyone acts.
Many findings share one cause
One missing setting or one outdated library can show up as a dozen lines. Fixing the shared cause removes them together, and fixing each line separately wastes time.
Severity labels are generic
A label like critical or high describes the type of weakness, not the damage in your business. A flaw on an admin page nobody can reach may matter less than a medium flaw on the page where customers pay.
The usual web weaknesses repeat
OWASP publishes the Top 10 as a reference standard for the most critical web application security risks, reflecting broad consensus among security professionals. Many reports are organised around it, so it is a good key for reading yours.
Confirmed by owasp.org
Fixes can break things
Tightening a browser security policy can block a script your site depends on, and upgrading a library can change behaviour. A fix that is rushed into production can cause the outage the report was meant to prevent.
What can I do about it today?
- Keep the report privateIt describes how to attack you. Share it only with the people fixing it, and not by open chat or public tickets.
- Ask the tester three questionsWhich findings did you confirm by hand, which are scanner output only, and what would you fix first on my business? A good tester will answer in plain words.
- Mark what is customer-facingNext to each finding, write whether it is on a page customers use, on an admin area, or on a test system. That order is a better start than the severity column.
- Take a backup before any changeConfirm with your host or developer that a fresh backup exists and that they have restored from one before. Do not let anyone patch production without it.
- Tell the customer a plan, not a promiseIf a deadline matters, reply with the order of work and a date for the retest. Do not claim the site is secure.
- Ask for a retestA report is closed when the tester checks the fix. Ask whether a retest is included, and keep the second report with the first.
When should I bring someone in?
- Nobody on your side can tell which findings are real.
- The fixes touch payments, logins or customer data.
- A contract or customer deadline depends on the result.
How do I handle it?
- Read the report with the tester's notesI go through each finding, mark it confirmed, unconfirmed or not applicable, and write a plain-language reason.
- Group by cause and by exposureFindings with one root cause are merged, then ranked by what a stranger on the internet could reach.
- Fix in small, reversible stepsOne change at a time with a backup before it, a check on the live system after it and a way back.
- Collect proof as I goEach fix gets a short note: what changed, how I checked it. Those notes are what you send the customer.
- Hand the report back to the testerOnce fixes are in, the tester retests, and anything left is written down as accepted risk with an owner and a reason.
The price depends on what I find, so I quote in writing after a short discovery call. Tell me what is happening and I reply within 24 hours.
What do owners ask about this?
Do I have to fix every finding?
No. Some are not exploitable in your setup, and some cost more to fix than the risk. Decide each one on purpose and write the reason down.
Can the person who built the site fix it?
Often yes. A second person to review the changes helps, because the same author may share the same blind spot.
Will fixing make me pass the customer's review?
I cannot promise that. The customer decides, and some reviews also ask for policies and records, not just fixes.
What does the OWASP Top 10 have to do with my report?
It is a reference list of the most critical web application risks. If your report cites it, that gives you the common name for each type of weakness.
Sources
How was this page made?
I drafted this page with AI assistance from the official documents listed under Sources. Found a mistake? Tell me.
For your developer: the technical fixes
found 12 vulnerabilities (3 low, 6 moderate, 3 high)npm audit fix runs, but the vulnerabilities are still listedRefused to load the script 'https://cdn.example.com/lib.js' because it violates the following Content Security Policy directive: "script-src 'self'".The browser blocks a script because of your Content Security Policy