Events Securing Finance in the Age of AI · Secure.com X ZIG Finance · London, 12 Nov Register now

Types of Vulnerabilities: 8 Security Flaws Attackers Exploit Most

Learn the 8 main types of vulnerabilities in cyber security, with real examples, the OWASP Top 10, key scan types, and simple fixes.

Cracked shield with an open padlock surrounded by code, cloud, and network icons

TL;DR

A vulnerability is a weak spot that lets an attacker in. The main types of vulnerabilities are code flaws, network gaps, misconfigurations, broken access, unpatched software, risky APIs, supply chain flaws, and human error. You can’t fix them all at once, so find every flaw, rank it by real risk, and close the ones attackers can reach first.

By mid-November 2026, VulnCheck had counted more than 100,000 new CVEs for the year. That’s more than double the total for all of 2025. However, attackers will only ever use a small share of them, so the hard part is knowing which ones matter to you. Learning the main types of vulnerabilities is the best place to start.

What Is a Vulnerability in Cyber Security?

A vulnerability is a flaw or weakness in a system that an attacker can use to cause harm. It can live in code, a network, a cloud setting, or a person’s habits. On its own, a vulnerability does nothing. The damage starts when someone exploits it.

Vulnerability vs Threat vs Exploit vs Risk

People mix up these four words all the time. Here’s how they differ.

TermWhat it meansSimple example
VulnerabilityA weak spot in a systemA login page with no limit on password tries
ThreatSomeone or something that could cause harmA criminal group that guesses passwords
ExploitThe method or code used to abuse the weak spotA script that tries thousands of passwords a minute
RiskThe chance of harm and how bad it would beHigh, if that login protects customer data

How Vulnerabilities Are Tracked: CVE, CWE and CVSS

Three labels show up in almost every security report. A CVE is an ID for one specific flaw in one product. Meanwhile, a CWE names the type of weakness behind it, such as SQL injection. CVSS scores the flaw’s severity on a scale of 0 to 10. For a closer look, read our guide on how CVE and CVSS work together.

8 Types of Vulnerabilities in Cyber Security

Most security flaws fit into one of these eight types of vulnerabilities. They overlap, and a single breach often chains two or three together.

Map of the 8 types of vulnerabilities grouped into code and apps, infrastructure, identity and people, and supply chain

1. Software and Code Vulnerabilities

These are mistakes developers make when they write an app. For example, the most common ones let an attacker slip their own commands into an app through a form, a link, or a file upload. Cross-site scripting and SQL injection hold the top two spots on MITRE’s 2025 CWE Top 25, the yearly list of the most dangerous software weaknesses.

Common examples:

  • SQL injection: an attacker types database commands into a search box or login field.
  • Cross-site scripting (XSS): a harmful script runs in another user’s browser.
  • Buffer overflow: an app takes in more data than it can hold, which can crash it or run attacker code.
  • Missing authorization: an app forgets to check whether a user is allowed to do something.

How to fix it: Check every input, use safe database queries, and scan code before it ships.

2. Network Security Vulnerabilities

These are gaps in how devices and systems connect. Open ports, old protocols like Telnet, weak Wi-Fi encryption, and flat networks with no zone separation all count. Edge devices are also a growing target. In Verizon’s 2025 DBIR, edge devices and VPNs accounted for 22% of vulnerability exploitation cases, up from 3% the year before.

How to fix it: Close unused ports, split the network into zones, and patch firewalls and VPNs first.

3. Security Misconfigurations

A misconfiguration is a setting that leaves a door open. For example, think of default admin passwords, a cloud storage bucket set to public, or error pages that show too much detail. OWASP ranks security misconfiguration #2 in its 2025 Top 10. However, version-based scanners often miss these because many misconfigurations never get a CVE.

How to fix it: Start with hardened settings, remove defaults, and check cloud settings regularly. Our guide on preventing cloud misconfiguration walks through the most common ones.

4. Broken Access and Identity Weaknesses

These flaws let people reach data or actions they shouldn’t. This includes users with too many permissions, shared admin accounts, weak MFA, and apps that skip permission checks. Broken access control holds the #1 spot in OWASP’s 2025 Top 10.

How to fix it: Give each person only the access they need, turn on MFA everywhere, and review permissions every quarter.

5. Unpatched and Outdated Software

This is the most avoidable type. The vendor already released a fix, but nobody installed it. In 2017, WannaCry ransomware used a Windows flaw called EternalBlue that Microsoft had patched about two months earlier, and it still spread to more than 150 countries. Patching is still slow today. In fact, Verizon’s 2026 DBIR found that organizations fully fixed only 26% of critical flaws in CISA’s Known Exploited Vulnerabilities catalog in 2025, down from 38% the year before.

How to fix it: Patch flaws that are already being exploited first, track software that has reached end of life, and set clear fix deadlines.

6. API and Third-Party Vulnerabilities

Modern apps talk to each other through APIs. A weak API can leak data when it skips checks on who is asking or what they’re allowed to see. Plugins and outside tools also bring their own flaws. In 2025, WordPress security firms Patchstack and Wordfence assigned 5.3 times more CVEs than Microsoft, Google, and Apple combined, most of them in third-party plugins.

How to fix it: Keep a list of every API you run, require strong authentication on each endpoint, and vet vendors before you connect them.

7. Supply Chain Vulnerabilities

A supply chain flaw lives in software you didn’t write but depend on. So when attackers hit a popular library or vendor, everyone who uses it inherits the problem. In 2020, attackers hid malware in a SolarWinds update that reached about 18,000 organizations. A year later, the Log4j flaw put many Java apps at risk overnight. As a result, OWASP added Software Supply Chain Failures as #3 in its 2025 Top 10.

How to fix it: Keep a software bill of materials (SBOM), verify the packages you pull in, and watch vendor security notices.

8. Human Vulnerabilities

People make mistakes, and attackers plan for that. For example, phishing emails, reused passwords, and fake login pages open doors that technology alone can’t close. Verizon’s 2026 DBIR found a human element in 62% of breaches.

How to fix it: Train staff with real examples, use phishing-resistant MFA, and make reporting a strange email quick and easy.

Known vs Zero-Day: How Vulnerabilities Get Exploited

A known vulnerability has a public CVE and usually a patch. In contrast, a zero-day is a flaw the vendor doesn’t know about yet, so no patch exists. Zero-days grab headlines, but known flaws are the bigger problem for most teams, because patches often sit uninstalled for months. For example, Verizon’s 2026 DBIR shows how much this matters: exploiting a vulnerability was the first step in 31% of breaches, which made it the top way attackers got in.

Vulnerability statistics: 100,222 CVEs published in 2026, 1.4% exploited, 31% of breaches started with an exploited vulnerability

What Makes a Security Vulnerability Exploitable?

Not every flaw is dangerous. A vulnerability turns into a real problem when:

  • An attacker can reach it, often straight from the internet.
  • Working exploit code is public or easy to write.
  • It needs little or no login and no user action.
  • It runs on a system that holds valuable data or provides broad access.

That’s why attackers use so few published CVEs. Of the 100,000+ CVEs published in 2026 so far, about 1.4% have made it onto VulnCheck’s list of known exploited vulnerabilities.

Remote, Local and Zero-Day Exploits

Exploit typeHow it worksExample
RemoteThe attacker abuses the flaw over a networkHitting an unpatched VPN from the internet
LocalThe attacker needs access to the device firstUsing a driver bug to become an admin
Zero-dayAttackers use the flaw before a patch existsA browser bug used before the vendor knows about it

The OWASP Top 10 (2025) at a Glance

OWASP’s Top 10 is the best-known list of web app security risks. Here is the 2025 version and where each item fits among the 8 types of vulnerabilities above.

RankOWASP categoryFits under
A01Broken Access ControlBroken access and identity
A02Security MisconfigurationMisconfigurations
A03Software Supply Chain FailuresSupply chain
A04Cryptographic FailuresSoftware and code
A05InjectionSoftware and code
A06Insecure DesignSoftware and code
A07Authentication FailuresBroken access and identity
A08Software or Data Integrity FailuresSupply chain
A09Security Logging and Alerting FailuresMisconfigurations
A10Mishandling of Exceptional ConditionsSoftware and code

Types of Vulnerability Scans

A scan is how you find these flaws before attackers do. Because each scan type covers a different layer, most teams run several to catch all types of vulnerabilities.

Scan typeWhat it checksWhat it finds
Network scanDevices, ports and servicesOpen ports, old protocols, missing patches
Host scanServers and laptops, from the insideOutdated software, weak local settings
Web app scanLive websites and APIsInjection, XSS, broken login flows
Cloud scanAWS, Azure and Google Cloud settingsPublic storage, risky permissions
Container scanContainer images and their packagesKnown CVEs in base images and libraries
Code and dependency scanSource code and open-source librariesCoding flaws and supply chain risk

Two more choices shape every scan. An authenticated scan logs in, so it can see missing patches deep inside a server. In contrast, an unauthenticated scan only sees what an outsider sees. External scans test what faces the internet, and internal scans test what an attacker could reach after getting inside. In short, a healthy program runs all four.

Vulnerability scan cheat sheet: network, host, web app, cloud, container, and code scans, plus authenticated, unauthenticated, external, and internal

Activity: Identify Types of Vulnerabilities

Many security courses ask students to match a scenario to the right vulnerability type. Try these five, then check your answers below.

  1. A company’s login page lets anyone guess passwords forever with no lockout.
  2. A cloud storage bucket full of customer files is set to public.
  3. A server still runs a software version that got a security fix last year.
  4. An employee types their password into a fake Microsoft 365 page.
  5. A popular open-source logging library inside your app has a newly found flaw.

Answers: 1 is broken access and identity, since there’s no limit on login tries. 2 is a security misconfiguration. 3 is unpatched software. 4 is a human vulnerability. 5 is a supply chain vulnerability.

How to Reduce Vulnerabilities Across Your Stack

You can’t fix all of these vulnerability types at once. Instead, these habits cut the most risk for the effort:

  • Know what you own. Keep a live list of devices, cloud accounts, apps, and APIs.
  • Fix what’s exploited first. CVSS alone can mislead you, so rank flaws by real exploit activity and business impact. Our exposure prioritization playbook shows how to build that queue.
  • Harden by default. Remove default passwords and lock down cloud settings from day one.
  • Limit access. Give people and apps only the permissions they need.
  • Scan all the time. A monthly scan misses everything that shows up in between.
  • Verify the fix. Rescan to confirm a patch worked before you close the ticket, and track metrics like time to fix by severity.

Stop Sorting Findings. Start Closing Them.

Most teams have plenty of scan results and not enough hours to sort them and push fixes through. Secure.com runs that work for you as one service on one bill: AI sorts the findings, and security experts decide what gets fixed.

  • Managed Exposure Management finds weak spots across your cloud, systems, and apps, and maps them to real CVEs.
  • We rank findings by exploit activity and business impact, so your team fixes the most critical issues first.
  • Experts work with your system owners to drive each fix, then rescan to verify the fix worked.
  • A vCISO turns the results into a security plan, compliance evidence, and board-ready reports, without a full-time hire.

You tell us what needs protecting, and we run the program.

FAQs

What are the four types of vulnerability?
In cybersecurity, one common split is network, operating system, human, and process vulnerabilities. Disaster management uses a different four: physical, social, economic, and environmental.
What are the six types of vulnerability?
Lists vary by source. A common set of six covers software flaws, network weaknesses, misconfigurations, API and third-party risks, supply chain flaws, and unpatched software. This guide adds broken access and human error to make eight types of vulnerabilities.
What are the top 10 common vulnerabilities according to OWASP?
Broken access control leads the OWASP Top 10 for 2025, followed by security misconfiguration and software supply chain failures. The full list of 10 is in the table above.
What are the 7 types of cybersecurity?
The seven areas people name most often are network, application, cloud, endpoint, data, identity, and IoT security. Each one protects a different layer where the vulnerabilities above can show up.
What is the difference between a vulnerability scan and a penetration test?
A vulnerability scan is automated and checks many systems for known flaws. In a penetration test, people try to exploit flaws the way a real attacker would, which shows which ones are truly dangerous. Most teams scan constantly and run penetration tests once or twice a year.