Vulnerability scanning explained: the complete guide to finding security weaknesses

Aperçu de la situation

  • Vulnerability exploitation now drives 20% of all data breaches — a 34% year-over-year increase — making automated scanning a non-negotiable security practice
  • Effective scanning follows a seven-step process from asset discovery through remediation verification, with risk-based prioritization (CVSS + EPSS + reachability) replacing severity-only scoring
  • Six distinct scan types cover different scopes and access levels, and organizations need a combination of approaches since no single scanner catches every weakness
  • Compliance frameworks including PCI DSS v4.0, NIST SP 800-53, CIS Controls, and NIS2 all mandate vulnerability scanning at defined frequencies
  • Scanning alone cannot catch zero-day vulnerabilities or post-exploitation behavior — pairing it with behavioral threat detection creates a complete defensive posture

What is vulnerability scanning?

Vulnerability scanning is the automated process of probing systems, networks, and applications for known security weaknesses by comparing discovered configurations and software versions against vulnerability databases such as the National Vulnerability Database (NVD) and Common Vulnerabilities and Exposures (CVE) entries. It is the systematic discovery step that tells an organization where its defenses have gaps.

The urgency behind scanning has never been greater. According to the Verizon 2025 Data Breach Investigations Report, vulnerability exploitation accounted for 20% of all data breaches — a 34% year-over-year increase. Attackers are moving faster, too. Research from VulnCheck shows the median time-to-exploit has dropped to five days, and 28.96% of entries in CISA's Known Exploited Vulnerabilities (KEV) catalog were exploited on or before the day the CVE was published. When the window between disclosure and exploitation is measured in hours, not months, continuous scanning becomes a survival requirement.

Vulnerability scanning is one component within the broader vulnerability management lifecycle, which encompasses discovery, prioritization, remediation, and verification. Understanding where scanning fits — and where it does not — is critical for building an effective security program.

Vulnerability scanning vs related activities

Security teams often use the terms scanning, assessment, penetration testing, and vulnerability management interchangeably. They are not the same thing.

How vulnerability scanning compares to assessment, penetration testing, and vulnerability management.

Activité Objectif Champ d'application Sortie
Vulnerability scanning Automated detection of known weaknesses Broad — systems, networks, applications Prioritized list of identified vulnerabilities
Vulnerability assessment Systematic evaluation of security posture Broader — includes scanning, configuration review, policy analysis Risk-rated findings with remediation recommendations
Penetration testing Validate exploitability through simulated attacks Targeted — specific systems or attack paths Proof-of-concept exploitation with business impact analysis
Gestion des vulnérabilités Ongoing lifecycle of discovery through remediation Enterprise-wide — continuous program Metrics on risk reduction, remediation rates, and compliance

Scanning identifies known weaknesses through automated tools. Assessment wraps scanning into a broader evaluation that includes manual review and contextual analysis. Penetration testing goes a step further by attempting to actually exploit vulnerabilities to validate real-world risk. Vulnerability management is the continuous program that governs all of these activities across the organization.

How vulnerability scanning works

The vulnerability scanning process follows seven sequential steps. Each builds on the previous one, and the cycle repeats continuously as environments change.

  1. Discover assets — identify all systems, services, and endpoints in scope
  2. Enumerate targets — determine open ports, running services, and software versions
  3. Detect vulnerabilities — compare findings against CVE databases and vendor advisories
  4. Validate findings — correlate results and eliminate false positives
  5. Score and prioritize — assign severity using CVSS, EPSS, and business context
  6. Generate reports — produce actionable findings with remediation guidance
  7. Verify remediation — rescan to confirm vulnerabilities have been resolved

The vulnerability scanning process follows seven sequential steps with rescanning for verification.

The scanner sends probes to target systems, collects responses about open ports and software versions, then compares those fingerprints against known vulnerability signatures. Is vulnerability scanning automated? Yes — the core detection process is automated, though validation, prioritization, and remediation require human judgment and organizational context.

What does a vulnerability scan detect? Scanners identify missing patches, misconfigured services, default credentials, outdated software, insecure protocols, and known application flaws. They cannot, however, detect vulnerabilities for which no signature exists — a critical limitation covered later in this guide.

CVSS v4.0 and risk-based prioritization

For years, organizations relied solely on the Common Vulnerability Scoring System (CVSS) to prioritize remediation. A critical score (9.0 to 10.0) went to the top of the queue. This approach is increasingly inadequate.

CVSS v4.0, released with a new Consumer Implementation Guide in January 2026, adds Environmental and Threat metrics that provide better context. But CVSS alone still does not answer the question that matters most: "Will this vulnerability actually be exploited in my environment?"

That is where complementary scoring comes in. The Exploit Prediction Scoring System (EPSS) estimates the probability a vulnerability will be exploited in the wild within 30 days. Reachability analysis determines whether an attacker can actually reach the vulnerable component through the network. Industry analyses indicate that 84% of 2025 breaches exploited "reachable" vulnerabilities — not necessarily those with the highest CVSS scores.

The reality underscores the urgency. According to the Verizon 2025 DBIR, only 54% of vulnerable devices were fully remediated within the year, with a median of 32 days to patch. Risk-based prioritization ensures limited remediation resources focus on what attackers are most likely to exploit.

How vulnerability scanners use CVE databases

Scanners correlate their findings against three primary data sources. The NVD serves as the main reference database, mapping CVE identifiers to CVSS scores and affected products. CISA's KEV catalog — which expanded 20% in 2025 to 1,484 entries — flags vulnerabilities confirmed to be actively exploited in the wild. Vendor-specific advisory feeds add product-level detail that generic databases sometimes miss.

Modern scanners pull from all three sources to build a layered view of risk. When a scanner identifies a software version on a target system, it checks that version against NVD entries, flags any matches in the KEV catalog for immediate attention, and cross-references vendor advisories for patches or workarounds.

Types of vulnerability scanning

Organizations need different scanning approaches depending on the environment, access level, and target. Six primary types cover the spectrum.

Six vulnerability scanning types and their appropriate use cases.

Scan type Description Best use case Limite principale
Network vulnerability scanning Scans IP ranges, ports, and network services for known weaknesses Discovering exposed services across network security infrastructure Limited application-layer visibility
Web application scanning (DAST) Tests running web applications for OWASP Top 10 and other flaws Internet-facing applications and APIs Cannot analyze source code or backend logic
Host-based scanning Agent-installed or local scans on individual hosts for OS and software vulnerabilities Endpoint and server hardening Requires deployment and maintenance per host
Credentialed (authenticated) scanning Logs into systems to inspect configurations, patches, and internals Deep assessment with full system visibility Requires credential management and access
Non-credentialed (unauthenticated) scanning Probes systems from the outside without credentials External attacker perspective; quick baseline Misses internal configuration issues
Cloud-native and agentless scanning API-based scanning of cloud workloads, containers, and serverless functions Ephemeral cloud security environments May lack depth for complex configurations

Credentialed vs non-credentialed scanning is one of the most impactful decisions in a scanning program. Credentialed scans find significantly more vulnerabilities because they can access system internals — installed software, registry settings, file permissions, and patch levels. Non-credentialed scans provide the attacker's external view, which is valuable for perimeter assessment but misses internal weaknesses.

Internal vs external scanning addresses different threat models. Internal scans assess the network from inside, identifying lateral movement opportunities and misconfigurations. External scans evaluate internet-facing assets the way an outside attacker would. Most security frameworks require both.

Cloud vulnerabilities contributed to 43% of all 2025 data breaches, and edge devices accounted for 22% of exploitation incidents — an eightfold year-over-year increase according to the Verizon 2025 DBIR. These numbers explain why cloud-native and agentless scanning has become essential.

Agent-based vs agentless scanning

Agent-based scanners install lightweight software on each host, providing deeper visibility and real-time monitoring. They excel at continuous assessment but require deployment, updates, and maintenance across every system.

Agentless scanners use APIs and network-based techniques to assess systems without installing anything. They deliver complete coverage without performance impact and are essential for cloud and ephemeral workloads where containers and serverless functions may exist for only minutes.

The industry consensus favors a hybrid approach. Traditional on-premises environments benefit from agent-based depth. Cloud-native environments increasingly favor agentless scanning for its coverage of ephemeral workloads and intrusion detection and prevention across dynamic infrastructure.

Vulnerability scanning in practice

The dual role of vulnerability scanning in MITRE ATT&CK

Vulnerability scanning occupies a unique position in the MITRE ATT&CK framework — it appears on both sides of the battlefield.

On the adversary side, T1595.002 (Active Scanning: Vulnerability Scanning) is a reconnaissance technique under 0043. Attackers use the same scanning tools defenders use to find exploitable weaknesses in target environments. The related technique T1046 (Network Service Scanning) under Discovery (0007) maps how attackers enumerate services after gaining initial access.

On the defensive side, M1016 (Vulnerability Scanning) is an official mitigation that addresses techniques including T1190 (Exploitation d'une application accessible au public), T1210 (Exploitation of Remote Services), and T1195 (Supply Chain Compromise). Proactive scanning reduces the attack surface that adversaries can exploit.

This dual role means defenders should both run their own scans and monitor for adversary scanning activity on their networks — unusual port sweeps, version probes, or enumeration patterns that indicate an attacker is mapping the environment.

How often should you scan?

The gap between compliance minimums and security best practice is dangerously wide.

  • PCI DSS minimum — quarterly internal and external scans
  • CIS Control 7 recommendation — weekly minimum for internet-facing assets
  • Security best practice — continuous or at minimum weekly, given the five-day median time-to-exploit

A risk-based frequency matrix helps organizations allocate scanning resources effectively.

Asset category Recommended frequency Rationale
Critical internet-facing systems Continuous Highest exposure, fastest exploitation timelines
Standard production systems Weekly Balances coverage with operational impact
Development and staging After each deployment Catches vulnerabilities before production
Low-risk internal systems Monthly to quarterly Minimum coverage for compliance

Quarterly scanning creates 45 to 90-day blind spots. When the median time-to-exploit is five days, a quarterly cadence means an organization may be exposed for weeks before the next scan even runs.

Études de cas concrets

MOVEit Transfer (CVE-2023-34362). A zero-day SQL injection vulnerability in the MOVEit file transfer application was exploited before disclosure, affecting 3,000+ U.S. organizations and 8,000+ worldwide. The lesson: internet-facing applications require continuous scanning, and even then, zero-days demand complementary detection approaches.

Log4Shell (CVE-2021-44228). When the Log4j vulnerability emerged, scanners were deployed at scale. Research analyzing 28 projects across 140 scans found scanners achieved 91.4% accuracy — impressive but insufficient. The gap came from transitive dependencies that scanners could not see. CISA's advisory emphasized using multiple scanning approaches for complete coverage.

Trivy supply chain attack (March 2026). An open-source vulnerability scanner was itself compromised through a multi-phase supply chain attack. The lesson: organizations must verify the integrity of their scanning tools, not just their scan results.

React2Shell (CVE-2025-55182). A CVSS 10.0 vulnerability affecting 592,000+ domains, with 172,000+ confirmed exploitable and 30,000+ already backdoored at discovery. The exploitation window collapsed to hours, reinforcing why continuous scanning with automated alerting is essential.

Detecting and preventing vulnerabilities effectively

Building an effective scanning program requires more than choosing the right tools. These best practices address the most common gaps.

  • Run continuous or at minimum weekly scans for internet-facing assets, following CIS Control 7 guidance
  • Use risk-based prioritization incorporating exploitability context (EPSS), reachability analysis, and threat intelligence — not just CVSS scores
  • Integrate scanning into CI/CD pipelines for shift-left security that catches vulnerabilities before production deployment
  • Maintain a comprehensive asset inventory as the foundation — you cannot scan what you do not know exists (NIST SP 800-53 RA-5)
  • Use multiple scanning tools for better coverage, since no single scanner catches everything
  • Combine scanning with endpoint detection and response to cover the gap between vulnerability discovery and active exploitation

The Verizon 2025 DBIR found that only 54% of vulnerable devices were fully remediated within the year, with a median of 32 days to patch. Edge devices accounted for 22% of exploitation incidents — an eightfold year-over-year increase — highlighting the need to extend scanning beyond traditional server and endpoint environments.

Scanning limitations and what scanners cannot find

Honest assessment of scanner blind spots prevents dangerous overconfidence.

  • Zero-day vulnerabilities. Scanners match against known signatures. If no CVE exists, no scanner will find it. This is where zero-day detection through behavioral analysis becomes critical.
  • Business logic flaws. Application-specific logic errors require human analysis and cannot be detected through signature matching.
  • Transitive dependencies. As Log4Shell demonstrated, nested dependencies in software supply chains can hide vulnerabilities that scanners miss.
  • Context-dependent vulnerabilities. Without credentialed access, scanners cannot assess configuration-specific weaknesses that depend on how software is deployed.
  • Active exploitation. Scanners find weaknesses. They do not detect whether an attacker is already exploiting them. That requires threat detection capabilities.

Managing false positives

False positives — scan results that incorrectly flag a vulnerability that does not actually exist — erode trust in scanning programs and waste analyst time.

Common causes include outdated signatures, version string mismatches where a patch was backported without changing the version number, and environmental differences between the scanner's test conditions and the actual system configuration.

Reduction strategies include running credentialed scans for deeper accuracy, keeping scanner signatures current, validating findings with contextual analysis, and correlating results across multiple tools. Credentialed scans are particularly effective because they can verify actual patch levels rather than relying on external version detection.

Vulnerability scanning and compliance

Major regulatory frameworks mandate vulnerability scanning at specific frequencies. This compliance mapping table provides a quick reference.

Compliance framework requirements for vulnerability scanning.

Le cadre Scanning requirement Minimum frequency Champ d'application
PCI DSS v4.0 (Req. 11.3) Internal and external scans; external by ASV Quarterly (four per year) Cardholder data environment
NIST SP 800-53 RA-5 Periodic scans and after significant changes; SCAP-compliant tools Per organizational risk assessment Federal information systems
CIS Controls v8 (Control 7) Automated internal (7.4), authenticated and unauthenticated (7.5), automated external (7.6) Internal quarterly+, external monthly+ Enterprise assets
ISO 27001:2022 (Annex A 8.8) Risk-based documented process for technical vulnerability management Per risk assessment Information assets in scope
Directive NIS2 Vulnerability management including frequent patching Per member state transposition Essential and important entities (EU)
HIPAA Security Rule Risk analysis and remediation including vulnerability scans Per risk assessment Electronic protected health information

EU organizations face additional pressure under NIS2, which maps 70 to 80% to ISO 27001 controls and carries penalties up to 10 million euros for noncompliance. PCI DSS v4.0 applies globally to any organization processing payment card data. Organizations operating across multiple frameworks benefit from aligning to the most stringent requirement — typically CIS Control 7's weekly to monthly cadence — to satisfy all frameworks simultaneously.

Modern approaches to vulnerability scanning

The vulnerability scanning landscape is evolving rapidly. Three shifts define the modern approach.

From periodic to continuous. Continuous vulnerability scanning replaces quarterly or monthly cadences with always-on monitoring. Given the five-day median time-to-exploit, any gap in scanning coverage represents an opportunity for attackers. Continuous scanning integrates with Continuous Threat Exposure Management (CTEM) frameworks, where scanning is one input into a broader program of attack surface assessment and risk reduction.

From CVSS-only to multi-factor prioritization. CVSS v4.0 improves scoring accuracy, but leading programs now layer EPSS, reachability analysis, and threat intelligence feeds to focus on vulnerabilities that are both exploitable and reachable in their specific environment.

From periodic scans to infrastructure-as-code analysis. Cloud-native environments demand new approaches — API-based scanning, container image analysis, and infrastructure-as-code review that catches misconfigurations before deployment.

How Vectra AI thinks about vulnerability scanning

Vectra AI's philosophy starts with a simple premise: assume compromise. When 28.96% of exploited vulnerabilities are weaponized on or before CVE publication day, even the best scanning program will face gaps. Vulnerability scanning identifies weaknesses. AI-driven threat detection identifies what happens when those weaknesses are exploited — the lateral movement, privilege escalation, and data staging that follow initial access. Attack Signal Intelligence does not replace vulnerability scanning. It catches what scanning cannot: zero-day exploitation, identity-based attacks, and post-compromise behaviors that no signature database contains. Together, vulnerability scanning and network detection and response form a complete defensive posture — finding weaknesses before exploitation and detecting attacker behavior when exploitation succeeds.

Tendances futures et considérations émergentes

The vulnerability scanning discipline faces significant evolution over the next 12 to 24 months, driven by three converging forces.

Exploits are accelerating. The collapse of time-to-exploit from weeks to days — and increasingly to hours — will push organizations toward real-time scanning capabilities integrated with automated response. React2Shell's exploitation timeline, where 30,000+ domains were backdoored before most organizations completed their first scan, previews this future.

AI is reshaping both offense and defense. Attackers are using AI to identify and exploit vulnerabilities faster. Defensive scanning tools are incorporating machine learning to improve detection accuracy, reduce false positives, and predict which vulnerabilities are most likely to be exploited next. EPSS adoption will continue to grow as organizations move beyond CVSS-only prioritization.

Regulatory pressure is increasing. NIS2 enforcement across the EU, PCI DSS v4.0's expanded requirements, and emerging frameworks around critical infrastructure will push organizations toward more frequent, more comprehensive scanning programs. Organizations that have built compliance-minimum programs will need to upgrade to continuous monitoring.

Scanning tools themselves become targets. The 2026 Trivy supply chain compromise demonstrated that vulnerability scanners are high-value targets for attackers. Expect increased scrutiny around scanner integrity verification, signed updates, and supply chain security for security tooling.

Organizations preparing for these shifts should invest in three areas: continuous scanning infrastructure that eliminates gaps between scan cycles, multi-factor prioritization that combines CVSS, EPSS, reachability, and threat intelligence, and behavioral detection capabilities that catch exploitation when scanning inevitably misses something.

Conclusion

Vulnerability scanning is no longer optional — it is a foundational security practice that every organization must implement effectively. With exploitation driving 20% of breaches, time-to-exploit collapsing to five days, and 50,000 new CVEs disclosed annually, the cost of neglecting scanning grows every quarter.

The path forward requires moving beyond compliance minimums. Build a risk-based program that scans continuously, prioritizes using CVSS alongside EPSS and reachability analysis, and covers the full environment including cloud workloads, containers, and edge devices. Combine multiple scanning approaches for coverage depth, and layer in behavioral detection to catch what scanners cannot.

Start with a clear asset inventory, implement continuous scanning for your most critical systems, and build from there. For organizations ready to close the gap between vulnerability discovery and active threat detection, explore how Vectra AI's Attack Signal Intelligence complements scanning programs by detecting attacker behaviors that no vulnerability database can capture.

Foire aux questions

What is vulnerability scanning?

How does vulnerability scanning work?

What is the difference between vulnerability scanning and penetration testing?

How often should vulnerability scans be performed?

What are the limitations of vulnerability scanners?

Is vulnerability scanning required for PCI DSS compliance?

What is continuous vulnerability scanning?