OWASP Top 10 -2025 Edition Breakdown

The number one vulnerability since 2021. Found in 94% of tested applications. Access control enforces that users cannot act outside their intended permissions. Broken access control includes IDOR (insecure direct object references), missing function-level access control, privilege escalation from user to admin, CORS misconfigurations that allow unauthorised API access, and JWT token manipulation to elevate privileges.

Previously called "Sensitive Data Exposure" -renamed to focus on the root cause rather than the symptom. Cryptographic failures cover: transmitting sensitive data in clear text (HTTP, SMTP without TLS), using weak or deprecated algorithms (MD5, SHA1, DES), hardcoded cryptographic keys, missing encryption at rest for sensitive data, and improper certificate validation.

SQL injection, OS command injection, LDAP injection, and Cross-Site Scripting (XSS) are all included here. Injection occurs when an application sends untrusted data to an interpreter as part of a command or query. SQL injection remains one of the most destructive vulnerabilities -a single injectable endpoint can lead to full database exfiltration. XSS allows attackers to execute scripts in victims' browsers, enabling session hijacking and credential theft.

New in the 2021 list and reinforced in 2025, insecure design focuses on design-level flaws rather than implementation bugs -missing security controls by design rather than misconfigured ones. An application that never implemented rate limiting on authentication endpoints has an insecure design flaw. No amount of patching fixes a design problem -it requires redesign. This is why threat modeling at the design phase is critical.

Found in 90% of tested applications. Security misconfiguration covers default credentials left unchanged, unnecessary features enabled, verbose error messages exposing stack traces, missing security headers (CSP, HSTS, X-Frame-Options), overly permissive CORS policies, and cloud storage buckets left publicly accessible. As applications increasingly run on cloud infrastructure, misconfiguration has become the dominant attack vector.

Applications using components -libraries, frameworks, operating systems -with known vulnerabilities are directly exploitable. Log4Shell was the most prominent example: a critical vulnerability in a logging library used by hundreds of thousands of applications globally. The 2025 edition adds explicit SBOM (Software Bill of Materials) requirements, reflecting regulatory pressure and the widespread adoption of SBOM as a supply chain security control.

Authentication failures include: permitting weak or default passwords, missing or ineffective MFA, exposing session IDs in URLs, not invalidating sessions on logout, credential stuffing enabled by missing rate limiting, and insecure password reset flows. In 2025, the category specifically calls out authentication bypass vulnerabilities in AI-powered chatbots and LLM-integrated applications as an emerging concern.

This category covers insecure deserialization and CI/CD pipeline integrity failures. An application that automatically updates without verifying the integrity of the update -using a compromised package registry, a tampered npm package, or a supply chain poisoning attack -falls here. The SolarWinds attack is the canonical example. The 2025 edition strengthens the connection to software supply chain attacks and adds guidance on securing AI model pipelines.

Without adequate logging and monitoring, breaches go undetected. The average dwell time -time between initial compromise and detection -is 197 days globally. Logging failures include: no logging of authentication events or high-value transactions, logs not monitored for suspicious activity, log tampering possible (logs stored on the same system as the application), and no incident response procedures triggered by log alerts.

SSRF vulnerabilities allow attackers to induce the server-side application to make HTTP requests to an arbitrary domain -including internal services that are not publicly accessible. In cloud environments, SSRF is particularly dangerous because it can be used to access instance metadata endpoints (AWS IMDSv1) and retrieve IAM credentials. SSRF is elevated to a standalone category in 2025 reflecting its increasing prevalence in cloud-native applications and its role in several high-profile breaches.