OWASP โ the Top 10 lists & the security toolbox
OWASP โ the Open Worldwide Application Security Project โ is a nonprofit community that produces free, vendor-neutral security resources: risk lists, standards, testing guides, cheat sheets and tools. It's the de-facto common language of application security, and you can adopt all of it at no cost.
The project What OWASP is
๐ Open & nonprofit
A global community; every project is free and open, funded by membership & donations โ no vendor agenda.
๐ Knowledge + tools
Documentation (Top 10, ASVS, WSTG, cheat sheets) and software (ZAP, Dependency-Check, Juice Shop).
๐ค Community-driven
Local chapters, conferences, and hundreds of contributor-led projects (flagship, production, incubator).
Flagship The OWASP Top 10 (Web, 2021)
A risk-ranked awareness list of the most critical web-app security risks, refreshed every few years from real-world data. It's an awareness document โ a starting point, not a complete checklist (that's ASVS).
| # | Risk | In a sentence |
|---|---|---|
| A01 | Broken Access Control | Users act outside their permissions (the #1 risk). |
| A02 | Cryptographic Failures | Sensitive data unprotected โ weak/missing encryption. |
| A03 | Injection | Untrusted input runs as code (SQLi, XSS, command). |
| A04 | Insecure Design | Missing security by design โ no threat modeling. |
| A05 | Security Misconfiguration | Insecure defaults, verbose errors, open settings. |
| A06 | Vulnerable & Outdated Components | Known-vulnerable libraries/dependencies. |
| A07 | Identification & Auth Failures | Weak login, sessions, credential handling. |
| A08 | Software & Data Integrity Failures | Unverified updates, insecure CI/CD, deserialization. |
| A09 | Logging & Monitoring Failures | Breaches go undetected โ no logs/alerts. |
| A10 | Server-Side Request Forgery (SSRF) | Server tricked into calling attacker-chosen URLs. |
Flagship OWASP API Security Top 10 (2023)
APIs have a distinct risk profile โ the theme is authorization, per request and per object.
| # | Risk | In a sentence |
|---|---|---|
| API1 | Broken Object Level Auth (BOLA) | Access another user's object by changing an ID (the #1 API risk). |
| API2 | Broken Authentication | Weak/missing auth on endpoints. |
| API3 | Broken Object Property Level Auth | Over-exposing or accepting fields (mass assignment). |
| API4 | Unrestricted Resource Consumption | No rate/size limits โ cost blowup & DoS. |
| API5 | Broken Function Level Auth | Regular users reaching admin functions. |
| API6 | Unrestricted Access to Sensitive Business Flows | Automating a flow (e.g. buy-all-stock) unchecked. |
| API7 | Server-Side Request Forgery | API fetches an attacker-supplied URL. |
| API8 | Security Misconfiguration | Insecure defaults, missing hardening. |
| API9 | Improper Inventory Management | Forgotten old/undocumented ("shadow") APIs. |
| API10 | Unsafe Consumption of APIs | Trusting third-party APIs' data blindly. |
Flagship OWASP Top 10 for LLM Applications (2025)
The newest list โ the risks unique to apps built on large language models and agents.
| # | Risk | In a sentence |
|---|---|---|
| LLM01 | Prompt Injection | Untrusted content hijacks the model's instructions (the #1 LLM risk). |
| LLM02 | Sensitive Information Disclosure | Model leaks secrets, PII or other users' data. |
| LLM03 | Supply Chain | Compromised models, datasets or plugins. |
| LLM04 | Data & Model Poisoning | Malicious training/RAG data corrupts behaviour. |
| LLM05 | Improper Output Handling | Trusting model output โ running it as code/SQL/HTML. |
| LLM06 | Excessive Agency | Over-privileged agents taking harmful actions. |
| LLM07 | System Prompt Leakage | Secrets/logic exposed via the system prompt. |
| LLM08 | Vector & Embedding Weaknesses | RAG/embedding attacks & cross-tenant leakage. |
| LLM09 | Misinformation | Confident hallucinations relied upon as fact. |
| LLM10 | Unbounded Consumption | Runaway token cost / denial-of-wallet. |
Beyond the lists The wider toolbox
The Top 10 raises awareness; these turn it into practice:
โ ASVS
Application Security Verification Standard โ hundreds of testable requirements at three levels. Your real security checklist & contract.
๐งช WSTG
Web Security Testing Guide โ how to actually test for each vulnerability, step by step.
๐ Cheat Sheet Series
Short, focused how-tos for specific topics (auth, CSRF, input validation, JWTโฆ).
๐ท๏ธ ZAP
Zed Attack Proxy โ a free, powerful DAST scanner/proxy for finding vulns in running apps.
๐ฆ Dependency-Check / Track
Software Composition Analysis โ flags known-vulnerable dependencies (maps to A06).
๐ SAMM
Software Assurance Maturity Model โ measure & improve your security program over time.
๐ง Juice Shop
A deliberately-vulnerable app to learn & practise finding the Top 10 safely (great for training/CTFs).
๐ Threat Dragon
A free threat-modeling tool for the design phase (pairs with A04 Insecure Design).
๐ฑ MASVS / MASTG
The mobile equivalents of ASVS & WSTG for iOS/Android apps.
Put it to work Using OWASP across the SDLC
Each resource has a natural home in your development lifecycle:
Cheat sheet
| Question | Short answer |
|---|---|
| What is OWASP? | A nonprofit producing free, vendor-neutral application-security resources. |
| The three Top 10 lists? | Web (2021), API Security (2023), LLM Applications (2025). |
| #1 in each? | Broken Access Control ยท Broken Object Level Auth (BOLA) ยท Prompt Injection. |
| Top 10 vs ASVS? | Top 10 = awareness list; ASVS = verifiable requirements (the real checklist). |
| Free tools? | ZAP (DAST), Dependency-Check (SCA), Juice Shop (practice), Threat Dragon. |
| How to test? | Follow the WSTG; automate with ZAP & dependency scanning in CI. |
Knowledge check
Six questions on OWASP. Pick an answer for each, then check your score.
Next steps
๐ก๏ธ Web ยท API ยท AI Security
The hands-on defenses โ headers, authN/authZ, and AI/agent guardrails for these risks.
๐ต๏ธ Cyber Security
The bigger picture โ mechanisms, attack prevention, and managing a security program.
๐ APIs vs MCP
The API surface โ where BOLA and broken auth live.
๐งฐ Agent Development Kits
Building agents that must resist the LLM Top 10.
๐ All tutorials
The hub โ LLMs, agents, APIs, cloud & more.