SSRF Explained: Server-Side Request Forgery (OWASP A10) — 2026
May 17, 2001 · by Pentevo
Server-Side Request Forgery (SSRF) tricks a server into making requests it shouldn't — to internal systems, cloud metadata services, or other resources the attacker can't reach directly. It's #10 on the OWASP Top 10 and has become especially dangerous in cloud environments.
The core idea
Many apps fetch URLs on the server's behalf — loading an image from a link, calling a webhook, importing a document from a URL. If the app takes a user-supplied URL and fetches it without restriction, an attacker can point it inward: at internal services, admin panels, or cloud APIs that trust requests coming from the server itself.
Why it's so dangerous in the cloud
Cloud platforms expose an internal metadata service that returns instance information — sometimes including credentials. A server with SSRF can be coerced into querying that endpoint and handing the response back to the attacker. That's how some major cloud breaches happened: SSRF → metadata → credentials → wider access.
Common scenarios
- Fetching a user-supplied URL (image/document import, webhooks).
- "URL preview" features.
- PDF/screenshot generators that render a given URL.
- Any "enter a URL and we'll connect to it" feature.
How to prevent it
- Allow-list destinations — only permit the specific hosts/protocols the feature genuinely needs. Deny everything else.
- Block internal ranges — reject requests to private IPs, loopback, and cloud metadata addresses.
- Validate and re-resolve — beware of redirects and DNS tricks that bounce an allowed URL to an internal one.
- Don't return raw responses to the user when avoidable.
- Network segmentation — limit what the server can even reach internally.
- Use hardened cloud metadata options (e.g., session-token-required metadata) where available.
How it's found
Testers supply URLs pointing at internal/metadata targets and watch the server's behavior — within an authorized pentest, using tools like Burp Suite. SSRF often chains with other issues, which is where reasoning-based AI testing shines.
Related: Broken Access Control, SQL Injection, XSS. Learn it all free at the Pentevo Academy.
Related reading
VPN Explained: What It Does (and Doesn't) — 2026
How a VPN actually works, what it protects you from, what it doesn't, and how to choose one — without the marketing hype.
FundamentalsWhat Is Penetration Testing? A Beginner's Guide (2026)
A plain-English guide to penetration testing: what it is, the five phases, the main types, and how it differs from a vulnerability scan.
FundamentalsZero-Day Vulnerabilities Explained (2026)
What a zero-day vulnerability is, why it's so dangerous, how zero-day exploits are used, and what defenders can do about the unknown.
FundamentalsWhat Is a CVE? Understanding Vulnerability IDs (2026)
What CVE means, how the numbering works, how CVSS severity and EPSS scores help you prioritize, and how to track the CVEs that matter.
Practice this hands-on
Pentevo Academy turns these concepts into guided lessons, videos and quizzes — free.
Start learning free