Skip to content
AdminKit

SSRF

What is SSRF?

SSRF (Server-Side Request Forgery) is a vulnerability where an attacker tricks a server, or an AI agent calling tools on its behalf, into requesting a destination the attacker chooses rather than the one intended, often reaching internal or metadata addresses. AdminKit's MCP hardening guide covers where this risk enters an agent's tool calls.

By AdminKit team, Netbase JSC · Updated

SSRF (Server-Side Request Forgery) is a vulnerability where an attacker tricks a server, or an AI agent calling tools on its behalf, into requesting a destination the attacker chooses rather than the one intended. MITRE’s CWE-918 catalogs it as a distinct weakness class. It’s a specific case of the broader confused deputy problem, where a more-privileged party — the server, or an agent calling an MCP tool — is tricked into using its own access on an attacker’s behalf. A classic example is a “fetch this image URL” feature pointed at an internal address like a cloud metadata endpoint, getting the server to reach places that aren’t reachable from the public internet at all. The risk has grown with AI agents and MCP tool integrations, because an agent that can fetch arbitrary URLs on a user’s behalf is itself a server-side requester, steerable by untrusted input the agent reads mid-task. Mitigations include allow-listing destinations, blocking requests to private/link-local and metadata addresses by default, and validating any URL from untrusted input before the request is made, not after. OWASP’s SSRF Prevention Cheat Sheet covers these mitigations in detail; the private and link-local ranges most allow-lists must block are listed in IANA’s IPv4 Special-Purpose Address Registry.

SSRF is part of AdminKit’s AI agents & MCP cluster; the full AdminKit glossary lists other related terms.

Frequently asked questions

Is SSRF only a risk for traditional web servers?

No -- the risk has grown with AI agents and MCP tool integrations, because an agent that can fetch arbitrary URLs on a user's behalf is itself a server-side requester, steerable by untrusted input the agent reads mid-task.

What's a classic example of an SSRF attack?

A "fetch this image URL" feature pointed at an internal address like a cloud metadata endpoint, which gets the server to reach places that aren't reachable from the public internet at all.

How is SSRF typically mitigated?

Common mitigations include allow-listing permitted destinations, blocking requests to private/link-local and metadata addresses by default, and validating any URL from untrusted input before the request is made, not after.

Which address ranges does an SSRF defense need to block by default?

The private and link-local ranges most allow-lists must block are listed in IANA's IPv4 Special-Purpose Address Registry.