CVE-2026-35459: pyLoad Download Manager SSRF Vulnerability

CVE-2026-35459 Overview

CVE-2026-35459 is a Server-Side Request Forgery (SSRF) vulnerability in pyLoad, a free and open-source download manager written in Python. The vulnerability exists in versions 0.5.0b3.dev96 and earlier where the SSRF fix implemented for CVE-2026-33992 can be bypassed through HTTP redirects. While IP validation was added to BaseDownloader.download() to check the hostname of initial download URLs, the underlying pycurl library is configured with FOLLOWLOCATION=1 and MAXREDIRS=10, causing it to automatically follow HTTP redirects without validating redirect targets against the SSRF filter.

Critical Impact
An authenticated user with ADD permission can exploit this vulnerability to perform server-side requests to internal network addresses by submitting a URL that redirects to an internal address, potentially accessing sensitive internal services or data.

Affected Products

pyLoad version 0.5.0b3.dev96 and earlier

Discovery Timeline

2026-04-06 - CVE CVE-2026-35459 published to NVD
2026-04-07 - Last updated in NVD database

Technical Details for CVE-2026-35459

Vulnerability Analysis

This vulnerability represents an incomplete fix for a previous SSRF vulnerability (CVE-2026-33992). The original patch added IP validation to prevent users from submitting URLs pointing to internal network addresses. However, the fix only validates the initial URL and fails to account for HTTP redirects that pycurl automatically follows.

The attack requires an authenticated user with ADD permission, but once that threshold is met, the attacker can completely bypass the SSRF protections. The impact is significant as it allows access to internal services that should not be reachable from the download manager, potentially exposing sensitive data, internal APIs, or cloud metadata endpoints.

Root Cause

The root cause is the incomplete validation of URL targets in the HTTP download chain. The BaseDownloader.download() function validates the hostname of the initial URL against the SSRF filter, but pycurl is configured to automatically follow up to 10 redirects (FOLLOWLOCATION=1, MAXREDIRS=10). These redirect targets are never validated, allowing an attacker to craft an external URL that redirects to internal addresses like 169.254.169.254 (cloud metadata), localhost, or other private IP ranges.

Attack Vector

The attack is network-based and requires low complexity to execute. An authenticated attacker with ADD permission can:

Host a malicious URL on an external server that returns an HTTP redirect (301/302) to an internal address
Submit this external URL through pyLoad's download interface
pyLoad validates the external hostname (which passes the SSRF filter)
pycurl follows the redirect to the internal address without validation
The attacker receives the response from the internal service

The security patch addresses this by adding an allow_private_ip flag to the HTTP chunk handling:

python

 
        self.code = 0  #: last http code, set by parent

+        self.allow_private_ip = False
        self.aborted = False  # indicates that the chunk aborted gracefully

        self.c = pycurl.Curl()

Source: GitHub Commit 33c55da

The patch also adds URL parsing capabilities to properly validate redirect targets:

python

import os
import time
+import urllib
from logging import getLogger

import pycurl
from pyload import APPID

-from ..exceptions import Abort
+from ..exceptions import Abort, Fail
from .http_chunk import ChunkInfo, HTTPChunk
from .http_request import BadHeader

Source: GitHub Commit 33c55da

Detection Methods for CVE-2026-35459

Indicators of Compromise

Unexpected outbound connections from the pyLoad server to external URLs followed by internal network requests
Download requests containing URLs that result in HTTP 3xx redirects to private IP ranges
Unusual access patterns to cloud metadata endpoints (e.g., 169.254.169.254) or internal services from the pyLoad process

Detection Strategies

Monitor pyLoad server logs for download attempts that trigger redirects, particularly those with external initial URLs
Implement network monitoring to detect connections from pyLoad to internal IP ranges that weren't explicitly requested
Deploy egress filtering alerts for pyLoad server processes attempting to access sensitive internal endpoints

Monitoring Recommendations

Enable detailed logging for all download requests including redirect chains
Set up alerts for pyLoad processes connecting to RFC1918 private address spaces or link-local addresses
Monitor for unusual patterns in ADD permission usage by authenticated users

How to Mitigate CVE-2026-35459

Immediate Actions Required

Update pyLoad to a version containing commit 33c55da084320430edfd941b60e3da0eb1be9443 or later
Review user accounts with ADD permission and restrict to trusted users only
Implement network segmentation to limit pyLoad server access to internal resources

Patch Information

The vulnerability has been addressed in the pyLoad repository. Users should apply the security patch available at the GitHub commit 33c55da. For detailed information about the vulnerability, refer to the GitHub Security Advisory GHSA-7gvf-3w72-p2pg.

Workarounds

Restrict ADD permission to only highly trusted users until the patch can be applied
Deploy a reverse proxy or firewall rules that prevent the pyLoad server from making connections to internal network ranges
Disable or limit HTTP redirect following at the network level if operationally feasible

bash

# Example: Block pyLoad server from accessing internal networks via iptables
# Replace pyload_user with the actual user running pyLoad
iptables -A OUTPUT -m owner --uid-owner pyload_user -d 10.0.0.0/8 -j DROP
iptables -A OUTPUT -m owner --uid-owner pyload_user -d 172.16.0.0/12 -j DROP
iptables -A OUTPUT -m owner --uid-owner pyload_user -d 192.168.0.0/16 -j DROP
iptables -A OUTPUT -m owner --uid-owner pyload_user -d 169.254.0.0/16 -j DROP