/

CVE-2026-42313

Severity:

HIGH RISK

(8.3)

Improper Authorization in pyLoad Core API

CVE-2026-42313

Analyzed

Vulnerability type

Category:

Improper Authorization

CWE:

CWE-441, CWE-863, CWE-918

Impact:

Outbound Traffic Hijacking and Data Interception

Severity:

HIGH RISK

(8.3)

CVSS version:

3.1

Catch this CVE while code is being written

Detect vulnerable code directly in pull requests, so CVEs are fixed before merge.

Summary

pyLoad's set_config_value() API exposes proxy routing options to non-admin users due to an incomplete allowlist for security-sensitive settings.

Any authenticated user with SETTINGS permission can enable and re-point pyLoad's outbound proxy to a host they control, allowing transparent interception and manipulation of all outbound HTTP traffic from the application.

Affected context

This vulnerability affects pyLoad deployments that:

  • Expose the core API with the set_config_value() method enabled
  • Use role-based permissions where non-admin users can hold the SETTINGS permission
  • Rely on the proxy configuration for outbound HTTP/S traffic (downloads, captcha retrieval, update checks, and plugin calls)

In such environments, security-sensitive proxy options are only partially restricted to administrators, leaving key routing parameters writable by lower-privileged users.

Exploit preconditions

  • Attack surface: pyLoad core API, specifically the set_config_value() method for configuration changes.
  • Required access: Valid authenticated user account with the SETTINGS permission, but without full administrative privileges.
  • Trigger point: Updating configuration keys for the "proxy" section (e.g., "enabled", "host", "port", "type") via API calls or any UI path that uses this API.
  • Why exploitation is possible: The ADMIN_ONLY_CORE_OPTIONS allowlist only protects proxy credentials ("username", "password"), but omits other critical routing fields ("enabled", "host", "port", "type"), enabling lower-privileged users to redirect traffic to attacker-controlled infrastructure.

Why this happens

The underlying issue is an improper authorization design around configuration management.

pyLoad implements a hand-maintained allowlist, ADMIN_ONLY_CORE_OPTIONS, to mark which configuration keys require administrator privileges.

While proxy credentials are correctly classified as admin-only, the associated proxy routing options are not, resulting in inconsistent privilege checks within the same logical security boundary.

This maps to:

  • CWE-441: Unintended Proxy or Intermediary ('Confused Deputy') because pyLoad can be coerced into acting as a traffic forwarder through an attacker-controlled proxy.
  • CWE-863: Incorrect Authorization since the system authorizes sensitive configuration changes based on an incomplete allowlist.
  • CWE-918: Server-Side Request Forgery (SSRF) behavior, as an attacker can redirect server-originated HTTP requests to arbitrary endpoints.

Conceptually, the bug is a fragile, manually curated authorization rule set for configuration keys, where not all security-relevant options are consistently protected.

Potential impact

  • An attacker-controlled user account can redirect all outbound downloads through a malicious proxy.
  • Transparent interception and modification of downloaded content, potentially enabling malware distribution or content tampering.
  • Interception of captcha traffic, plugin HTTP communication, and update checks, enabling downgrade or update poisoning scenarios.
  • Use of pyLoad as an SSRF pivot to reach internal or restricted network resources via attacker-chosen proxy endpoints.
  • Loss of integrity and confidentiality of data fetched by pyLoad on behalf of legitimate users.

General mitigation guidance

Effective mitigation centers on hardening authorization boundaries around configuration changes and treating routing options as security-sensitive:

  • Classify all settings that control where the application sends traffic (hostnames, ports, proxy selectors, protocols) as privileged, not just credentials.
  • Replace ad-hoc, hand-maintained allowlists with structured policy mechanisms, such as role-based access control (RBAC) rules or configuration schemas that explicitly mark security-critical fields.
  • Introduce a centralized authorization layer for configuration writes, rather than scattering checks across individual options.
  • Consider limiting or disabling runtime changes to proxy and network routing settings in multi-user or untrusted-tenant deployments.
  • Apply least-privilege principles to user roles so that SETTINGS-like permissions are not granted more broadly than necessary.

How teams detect issues like this

Organizations can detect and monitor similar authorization issues using a combination of static and operational techniques:

  • Static analysis: Search for configuration APIs that gate access via hard-coded allowlists or option-name checks, especially where some keys are tagged as "admin-only".
  • Code review: Flag patterns where only a subset of related options (e.g., some but not all proxy settings) are protected, and ensure that routing and network endpoints are reviewed as security-sensitive.
  • Policy enforcement: Establish guidelines requiring that any configuration impacting authentication, encryption, routing, or external communication go through a centralized authorization policy.
  • Runtime monitoring: Log configuration changes to network and proxy options and alert on changes made by non-admin or low-privileged accounts.
  • Automated tools: Integrate SAST or configuration-policy linters into CI pipelines to highlight authorization checks based on fragile string allowlists.

Sources

SourceLink
github.comgithub.com/pyload/pyload/security/advisories/GHSA-pg67-9wjv-mr85

Frequently asked

questions

Exploitation is limited to environments where non-admin users have valid accounts with the SETTINGS permission, but in such setups, any such user can redirect and intercept all outbound traffic from pyLoad.

No. The core problem is that proxy routing options are not restricted to administrators, so an authenticated user with SETTINGS-level access is sufficient to exploit the flaw.

Treat all configuration settings that affect outbound communication—hosts, ports, proxies, protocols—as security-sensitive and protect them with strong authorization policies, preferably using centralized RBAC or policy engines rather than manual allowlists.

Avoid scattered, hand-maintained lists of "sensitive" options, define explicit configuration schemas with sensitivity classifications, enforce consistent authorization checks for all keys in a security domain (like networking), and validate these rules with security-focused code reviews and automated checks.

Find out what's already
exploitable in your codebase.