/

CVE-2026-68004

Severity:

CRITICAL RISK

(9.8)

Improper Access Control (CWE-284) in OSSRS SRS RTMP

CVE-2026-68004

Deferred

Vulnerability type

Category:

Improper Access Control

CWE:

CWE-284

Impact:

Remote Code Execution

Severity:

CRITICAL RISK

(9.8)

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

A critical access-control vulnerability in OSSRS SRS can allow a remote attacker to bypass RTMP publish authorization and execute arbitrary code under affected security configurations. The risk arises when RTMP listeners process publish requests through vulnerable vhost-level security checks.

Affected context

The issue affects SRS deployments using RTMP listeners and vhost-level security controls governed by the security.enabled setting. It becomes risky when publish authorization is relied upon to restrict who may publish and the security-checking path does not consistently enforce that boundary.

Exploit preconditions

  • Attack surface: Network-accessible SRS RTMP listener
  • Required access: Ability to connect to the RTMP service and submit a publish request
  • Trigger point: RTMP publish authorization and the associated vhost security check
  • Why exploitation is possible: An authorization-control weakness may allow an untrusted publisher to reach functionality intended only for authorized clients, potentially enabling command execution

Why this happens

The vulnerability is rooted in improper access control in the RTMP publish authorization flow. Security decisions involving vhost configuration and SrsSecurity::check() may fail to reliably enforce the intended authorization boundary before sensitive publish-related processing occurs. This maps to CWE-284, where access restrictions are missing, inconsistent, or ineffective.

Potential impact

  • An unauthenticated remote attacker may execute arbitrary code
  • The SRS process or hosting system could be compromised
  • Streaming services and published media could be disrupted or manipulated
  • Credentials, tokens, or other accessible sensitive data may be exposed
  • A compromised service could be used for further network intrusion

General mitigation guidance

Mitigation generally focuses on making authorization checks complete,

consistent, and fail-closed across every RTMP publish path.

Common defensive patterns include:

  • Enforcing authentication and authorization before accepting publish operations
  • Treating security.enabled and related vhost settings as security-sensitive policy
  • Denying publication by default when authorization state is missing or ambiguous
  • Restricting RTMP listener exposure through network segmentation and access controls
  • Running the service with the least privileges necessary for its role
  • Separating streaming processes from sensitive hosts and administrative networks
  • Limiting the process's filesystem, operating-system, and network permissions

How teams detect issues like this

Organizations can look for this class of issue through:

  • Static analysis of authorization branches around RTMP publish handlers and security checks
  • Code review for configuration-controlled bypasses, fail-open behavior, and checks performed after sensitive processing
  • Tests covering unauthorized, malformed, and boundary-condition publish requests across vhost configurations
  • Runtime monitoring for unexpected publish activity, authorization failures, and child-process creation by the streaming service
  • Network policy enforcement that limits which clients can reach RTMP listener ports
  • Automated security review tools that flag inconsistent access-control enforcement

Sources

SourceLink
github.comgithub.com/ossrs/srs/releases/tag/v5.0-r3
github.comgithub.com/xuwu-xuwu/CVE-2026-68004

Frequently asked

questions

Yes. The stated attack surface is the network-facing RTMP listener, although

successful exploitation depends on the deployment's exposure and the affected

authorization configuration.

Not necessarily. The vulnerability concerns RTMP publish authorization, so an

attacker may only need network access if the authorization control can be

bypassed or incorrectly enforced.

Ensure that every publish request is subject to consistent, fail-closed

authorization before sensitive processing, and restrict RTMP listener exposure

to trusted clients.

Use deny-by-default policies, test all authorization paths and configuration

combinations, apply least privilege to the service, and monitor unexpected

publishing and process activity.

Find out what's already
exploitable in your codebase.