Summary
An authorization bypass in AVideo allows an unauthenticated attacker to use a known stream-key value to publish to another user's RTMP stream. The issue becomes risky when affected live-stream endpoints trust a session cookie value over the request's intended authorization parameter.
Affected context
The vulnerability occurs in AVideo live-stream publishing functionality, including saveLive.php and related endpoints, when request authorization relies on the key parameter. Deployments may be exposed when the application accepts a predictable cookie value that overrides the request parameter without verifying ownership of the targeted stream.
Exploit preconditions
- Attack surface: Publicly reachable live-stream publishing endpoints
- Required access: No valid user authentication may be required when the affected authorization path accepts the known cookie value
- Trigger point: Processing requests to
saveLive.phpor related stream-management endpoints - Why exploitation is possible: A session cookie named
keycan override$_REQUEST['key'], allowing the endpoint to treat a constant value as sufficient authorization instead of validating stream ownership
Why this happens
The underlying bug is an authorization decision based on attacker-controllable request state. When a cookie value takes precedence over the intended request parameter and is not bound to an authenticated identity or specific stream owner, the application can accept a fixed value as authorization.
This behavior maps to CWE-287: Improper Authentication, because the application fails to reliably establish the identity or authorization context associated with the requested live stream before permitting a privileged action.
Potential impact
- Attackers may publish unauthorized content to another user's RTMP stream
- Live broadcasts could be hijacked, disrupted, or replaced
- Viewers may receive misleading or malicious broadcast content
- Stream owners and platform operators could experience reputational or operational harm
- Related endpoints could permit additional unauthorized stream-management actions if they share the same authorization logic
General mitigation guidance
Mitigation generally focuses on ensuring that stream authorization is derived from a trusted, authenticated identity rather than from arbitrary request or cookie values.
Common defensive patterns include:
- Avoiding precedence rules that allow cookies or generic request parameters to override security-sensitive inputs
- Binding stream keys and publishing permissions to authenticated users and specific stream resources
- Enforcing server-side ownership and capability checks for every live-stream operation
- Using unpredictable, scoped, and rotatable stream credentials
- Applying least-privilege permissions to streaming accounts and service identities
- Rejecting ambiguous requests when the same authorization value appears in multiple input locations
How teams detect issues like this
Organizations can look for this class of issue through:
- Static analysis rules identifying authorization decisions based on
$_REQUEST, cookies, or other mixed input sources - Code review checks for security-sensitive parameters that can be overridden by alternate request locations
- Tests verifying that unauthenticated users cannot publish to existing streams or operate on resources they do not own
- Runtime monitoring for unexpected publishing activity, repeated use of identical stream keys, and publishing attempts across multiple users' streams
- Policy enforcement requiring explicit identity, ownership, and capability checks before stream operations
Automated security review tools may help identify unsafe input precedence and missing authorization checks, but endpoint-level authorization testing remains important.

