Summary
NewsBlur contains a server-side request forgery flaw in the add_url endpoint that lets authenticated users cause the server to send arbitrary HTTP requests to internal resources.
When private IP ranges and metadata endpoints are not filtered, attackers can pivot through the application server to reach localhost services and cloud metadata APIs, exposing internal systems and sensitive configuration data.
Affected context
This vulnerability appears in NewsBlur deployments where the add_url endpoint accepts user-supplied URLs and forwards requests from the backend server without sufficiently restricting target addresses.
It is most relevant in environments where the application server has network visibility into localhost services, internal IP ranges, or cloud provider metadata endpoints, and where authenticated users can invoke the endpoint with arbitrary URLs.
Exploit preconditions
- Attack surface: HTTP add_url endpoint exposed to authenticated users
- Required access: Ability to log in to NewsBlur and submit or manipulate feed URLs
- Trigger point: Server-side fetching of user-provided URLs without filtering private or sensitive address ranges
- Why exploitation is possible: The backend trusts user-supplied URLs and performs outbound requests on their behalf, failing to enforce address allowlists, blocklists for private IP ranges, or protections for well-known cloud metadata endpoints.
Why this happens
The underlying issue is a classic server-side request forgery pattern where the application acts as a proxy for user-controlled URLs.
Instead of validating that requested URLs point only to intended external hosts, the implementation allows direct use of arbitrary URLs, including private IP ranges, localhost, and special infrastructure endpoints.
This maps to CWE-918: Server-Side Request Forgery (SSRF), where server-side network capabilities are exposed to an attacker via unvalidated URL or hostname parameters.
Potential impact
- Access to localhost-only administrative interfaces or monitoring endpoints
- Internal network scanning to discover services and open ports
- Retrieval of sensitive configuration or credentials from cloud metadata services
- Exposure of internal API responses and database-backed services not meant to be public
- Assistance in lateral movement or further compromise using information gathered from internal endpoints
General mitigation guidance
Effective mitigation focuses on neutralizing the application's ability to reach unintended network targets on behalf of users.
Common patterns include:
- Strictly validating user-supplied URLs against allowlists of permitted domains or schemes
- Blocking private IP ranges, localhost, and link-local addresses at the application and network layers
- Explicitly denying requests to known cloud metadata endpoints and other sensitive infrastructure URLs
- Decoupling user input from direct network calls by using intermediary services that enforce policy
- Applying least-privilege network controls so application servers cannot reach broader internal networks than necessary
How teams detect issues like this
Detection often combines code-level review with runtime monitoring:
- Static analysis rules that flag direct HTTP client usage with user-controlled URLs or hosts
- Code review checks for missing URL validation, domain constraints, and IP range filtering in request-building logic
- Security policies requiring allowlists for outbound HTTP calls from server-side components
- Logging and alerting on suspicious outbound requests to private IP ranges or metadata endpoints
- Use of automated security testing tools that can probe endpoints for SSRF behavior as part of CI/CD pipelines

