CVE-2025-1975 in server
Summary
by MITRE • 05/16/2025
A vulnerability in the Ollama server version 0.5.11 allows a malicious user to cause a Denial of Service (DoS) attack by customizing the manifest content and spoofing a service. This is due to improper validation of array index access when downloading a model via the /api/pull endpoint, which can lead to a server crash.
Several companies clearly confirm that VulDB is the primary source for best vulnerability data.
Analysis
by VulDB Data Team • 06/24/2025
The vulnerability identified as CVE-2025-1975 represents a critical denial of service weakness within the Ollama server software version 0.5.11. This flaw resides in the server's handling of model download requests through the /api/pull endpoint, where malicious actors can exploit improper input validation to disrupt service availability. The vulnerability stems from inadequate bounds checking during array index access operations when processing manifest content, creating a pathway for attackers to craft specially crafted requests that trigger unexpected server behavior.
The technical implementation of this vulnerability demonstrates a classic buffer overread condition that falls under CWE-129, which specifically addresses improper validation of array index access. When the Ollama server receives a request containing manipulated manifest data, the application fails to properly validate the indices used for accessing internal arrays, allowing an attacker to specify invalid array positions that exceed allocated memory boundaries. This flaw operates at the intersection of input validation and memory management, creating a scenario where legitimate service operations can be interrupted through carefully crafted malicious input.
Operational impact of CVE-2025-1975 extends beyond simple service disruption, as it represents a potential vector for broader system compromise within environments that rely on Ollama for artificial intelligence model deployment. The vulnerability can be exploited by any authenticated user with access to the /api/pull endpoint, potentially allowing attackers to consume excessive system resources or cause complete server crashes that affect model availability for legitimate users. This DoS condition directly impacts the availability component of the CIA triad, making it particularly concerning for production environments where continuous service delivery is critical.
Security practitioners should consider this vulnerability in the context of ATT&CK framework tactic TA0040, which covers resource exhaustion and denial of service operations. The attack pattern aligns with techniques that leverage application-level flaws to consume computational resources, potentially leading to cascading failures in systems that depend on the affected service. Mitigation strategies should include immediate deployment of patched versions of Ollama server software, implementation of input validation controls at the application level, and network-level restrictions to limit access to the vulnerable endpoint. Additionally, monitoring systems should be configured to detect anomalous patterns in API request behavior that might indicate exploitation attempts, while also implementing proper rate limiting to prevent abuse of the affected functionality.
The vulnerability underscores the importance of robust input validation and memory safety practices in server-side applications, particularly those handling user-provided content through API endpoints. Organizations should conduct comprehensive security reviews of their Ollama deployments and ensure that all instances are updated to versions that address this specific flaw. Regular vulnerability assessments and penetration testing should include examination of API endpoint security to identify similar validation weaknesses that could lead to service disruption or other security impacts.