PyTorch torch.nn.utils.rnn.pad_packed_sequence memory corruption (CVE-2025-2998)
A critical vulnerability in PyTorch version 2.6.0 affecting the function torch.nn.utils.rnn.pad_packed_sequence can lead to memory corruption. Exploitation requires local access. The vulnerability has been publicly disclosed, but no known exploits are reported in the wild. A patch is available to address this issue.
AI Analysis
Technical Summary
CVE-2025-2998 is a memory corruption vulnerability found in PyTorch 2.6.0 within the torch.nn.utils.rnn.pad_packed_sequence function. The flaw allows an attacker with local access to manipulate the function in a way that causes memory corruption. The vulnerability has been publicly disclosed, indicating potential risk of exploitation, although no active exploits have been observed. A patch exists for affected versions.
Potential Impact
Successful exploitation of this vulnerability can lead to memory corruption, which may cause application crashes or potentially allow arbitrary code execution depending on the context. However, exploitation requires local access to the affected system. No known active exploitation in the wild has been reported.
Mitigation Recommendations
A patch is available for this vulnerability. Users should upgrade PyTorch to a fixed version beyond 2.6.x to remediate the issue. Since local access is required for exploitation, restricting local user permissions can also reduce risk.
PyTorch torch.nn.utils.rnn.pad_packed_sequence memory corruption (CVE-2025-2998)
Description
A critical vulnerability in PyTorch version 2.6.0 affecting the function torch.nn.utils.rnn.pad_packed_sequence can lead to memory corruption. Exploitation requires local access. The vulnerability has been publicly disclosed, but no known exploits are reported in the wild. A patch is available to address this issue.
Affected software
Run on your own infrastructure? Check whether these packages are installed with threat-finder — our free open-source scanner.
AI-Powered Analysis
Machine-generated threat intelligence
Technical Analysis
CVE-2025-2998 is a memory corruption vulnerability found in PyTorch 2.6.0 within the torch.nn.utils.rnn.pad_packed_sequence function. The flaw allows an attacker with local access to manipulate the function in a way that causes memory corruption. The vulnerability has been publicly disclosed, indicating potential risk of exploitation, although no active exploits have been observed. A patch exists for affected versions.
Potential Impact
Successful exploitation of this vulnerability can lead to memory corruption, which may cause application crashes or potentially allow arbitrary code execution depending on the context. However, exploitation requires local access to the affected system. No known active exploitation in the wild has been reported.
Mitigation Recommendations
A patch is available for this vulnerability. Users should upgrade PyTorch to a fixed version beyond 2.6.x to remediate the issue. Since local access is required for exploitation, restricting local user permissions can also reduce risk.
Technical Details
- Gcve Source
- db.gcve.eu
- Osv Id
- BIT-pytorch-2025-2998
- Osv Schema Version
- 1.6.2
- Aliases
- ["CVE-2025-2998"]
- Ecosystems
- ["Bitnami"]
- Database Specific Severity
- Medium
Threat ID: 6aa005ceacd9273b49ab5e1a
Added to database: 09/08/2026, 12:55:42 UTC
Last enriched: 09/08/2026, 13:24:09 UTC
Last updated: 09/10/2026, 19:36:49 UTC
Views: 4
Community Reviews
0 reviewsCrowdsource mitigation strategies, share intel context, and vote on the most helpful responses. Sign in to add your voice and help keep defenders ahead.
Want to contribute mitigation steps or threat intel context? Sign in or create an account to join the community discussion.
Actions
Updates to AI analysis require Pro Console access. Upgrade inside Console → Billing.
Need more coverage?
Upgrade to Pro Console for AI refresh and higher limits.
For incident response and remediation, OffSeq services can help resolve threats faster.
Latest Threats
Check if your credentials are on the dark web
Instant breach scanning across billions of leaked records. Free tier available.