Security Digest · Vulnerability record
CVE-2025-49655: Keras framework vulnerable to deserialization of untrusted data
Deserialization of untrusted data can occur in versions of the Keras framework running versions 3.11.0 up to but not including 3.11.3, enabling a maliciously uploaded Keras file containing a TorchModuleWrapper class to run arbitrary code on an end user’s system when loaded despite safe mode being enabled.
CVSS 9.8 Critical EPSS 0.8% · p54
Should you act?
Critical and remotely reachable
No known exploitation yet, but it is network-reachable without credentials. This signal is computed from CISA KEV, EPSS and CVSS data; it is not an editorial verdict.
- Not in CISA KEV, but CISA’s SSVC assessment records a public proof-of-concept, so attack code exists.
- EPSS 0.8%, percentile 54: FIRST’s model estimates a 0.8% chance of exploitation activity in the next 30 days, higher than 54% of all scored CVEs (moderate). EPSS is a probability, not a verdict.
- Remotely reachable with no login and no user interaction. If the vulnerable service is exposed to the internet, prioritize it.
- Only applies if your dependency tree includes
keras(PyPI) >= 3.11.0, < 3.11.3, fixed in 3.11.3. Check your lockfile, including transitive dependencies.
What it is
Deserialization of untrusted data can occur in versions of the Keras framework running versions 3.11.0 up to but not including 3.11.3, enabling a maliciously uploaded Keras file containing a TorchModuleWrapper class to run arbitrary code on an end user’s system when loaded despite safe mode being enabled. The vulnerability can be triggered through both local and remote files.
From the GitHub advisory
Deserialization of untrusted data can occur in versions of the Keras framework running versions 3.11.0 up to but not including 3.11.3, enabling a maliciously uploaded Keras file containing a TorchModuleWrapper class to run arbitrary code on an end user’s system when loaded despite safe mode being enabled. The vulnerability can be triggered through both local and remote files.
Weakness type: CWE-502.
Product: keras (PyPI) · Assigned by: HiddenLayer · Published: 2025-10-17
Affected and fixed versions
| Ecosystem | Package | Vulnerable | First fixed |
|---|---|---|---|
| PyPI | keras | >= 3.11.0, < 3.11.3 | 3.11.3 |
| Vendor | Product | Affected versions (per CVE record) |
|---|---|---|
| Keras | Keras | 3.11.0 < 3.11.3 |
Severity scores
| Scored by | CVSS | Score | Vector |
|---|---|---|---|
| HiddenLayer | 3.1 | 9.8 Critical | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H |
| GitHub | 3.1 | 9.8 | CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H |
- Attack vector: network (remotely reachable)
- Attack complexity: low
- Privileges required: none (no login)
- User interaction: none
- Scope: unchanged
CVSS measures how bad a flaw is if exploited, not how likely exploitation is. Read it next to the exploitation data below.
Exploitation and urgency
- CISA KEV: not listed (no confirmed in-the-wild exploitation known to CISA).
- CISA SSVC assessment: exploitation poc, automatable yes, technical impact total.
- FIRST EPSS (model of 2026-10-21): 0.75% probability of exploitation activity in the next 30 days; percentile 54.
Timeline
- — CVE record last updated
- — CVE record published by HiddenLayer
- — GitHub-reviewed advisory GHSA-cvhh-q5g5-qprp published (critical)
Sources and references
Original sources first. Version and fix details should always be confirmed against the vendor or project advisory.
Vulnerability records
- GitHub Advisory GHSA-cvhh-q5g5-qprp github.com
- CVE.org record CVE-2025-49655 cve.org
- NVD entry CVE-2025-49655 nvd.nist.gov
Other references from the CVE record and advisory
- hiddenlayer.com hiddenlayer.com
- github.com github.com