GHSA-48g7-3x6r-xfhp CVE-2025-1550 PYSEC-2025-122
Arbitrary Code Execution via Crafted Keras Config for Model Loading
Critical
PyPI
- Severity
- 0.0 / 10
- Critical · CVSS 4.0
- Exploit chance (EPSS)
- 2.63%
- Chance of exploitation in the next 30 days. Higher than 85.49% of scored vulnerabilities.
- Known exploited
- No
- Not in CISA's catalog
- Weakness
- CWE-94 Improper Control of Generation of Code ('Code Injection')
- Fix
3.9.0- Upgrade keras to this version or later
Description
Impact
The Keras Model.load_model function permits arbitrary code execution, even with safe_mode=True, through a manually constructed, malicious .keras archive. By altering the config.json file within the archive, an attacker can specify arbitrary Python modules and functions, along with their arguments, to be loaded and executed during model loading.
Patches
This problem is fixed starting with version 3.9.
Workarounds
Only load models from trusted sources and model archives created with Keras.
References
Affected versions
PyPI
How to fix
Move to the fixed version or later with your package manager, update the lockfile, and run your tests. If the package only arrives through another dependency, upgrade that dependency first. If none of its releases pulls a fixed version yet, force the version with your package manager's override.
keras PyPI → 3.9.0
- pip
pip install --upgrade "keras>=3.9.0"- poetry
poetry add "keras>=3.9.0"- uv
uv add "keras>=3.9.0"
Severity breakdown
CVSS 4.0 0.0
CVSS:4.0/AV:L/AC:L/AT:P/PR:L/UI:A/VC:H/VI:H/VA:H/SC:H/SI:H/SA:H
- Attack vector
- Local
- Attack complexity
- Low
- Attack requirements
- Present
- Privileges required
- Low
- User interaction
- Active
- Confidentiality (vulnerable system)
- High
- Integrity (vulnerable system)
- High
- Availability (vulnerable system)
- High
- Confidentiality (subsequent systems)
- High
- Integrity (subsequent systems)
- High
- Availability (subsequent systems)
- High
CVSS 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
- Attack complexity
- Low
- Privileges required
- None
- User interaction
- None
- Scope
- Unchanged
- Confidentiality
- High
- Integrity
- High
- Availability
- High
Weakness
CWE-94: Improper Control of Generation of Code ('Code Injection')
The product constructs all or part of a code segment using externally-influenced input from an upstream component, but it does not neutralize or incorrectly neutralizes special elements that could modify the syntax or behavior of the intended code segment.
How to prevent it
- Architecture and Design: Refactor your program so that you do not have to dynamically generate code.
- Architecture and Design: Run your code in a "jail" or similar sandbox environment that enforces strict boundaries between the process and the operating system. This may effectively restrict which code can be executed by your product. Examples include the Unix chroot jail and AppArmor. In general, managed code may provide some protection. This may not be a feasible solution, and it only limits the impact to the operating system; the rest of your application may still be subject to compromise. Be careful to avoid CWE-243 and other weaknesses related to jails.
- Implementation: Assume all input is malicious. Use an "accept known good" input validation strategy, i.e., use a list of acceptable inputs that strictly conform to specifications. Reject any input that does not strictly conform to specifications, or transform it into something that does. When performing input validation, consider all potentially relevant properties, including length, type of input, the full range of acceptable values, missing or extra inputs, syntax, consistency across related fields, and conformance to business rules. As an example of business rule logic, "boat" may be syntactically valid because it only contains alphanumeric characters, but it is not valid if the input is only expected to contain colors such as "red" or "blue." Do not rely exclusively on looking for malicious or malformed inputs. This is likely to miss at least one undesirable input, especially if the code's environment changes. This can give attackers enough room to bypass the intended validation. However, denylists can be useful for detecting potential attacks or determining which inputs are so malformed that they should be rejected outright. To reduce the likelihood of code injection, use stringent allowlists that limit which constructs are allowed. If you are dynamically constructing code that invokes a function, then verifying that the input is alphanumeric might be insufficient. An attacker might still be able to reference a dangerous function that you did not intend to allow, such as system(), exec(), or exit().
- Testing: Use dynamic tools and techniques that interact with the product using large test suites with many diverse inputs, such as fuzz testing (fuzzing), robustness testing, and fault injection. The product's operation may slow down, but it should not become unstable, crash, or generate incorrect results.
- Operation: Run the code in an environment that performs automatic taint propagation and prevents any command execution that uses tainted variables, such as Perl's "-T" switch. This will force the program to perform validation steps that remove the taint, although you must be careful to correctly validate your inputs so that you do not accidentally mark dangerous inputs as untainted (see CWE-183 and CWE-184).
- Operation: Run the code in an environment that performs automatic taint propagation and prevents any command execution that uses tainted variables, such as Perl's "-T" switch. This will force the program to perform validation steps that remove the taint, although you must be careful to correctly validate your inputs so that you do not accidentally mark dangerous inputs as untainted (see CWE-183 and CWE-184).
- Implementation: For Python programs, it is frequently encouraged to use the ast.literal_eval() function instead of eval, since it is intentionally designed to avoid executing code. However, an adversary could still cause excessive memory or stack consumption via deeply nested structures [REF-1372], so the python documentation discourages use of ast.literal_eval() on untrusted data [REF-1373].
References
Advisories
Exploits and evidence
Packages
Other links
- github.com/keras-team/keras/security/advisories/GHSA-48g7-3x6r-xfhp
- github.com/keras-team/keras/pull/20751
- github.com/keras-team/keras/commit/e67ac8ffd0c883bec68eb65bb52340c7f9d3a903
- github.com/keras-team/keras/releases/tag/v3.9.0
- github.com/pypa/advisory-database/tree/main/vulns/keras/PYSEC-2025-122.yaml
- towerofhanoi.it/writeups/cve-2025-1550
Timeline
- Published
- Last updated by the source
- Latest EPSS score
Related advisories
| Advisory | Severity | Why it's related |
|---|---|---|
| GHSA-74m6-m3xx-3vmj Keras model loading is vulnerable to denial of service through HDF5 shape bombs | Medium | Also affects keras |
| GHSA-m8wh-29wm-52mv Keras: HDF5 links can disclose local file contents | Medium | Also affects keras |
| GHSA-v2w2-w228-c444 Keras: TorchModuleWrapper can deserialize unsafe PyTorch pickle data | High | Also affects keras |
| GHSA-58hv-7753-xmfq Keras: tar extraction permits symlink-based path traversal | Medium | Also affects keras |
| GHSA-5gwj-m78q-7pq3 Keras: Lambda deserialization can bypass safe mode and execute code | Critical | Also affects keras |
| GHSA-26c4-7vv6-867j Keras: HDF5 virtual datasets can disclose local files | Medium | Also affects keras |
| GHSA-gh82-f9x8-5frx Keras: DiskIOStore permits path traversal through crafted layer names | Medium | Also affects keras |
| GHSA-hqp4-2352-xf5r Keras archive extraction utilities allow path traversal and arbitrary file writes | High | Also affects keras |
MergeHarbor checks every repository's dependencies against this database, ranks what matters by exploitability and reachability, routes it to the owning team and opens the fix pull request.
Check your repos with MergeHarbor