Description

openssl_encrypt (pip: openssl-encrypt) versions 1.4.8 and earlier fail to validate the 36-bit STREAMINFO total_samples field of FLAC files before using it to size an allocation (np.random.randint(size=(total_samples, channels))). A ~50-byte crafted FLAC file declaring ~100 million samples causes a multi-gigabyte memory allocation, leading to out-of-memory denial of service during 'decrypt --stego-extract'. The issue is fixed in 1.4.9; both the 1.4.x and 1.5.x lines are affected.

Executive Summary

Technical Details

Mitigations

References

Additional Metadata

{
  "score": "7.5",
  "pubDate": "2026-08-27T17:20:59.140Z",
  "pubdate": "2026-08-27T17:20:59.140Z",
  "executiveSummary": "The openssl-encrypt Python package contains an improper input validation vulnerability within its FLAC file processing logic, specifically concerning the STREAMINFO header. The vulnerability exists because the library uses the 36-bit 'total_samples' field from the FLAC metadata to determine the size of a memory allocation without performing bounds checking or sanitization. By crafting a malicious FLAC file with a significantly inflated 'total_samples' value, an attacker can trigger a massive, uncontrolled memory allocation upon invoking 'decrypt --stego-extract'. This results in an out-of-memory (OOM) condition, causing the host application to crash or become unresponsive, effectively facilitating a Denial of Service (DoS) attack. The vulnerability affects both the 1.4.x and 1.5.x branches, with a fix introduced in version 1.4.9. Successful exploitation requires an attacker to convince a user or automated system to process a maliciously crafted FLAC file. Given that the allocation is based on user-supplied metadata rather than actual stream size, the threat is significant for any service parsing untrusted audio files using this library.",
  "technicalDetails": "The vulnerability resides in the core decoding logic of openssl-encrypt where FLAC file metadata is parsed. Specifically, the library extracts the 'total_samples' field from the STREAMINFO block. This field is defined as a 36-bit value in the FLAC specification, allowing for theoretically large sample counts. The application performs a direct mapping of this value to the 'size' parameter of the 'np.random.randint' function (NumPy) used for allocating buffer space for decryption or steganographic extraction.\nThe root cause is a lack of validation and sanitization of the 'total_samples' value. The library assumes that the metadata in the FLAC header is trustworthy and reflects the actual intended size of the audio content. Because there is no check to ensure the total_samples value corresponds to the actual data payload or fits within reasonable hardware constraints, an attacker can define a nominal value (e.g., 100 million samples) within a file that is only approximately 50 bytes in total size.\nThe attack flow follows a predictable pattern: 1) The attacker constructs a malicious FLAC file containing a fabricated STREAMINFO block with an exceptionally large 'total_samples' integer. 2) The target environment is coerced into running the 'decrypt --stego-extract' routine on this file. 3) The library parses the 36-bit total_samples value and feeds it into the memory allocation routine. 4) The system attempts to allocate a multi-gigabyte buffer based on the attacker's value, which leads to immediate system resource depletion. 5) If the system cannot fulfill the allocation request, the process terminates due to OOM conditions, resulting in an application crash.\nThe impact of this vulnerability is a high-availability risk. Because the memory allocation is handled synchronously before file validation is completed, the crash is immediate upon header processing. There are no authentication requirements to trigger this vulnerability, as the processing routine is invoked directly by the provided input. Exposure is dependent on the context in which openssl-encrypt is utilized; if it is deployed as part of a web service or a file conversion pipeline, the potential for automated exploitation is high. Affected versions include all instances within the 1.4.x and 1.5.x lines prior to the 1.4.9 release."
}