When an attacker provides a crafted file where these size fields are set to zero, the library's extent-parsing loop fails to advance the buffer offset while simultaneously pushing new objects onto the item.extents array for every iteration defined by the itemCount and extentCount headers. This leads to unbounded memory consumption. A malicious file as small as 652 bytes can trigger 400MB of heap growth, while files in the kilobyte range result in an out-of-memory (OOM) crash, effectively terminating the Node.js process. This poses a significant risk to any service or application utilizing this library for processing user-provided imagery.

Attack Chain

1. Attacker crafts a malicious HEIC/AVIF file container following the ISO-BMFF specification.
2. The iloc box is populated with an itemCount and extentCount set to the maximum allowed values (up to 65535 each).
3. The offsetSize, lengthSize, and baseOffsetSize fields in the iloc header are set to zero, signaling that these values are absent in the byte stream.
4. The victim application receives the malicious image and passes the buffer to ExifReader.load().
5. The library's getItems() function initiates a nested loop, iterating based on the attacker-controlled itemCount and extentCount.
6. Within the loop, the parser performs unconditional object allocations for each extent without advancing the buffer offset.
7. The heap usage of the Node.js process expands rapidly until it reaches the memory limit defined by the environment.
8. The process crashes with a JavaScript heap out-of-memory error, resulting in a successful denial-of-service against the host application.

Impact

Successful exploitation results in an immediate denial-of-service for the application process. This vulnerability affects any service (such as web servers, backend media processors, or mobile app backends) that accepts and processes HEIC or AVIF images using ExifReader 4.41.0. Given that a crash-inducing payload is less than 2KB, this is highly effective against internet-facing services, potentially causing significant downtime or resource contention across shared infrastructure.

Recommendation

1. Upgrade to a patched version of ExifReader once available.
2. Implement a per-item or global extent allocation limit in src/image-header-iso-bmff-iloc.js to bound the memory growth.
3. Apply an input validation gate to skip the extent parsing loop if offsetSize, lengthSize, and baseOffsetSize are all zero, as per the suggested fix in the advisory.
4. Implement memory resource monitoring and limits for Node.js worker processes that handle image parsing to mitigate potential OOM crashes.