CVE-2026-88047

CVE-2026-88047 is a stack buffer overflow (CWE-121) in the Classify::ReadNormProtos function of Tesseract OCR's legacy engine, where a whitespace-delimited token from the NORMPROTO component of a .traineddata file is read via std::istream::operator>>(char*) into a fixed 61-byte stack buffer without a stream width limit. Because the 100-byte line buffer can carry tokens of up to 99 characters, a token longer than 60 characters encountered during TessBaseAPI::Init writes up to 39 attacker-controlled bytes past the buffer. An attacker who can supply a malicious .traineddata file to an application that initializes Tesseract's legacy engine can cause stack corruption and denial of service, and potentially control-flow hijacking on affected standard-library implementations. All Tesseract versions 5.5.3 and earlier are affected on typical libstdc++ builds (as commonly used on Linux and Windows), while builds using Apple's libc++ C++20 bounded array overload are incidentally protected. There is no fixed release yet, no public proof-of-concept, and no known in-the-wild exploitation.

What to do: No fixed release exists yet, so monitor the tesseract-ocr GitHub repository and apply the upstream patch promptly when published. Until then, avoid loading .traineddata files from untrusted sources into applications that use Tesseract's legacy engine. Check your build toolchain: typical libstdc++ builds (most Linux/Windows) are affected, while builds using Apple libc++ with the C++20 bounded array overload are incidentally protected.