{"description":"Trending threats, MITRE ATT\u0026CK coverage, and detection metadata. Fed continuously.","favicon":"https://feed.craftedsignal.io/favicon-32x32.png","feed_url":"https://feed.craftedsignal.io/products/doctr--1.1.0/feed.json","home_page_url":"https://feed.craftedsignal.io/","icon":"https://feed.craftedsignal.io/apple-touch-icon.png","items":[{"_cs_actors":[],"_cs_cpes":["cpe:2.3:a:mindee:doctr:*:*:*:*:*:*:*:*"],"_cs_cves":[{"cvss":7.5,"id":"CVE-2026-112304"}],"_cs_exploited":false,"_cs_has_poc":false,"_cs_poc_references":[],"_cs_products":["docTR (\u003c= 1.1.0)"],"_cs_severities":["high"],"_cs_tags":["deserialization","remote-code-execution","python","supply-chain"],"_cs_type":"advisory","_cs_vendors":["Mindee"],"content_html":"\u003cp\u003eMindee docTR (Document Text Recognition) versions through 1.1.0 contain a critical unsafe deserialization vulnerability identified as CVE-2026-112304. The vulnerability exists within the \u003ccode\u003eload_pretrained_params()\u003c/code\u003e function, which invokes \u003ccode\u003etorch.load\u003c/code\u003e to process model checkpoints. By default, the library fails to set the \u003ccode\u003eweights_only=True\u003c/code\u003e parameter, which is a required security configuration for loading untrusted PyTorch models.\u003c/p\u003e\n\u003cp\u003eAn attacker who can provide a malicious checkpoint file - either via local storage or a remote URL - can embed a malicious pickle payload within the file. When \u003ccode\u003eload_pretrained_params()\u003c/code\u003e processes this file, the underlying PyTorch mechanism (on versions prior to 2.6) will deserialize the payload, resulting in arbitrary Python code execution under the security context of the user or service running the docTR library. This vulnerability is significant for environments where document processing pipelines ingest user-supplied model files or rely on potentially compromised remote storage.\u003c/p\u003e\n\u003ch2 id=\"impact\"\u003eImpact\u003c/h2\u003e\n\u003cp\u003eSuccessful exploitation of this vulnerability allows an attacker to achieve arbitrary code execution on the system running the docTR library. This could lead to a full compromise of the application server, data exfiltration from the processing environment, or lateral movement within the network. The scope affects any application utilizing docTR version 1.1.0 or earlier for OCR tasks.\u003c/p\u003e\n\u003ch2 id=\"recommendation\"\u003eRecommendation\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eUpdate docTR to a version that enforces \u003ccode\u003eweights_only=True\u003c/code\u003e in \u003ccode\u003eload_pretrained_params()\u003c/code\u003e.\u003c/li\u003e\n\u003cli\u003eIf patching is not immediately feasible, audit the code to ensure \u003ccode\u003eload_pretrained_params()\u003c/code\u003e is only ever called with trusted, verified local model files.\u003c/li\u003e\n\u003cli\u003eVerify the PyTorch version in the environment and ensure it is updated to 2.6 or later to benefit from improved default deserialization safety.\u003c/li\u003e\n\u003cli\u003eImplement file integrity monitoring on any directory where model checkpoints are stored to detect unauthorized modifications.\u003c/li\u003e\n\u003c/ul\u003e\n","date_modified":"2026-10-19T18:08:30Z","date_published":"2026-10-19T18:08:30Z","id":"https://feed.craftedsignal.io/briefs/2026-10-doctr-deserialization/","summary":"docTR versions through 1.1.0 are vulnerable to arbitrary code execution via unsafe deserialization of pickle payloads within the load_pretrained_params() function.","title":"Unsafe Deserialization Vulnerability in docTR","url":"https://feed.craftedsignal.io/briefs/2026-10-doctr-deserialization/"}],"language":"en","title":"CraftedSignal Threat Feed - DocTR (\u003c= 1.1.0)","version":"https://jsonfeed.org/version/1.1"}