{"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/tags/cve-2026-48705/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:monai:monailabel:*:*:*:*:*:*:*:*","cpe:2.3:a:pydicom:pydicom:*:*:*:*:*:*:*:*"],"_cs_cves":[{"id":"CVE-2026-48705"}],"_cs_exploited":false,"_cs_has_poc":false,"_cs_poc_references":[],"_cs_products":["MONAILabel (\u003c 1.6.1)","MONAI Core (\u003c 1.6.1)","pydicom (0.9.7 - 3.0.2)"],"_cs_severities":["high"],"_cs_tags":["medical-imaging","dicom","dos","supply-chain","cve-2026-48705"],"_cs_type":"advisory","_cs_vendors":["MONAI","pydicom"],"content_html":"\u003cp\u003eMONAILabel and MONAI Core are vulnerable to input validation flaws originating from pydicom's failure to reject non-numeric values in Decimal String (DS) DICOM fields. An unauthenticated attacker can upload malformed DICOM files (approx. 132 KB) via the REST API, which are then persisted in the datastore. These invalid values (NaN/inf) propagate through spatial transforms in the MONAI inference pipeline, resulting in server crashes (Denial of Service) or the generation of geometrically incorrect AI inference results. The issue affects MONAI versions prior to 1.6.1 and pydicom versions 0.9.7 through 3.0.2.\u003c/p\u003e\n\u003ch2 id=\"attack-chain\"\u003eAttack Chain\u003c/h2\u003e\n\u003col\u003e\n\u003cli\u003eAttacker identifies a target MONAILabel server with an exposed REST API (default port 8000).\u003c/li\u003e\n\u003cli\u003eAttacker crafts a DICOM file with NaN or infinity values injected into mandatory spatial DS tags (e.g., PixelSpacing, ImagePositionPatient).\u003c/li\u003e\n\u003cli\u003eAttacker submits the crafted file to the server's \u003ccode\u003ePUT /datastore/\u003c/code\u003e endpoint without requiring authentication.\u003c/li\u003e\n\u003cli\u003eThe MONAILabel server stores the malicious DICOM file in its permanent datastore.\u003c/li\u003e\n\u003cli\u003eAttacker triggers a standard inference request via \u003ccode\u003ePOST /infer/\u003c/code\u003e targeting the stored malicious image ID.\u003c/li\u003e\n\u003cli\u003eThe MONAI \u003ccode\u003ePydicomReader\u003c/code\u003e and downstream transforms (\u003ccode\u003eLoadImaged\u003c/code\u003e/\u003ccode\u003eSpacingd\u003c/code\u003e) process the NaN/inf values without validation.\u003c/li\u003e\n\u003cli\u003eThe inference pipeline crashes with a \u003ccode\u003eRuntimeError\u003c/code\u003e or produces corrupted AI output depending on the transform stage.\u003c/li\u003e\n\u003c/ol\u003e\n\u003ch2 id=\"impact\"\u003eImpact\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003eDenial of Service: Repeated crashes of the inference service consuming server CPU and memory.\u003c/li\u003e\n\u003cli\u003eData Corruption: Silent production of incorrect AI segmentation masks when the pipeline does not immediately crash.\u003c/li\u003e\n\u003cli\u003ePersistence: Malicious DICOM files remain in the datastore, causing recurrent failures upon future inference requests.\u003c/li\u003e\n\u003c/ul\u003e\n\u003ch2 id=\"recommendation\"\u003eRecommendation\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003ePatch MONAILabel and MONAI Core to version 1.6.1 or later immediately.\u003c/li\u003e\n\u003cli\u003eImplement strict input validation for spatial metadata on all API-exposed upload endpoints before storing files.\u003c/li\u003e\n\u003cli\u003eRestrict network access to the MONAILabel REST API to authorized internal networks only.\u003c/li\u003e\n\u003cli\u003eImplement Authentication and Authorization (RBAC) for all MONAILabel API endpoints to prevent unauthenticated uploads.\u003c/li\u003e\n\u003cli\u003eDeploy monitoring for \u003ccode\u003eRuntimeError\u003c/code\u003e and \u003ccode\u003eitkGDCMImageIO\u003c/code\u003e related exceptions in inference server logs.\u003c/li\u003e\n\u003c/ul\u003e\n","date_modified":"2026-10-29T00:11:39Z","date_published":"2026-10-29T00:11:39Z","id":"https://feed.craftedsignal.io/briefs/2026-10-monailabel-dicom-corruption/","summary":"An unauthenticated remote attacker can crash the MONAILabel inference pipeline or cause silent medical AI data corruption by uploading crafted DICOM files containing invalid NaN or infinity values.","title":"MONAILabel: Remote Denial of Service and Inference Corruption (CVE-2026-48705)","url":"https://feed.craftedsignal.io/briefs/2026-10-monailabel-dicom-corruption/"}],"language":"en","title":"CraftedSignal Threat Feed - Cve-2026-48705","version":"https://jsonfeed.org/version/1.1"}