{"description":"Trending threats, MITRE ATT\u0026CK coverage, and detection metadata. Fed continuously.","feed_url":"https://feed.craftedsignal.io/products/open-webui-pip/open-webui--0.7.0/feed.json","home_page_url":"https://feed.craftedsignal.io/","items":[{"_cs_actors":[],"_cs_cpes":[],"_cs_cves":[],"_cs_exploited":false,"_cs_has_poc":false,"_cs_poc_references":[],"_cs_products":["Open WebUI (pip/open-webui \u003c 0.7.0)"],"_cs_severities":["high"],"_cs_tags":["xss","web-application","vulnerability","client-side-injection","open-webui"],"_cs_type":"advisory","_cs_vendors":[],"content_html":"\u003cp\u003eOpen WebUI is vulnerable to a high-severity stored Cross-Site Scripting (XSS) vulnerability, identified as CVE-2026-26192, stemming from insecure iFrame rendering in its citation model. An authenticated attacker can exploit this by intercepting and modifying HTTP requests when saving chat history to inject malicious HTML content into document metadata, specifically by adding \u003ccode\u003ehtml: true\u003c/code\u003e and embedding an XSS payload. When another user, particularly an administrator, views a citation containing this weaponized document within a shared chat, the vulnerable Open WebUI frontend renders the content in an iFrame with insufficient sandboxing (\u003ccode\u003eallow-scripts\u003c/code\u003e and \u003ccode\u003eallow-same-origin\u003c/code\u003e are hardcoded), allowing the XSS payload to execute. This can lead to session hijacking, exfiltration of sensitive information, or, in the case of an administrator, potential server-side Remote Code Execution (RCE) by leveraging other known vulnerabilities. The vulnerability affects \u003ccode\u003epip/open-webui\u003c/code\u003e versions prior to 0.7.0.\u003c/p\u003e\n\u003ch2 id=\"attack-chain\"\u003eAttack Chain\u003c/h2\u003e\n\u003col\u003e\n\u003cli\u003eAn authenticated attacker logs into Open WebUI and initiates a new chat session.\u003c/li\u003e\n\u003cli\u003eThe attacker attaches an arbitrary file as a \u0026quot;document source\u0026quot; to the chat.\u003c/li\u003e\n\u003cli\u003eWhile saving the chat history or message, the attacker uses an HTTP proxy tool (e.g., Burp Suite, Caido, ZAP) to intercept the outgoing save request.\u003c/li\u003e\n\u003cli\u003eWithin the intercepted request's JSON body, the attacker locates the object corresponding to the document source within the \u003ccode\u003ehistory\u003c/code\u003e and \u003ccode\u003emessages\u003c/code\u003e objects.\u003c/li\u003e\n\u003cli\u003eThe attacker modifies this object by adding \u003ccode\u003ehtml: true\u003c/code\u003e to its metadata and injects an XSS payload (e.g., \u003ccode\u003e\u0026lt;script\u0026gt;alert(document.cookie)\u0026lt;/script\u0026gt;\u003c/code\u003e) into the document content field.\u003c/li\u003e\n\u003cli\u003eThe attacker forwards the modified HTTP request, causing the Open WebUI server to store the malicious chat history with the embedded XSS payload.\u003c/li\u003e\n\u003cli\u003eThe attacker shares the link to this weaponized chat with a victim user (e.g., via a phishing link).\u003c/li\u003e\n\u003cli\u003eWhen the victim accesses the shared chat and clicks on the malicious document citation, the Open WebUI frontend renders the content in an insecure iFrame, executing the attacker's JavaScript payload in the victim's browser context.\u003c/li\u003e\n\u003c/ol\u003e\n\u003ch2 id=\"impact\"\u003eImpact\u003c/h2\u003e\n\u003cp\u003eSuccessful exploitation of this vulnerability allows an attacker to execute arbitrary client-side JavaScript code in the context of the victim's browser session. For low-privilege users, this can lead to session takeover, allowing the attacker to read session tokens from local storage and exfiltrate them to an attacker-controlled server. If an administrator is targeted and views the malicious citation, the XSS payload can be used to bypass security controls and potentially achieve server-side Remote Code Execution (RCE) by chaining with other known vulnerabilities, as described in GHSA-w7xj-8fx7-wfch. This poses a significant risk to the integrity and confidentiality of data within the Open WebUI environment.\u003c/p\u003e\n\u003ch2 id=\"recommendation\"\u003eRecommendation\u003c/h2\u003e\n\u003cul\u003e\n\u003cli\u003e\u003cstrong\u003ePatch CVE-2026-26192 immediately:\u003c/strong\u003e Upgrade all Open WebUI installations to version 0.7.0 or newer to mitigate the vulnerability.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eImplement Web Application Firewall (WAF) rules:\u003c/strong\u003e Configure a WAF to inspect \u003ccode\u003ePOST\u003c/code\u003e requests to chat history save endpoints for unusual modifications like the addition of \u003ccode\u003ehtml: true\u003c/code\u003e in JSON bodies, and block requests containing common XSS payload patterns in document content, though this may require product-specific WAF configuration for JSON body inspection.\u003c/li\u003e\n\u003cli\u003e\u003cstrong\u003eEducate users on phishing awareness:\u003c/strong\u003e Warn users about suspicious shared chat links or messages that encourage clicking on document citations from unknown or untrusted sources.\u003c/li\u003e\n\u003c/ul\u003e\n","date_modified":"2026-07-07T16:56:32Z","date_published":"2026-07-07T16:56:32Z","id":"https://feed.craftedsignal.io/briefs/2026-07-open-webui-stored-xss/","summary":"An authenticated attacker can achieve stored Cross-Site Scripting (XSS) in Open WebUI by manually modifying chat history requests to inject malicious HTML into citation document metadata, leading to session takeover or potential Remote Code Execution (RCE) on the server if an administrator is targeted.","title":"Open WebUI Vulnerable to Stored Cross-Site Scripting (XSS) via iFrame in Citations Model","url":"https://feed.craftedsignal.io/briefs/2026-07-open-webui-stored-xss/"}],"language":"en","title":"CraftedSignal Threat Feed - Open WebUI (Pip/Open-Webui \u003c 0.7.0)","version":"https://jsonfeed.org/version/1.1"}