CVE-2026-57957: Papermark CORS Misconfiguration Allows Silent File Uploads

Technical summary

CVE-2026-57957 is a CORS misconfiguration in Papermark's TUS-based viewer upload endpoint. The endpoint reflects the Origin header from incoming requests and sets Access-Control-Allow-Credentials to true, enabling cross-origin credentialed requests from any origin. An attacker crafts a webpage that, when visited by an authenticated Papermark user, issues fetch() or XMLHttpRequest calls to the vulnerable endpoint. The browser permits these requests because the endpoint mirrors the attacker's origin as allowed and signals credential inclusion is permitted. This allows file upload into victim datarooms and exfiltration of credentialed API responses. The vulnerability requires user interaction (victim must visit the malicious page while authenticated to Papermark) but succeeds even against users who never leave their Papermark session open in the background.

Business impact

Organizations relying on Papermark for secure dataroom sharing face unauthorized file uploads and potential data exfiltration. Compromised datarooms could contain M&A documents, legal contracts, IP, or financial records. The attack is silent and difficult to detect in real time. While the CVSS score is MEDIUM, the business impact depends on dataroom contents and threat landscape; a dataroom holding due-diligence materials for an acquisition carries different risk than one used for general document sharing.

Affected systems

Papermark versions through 0.22.0 are affected. The vulnerability resides in the TUS-based viewer upload endpoint, so any Papermark instance running an unpatched version that accepts uploads from authenticated users is in scope. This includes self-hosted and cloud-hosted deployments.

Exploitability

Exploitability is straightforward for an attacker but requires user interaction. The attacker must craft a malicious webpage and trick an authenticated Papermark user into visiting it—via phishing, watering hole, or social engineering. The attack requires no specialized tools or zero-days. Detection by the victim is unlikely because the upload occurs silently in the background. However, the attack is limited to the scope of the victim's Papermark account and permissions; it does not grant access to other users' datarooms unless the victim themselves has that access.

Remediation

Upgrade Papermark to a patched version released after 0.22.0. Verify the exact patched version against the official Papermark project repository or security advisory. Additionally, implement a content security policy (CSP) to restrict cross-origin requests and educate users about not visiting untrusted links while authenticated to sensitive applications.

Patch guidance

Contact Papermark's maintainers or check the official GitHub repository for patch availability and version numbers. Most organizations should prioritize patching this within 30 days given the medium severity and the ease of exploitation. If patching is delayed, compensating controls (e.g., IP whitelisting, additional authentication for sensitive operations) may reduce exposure.

Detection guidance

Monitor web server and application logs for suspicious upload requests originating from unexpected origins or user-agents. Look for uploads that occur outside normal user behavior patterns. Check dataroom audit logs for unauthorized file additions. Correlate upload activity with user login sessions to identify anomalies. In a WAF or proxy, flag requests to the TUS endpoint that lack proper origin validation.

Why prioritize this

Although the CVSS score is MEDIUM (4.7), the vulnerability merits attention because it directly undermines user trust in document confidentiality. Dataroom applications are often deployed specifically to prevent unauthorized access to sensitive materials. An attacker can silently compromise this assurance. The exploit path is simple and requires only social engineering. Organizations with high-value datarooms should deprioritize below critical business initiatives, but this should not remain unpached indefinitely.

Risk score, explained

CVSS 3.1 score of 4.7 reflects: network-accessible endpoint (AV:N), low attack complexity (AC:L), no privilege required (PR:N), but user interaction required (UI:R), changed scope (S:C), no confidentiality impact (C:N), limited integrity impact (I:L—files can be uploaded but not deleted or modified at scale), and no availability impact (A:N). The score is conservative because the attack does not lead to direct data exfiltration in the traditional sense; instead, it allows unauthorized file writes. In practice, unauthorized file uploads to a dataroom can enable follow-up attacks (e.g., social engineering via planted documents) and damage integrity.

Frequently asked questions