An attacker needs a valid public survey hash, rather than a REDCap account, to reach a code-execution flaw in the research platform. CVE-2026-90817 entered CISA’s Known Exploited Vulnerabilities catalog on November 18. The warning puts teams running Vanderbilt University’s REDCap research-data system on a short clock: CISA’s listed remediation date is November 21.

This is a server-side risk for institutions that operate REDCap and publish surveys. A successful attack could run arbitrary code on the REDCap server. The CVE Program record, assigned by Securifera, credits researcher Ryan Wincey and rates the flaw 9.8 under CVSS v3.1. Scope exposure to REDCap instances with reachable public surveys and valid hashes; those are the conditions for the route described in the record. A hash is a locator for a survey, not an authenticated administrator session.

Survey routing reaches import processing

REDCap gathers study data through public surveys as well as authenticated research workflows. Here, the CVE record places the flaw at the junction of survey passthrough routing and Data Import processing. An attacker with a valid public survey hash can manipulate HTTP requests from that public context to reach an unintended controller route, then supply a crafted file-path or stream parameter during import handling. The path crosses from a survey meant for public submission into a server operation that handles imported data. The source classifies the weakness as external control of a file name or path (CWE-73) and code injection (CWE-94).

CISA’s description says the requests “could allow an unauthenticated attacker to execute arbitrary code on the server.” CISA lists the vulnerability as known exploited; its CVE enrichment also marks exploitation active. Treat an exposed instance as an incident-triage priority alongside the application update. Treat a matching route as a lead rather than an exploit signature. Preserve web and application logs before rotation, then examine unusual transitions from public survey to import handling alongside server process activity.

Fixed releases span three branches

The CVE record describes REDCap 13.3.0 and higher in its affected description, and separately identifies 16.0.49 LTS, 17.3.10 LTS and 17.4.4 Standard as unaffected versions. Its stated solution is to upgrade to the appropriate one of those releases. Read those branch-specific fixed versions together: the phrase “and higher” in the broad description is not a reason to label a listed fixed release vulnerable. Match each installation’s release branch to its applicable fixed version; record the installed build before and after the change.

CISA directs operators to follow vendor mitigations and its BOD 26-04 guidance, including forensic triage. Vanderbilt’s REDCap Community site is the technical support and download channel for registered administrators at partner institutions. Have your institution’s REDCap administrator obtain the applicable vendor release and deployment guidance there. If the site is hosted for you, ask the operator for the deployed REDCap version and remediation status rather than assuming your study team’s access to a survey means it owns the server.

Trace exposure through survey access

Start with your REDCap inventory: identify instances that serve public surveys, the internet-facing routes that reach them, the installed versions and the people who manage upgrades. Prioritize a reachable instance on an affected branch, especially one whose public survey hash is in use. Review reverse-proxy and application logs around survey requests and import processing, together with web-server child processes and unexpected files or outbound activity from the REDCap host. A quiet log may reflect retention or visibility limits, so give the forensic team the logs you have before changing the system.

If immediate upgrade is blocked, restrict access to the affected application from networks that do not need it and work with the vendor administrator on supported mitigations. Where a public survey must stay open, coordinate access changes with the study owner so patients or participants are not silently cut off. CISA calls for vendor instructions and forensic triage; it also says to follow applicable BOD 26-04 cloud-service guidance or discontinue use when mitigations are unavailable. The November 21 date is CISA’s KEV remediation date; institutions outside that directive can use it as an urgency marker without treating it as their own statutory deadline.

Verify the deployed REDCap release

Have the administrator move each affected installation to the appropriate fixed branch release: 16.0.49 LTS, 17.3.10 LTS or 17.4.4 Standard. In parallel, preserve proxy, application and host evidence from the period when public surveys were reachable, and investigate suspicious requests in the survey-to-import path with the incident team. A version change removes the documented vulnerability from the patched instance; it cannot erase activity that happened earlier.

After the update, have the administrator read the installed REDCap version from the running application, compare it with the intended branch release, and confirm a legitimate public survey still submits through the approved route. That test checks the fix and the research workflow together. If the release is behind, restrict access and schedule the vendor-supported update rather than treating the inventory entry as proof that the live server changed.