On this page (7 sections)

A critical code injection flaw in Vanderbilt University's Research Electronic Data Capture (REDCap) platform, tracked as CVE-2026-90817, has been added to the CISA Known Exploited Vulnerabilities (KEV) catalog on 2026-11-18. The flaw can let an unauthenticated attacker run arbitrary code on the REDCap server through crafted HTTP requests.

If your organization hosts REDCap, especially with public surveys reachable from the internet, treat this as an urgent item. REDCap often holds sensitive research and health-related data, so a compromised server can have serious consequences.

What happened

The NVD record describes a code injection vulnerability in REDCap that could allow an unauthenticated attacker to execute arbitrary code on the server. The GitHub advisory was published on Sep 20, 2026. On 2026-11-18, CISA added the issue to its KEV catalog, which is the signal that exploitation has been observed. The sources we reviewed do not say who is exploiting it, how many systems are affected, or whether ransomware groups are involved; CISA lists the ransomware field as Unknown.

CISA's KEV entry gives a due date of 2026-11-21 and asks for mitigations in line with vendor instructions, or discontinuing use of the product if mitigations are unavailable. That due date formally applies to US federal agencies, but it is a useful benchmark for everyone else.

Technical details in plain terms

According to the advisory, two weaknesses combine. First, the survey passthrough routing can be manipulated so that a request coming from a public survey context reaches a controller route it was never meant to reach. Second, the Data Import processing logic accepts a crafted file-path or stream parameter. Together, these can lead to arbitrary code execution on the REDCap server.

The two weakness classes explain the risk. CWE-73 (external control of file name or path) means user input influences which file or stream the server touches. CWE-94 (code injection) means attacker-controlled content ends up being run as code. When both are reachable without logging in, the result is a remote code execution bug, which is why the score is 9.8.

Exploitation requires knowledge of a valid public survey hash. This is a real precondition, but it should not be read as strong protection. Survey links are designed to be shared, so the hash of a public survey is often visible to anyone who has received or seen the link.

Who is exposed

  • Organizations running self-hosted REDCap 13.3.0 or any later version, per the advisory wording "13.3.0 and higher".
  • Servers where public survey URLs are reachable from the internet. The attacker needs a valid public survey hash, so instances that publish surveys are the most realistic targets.
  • Installations that cannot be patched quickly, such as those with change freezes or shared institutional hosting.

The sources do not list a fixed version or patched release. They also do not say whether REDCap versions below 13.3.0 are unaffected beyond stating that the flaw exists in 13.3.0 and higher. Confirm both points with your REDCap administrator or the consortium's official communications rather than assuming.

How to check your exposure

Start by finding out which version you run and which surveys are public. The version is shown in the REDCap Control Center for administrators. Then review your web server logs for unusual activity around survey and import requests.

This read-only command searches an Apache or Nginx access log for requests to survey URLs from the last log file. Run it on the REDCap web server and look for unusual parameters, unexpected routes or spikes from single IP addresses. Adjust the path to your log location.

Search web access logs for survey requests (Linux, adjust log path)Shell
grep -i "surveys" /var/log/nginx/access.log | tail -n 200

Also review the server itself for signs of compromise: unexpected files in the web root or temporary directories, new scheduled jobs, new user accounts and unusual outbound connections from the web server process.

What to do now

  1. Identify every REDCap instance you operate and record its version. Anything at 13.3.0 or higher should be treated as vulnerable until the vendor says otherwise.
  2. Check the official REDCap community and Vanderbilt communications for a patch or mitigation guidance, and apply it as soon as it is available. The research we reviewed does not name a fixed version.
  3. If no fix or mitigation is available to you yet, restrict access: place the instance behind a VPN or IP allowlist, or take public surveys offline until you can update.
  4. Where you must keep surveys public, put a web application firewall or reverse proxy in front and review rules for requests to survey and import routes.
  5. Review web server and application logs from before 2026-11-18 for suspicious survey and import requests, since exploitation may predate the KEV listing.
  6. If you find signs of compromise, isolate the server, preserve logs and disk images, rotate credentials and database passwords, and follow your incident response and data breach notification obligations.
  7. Run the web server and PHP processes with least privilege, and restrict outbound connections from the server to limit what code execution can achieve.
Caution. Restricting access or taking surveys offline will interrupt data collection. Coordinate with research teams first, and keep a record of changes so you can restore access safely after patching.

Detection and monitoring afterwards

After patching or restricting access, keep watching. Alert on new or modified files in the application directory, unexpected child processes spawned by the web server, and outbound connections to unfamiliar hosts. Keep access logs for at least several weeks so you can look back if new indicators are published.

From the outside, you can track changes to your public-facing endpoints with Webforta Monitor, which watches uptime, certificates, DNS and header changes. Outside-in checks cannot see inside the server, so they complement, rather than replace, log review and host-level monitoring. For background on what that kind of scanning can and cannot find, see what outside-in malware detection can see.

If you suspect a compromise, our guide on what to do when a website is hacked covers containment and recovery steps.

What is still unknown

  • A fixed or patched REDCap version is not listed in the GitHub advisory or NVD entry.
  • The identity of the attackers, the scale of exploitation and any link to ransomware are not stated in the sources.
  • Whether versions below 13.3.0 are affected is not addressed beyond the "13.3.0 and higher" wording.
  • The GitHub advisory is marked unreviewed and lists no package, so dependency scanners will not flag it automatically. You need to check your REDCap version manually.

The GitHub advisory also shows an EPSS score of 3.749% (90th percentile), an estimate of exploitation probability. Given CISA's KEV listing, treat exploitation as confirmed rather than a probability.

Sources

This article is based on the public sources below. Details can change as vendors and researchers publish more; check the original advisories before acting on version numbers.

  1. CVE-2026-90817 - GitHub Advisory Databasegithub.com
  2. Vanderbilt University Research Electronic Data Capture (REDCap) Code Injection Vulnerability (CVE-2026-90817) is being actively exploitednvd.nist.gov ·
  • REDCap
  • CVE-2026-90817
  • CISA KEV
  • RCE
  • Vanderbilt University