Plattform
python
Komponente
gpt-researcher
Behoben in
3.4.1
3.4.2
3.4.3
3.4.4
A code injection vulnerability has been identified in assafelovic's gpt-researcher, impacting versions 3.4.0 through 3.4.3. This flaw allows remote attackers to inject malicious code by manipulating arguments within the extractcommanddata function of the backend/server/server_utils.py file. The vulnerability is considered HIGH severity and has been publicly disclosed, posing a significant risk to deployments. A fix is pending from the project.
Successful exploitation of CVE-2026-5631 allows an attacker to execute arbitrary code on the server hosting the gpt-researcher instance. This could lead to complete system compromise, including data exfiltration, modification, or deletion. The remote nature of the vulnerability means an attacker does not need local access to exploit it. Given the potential for code execution, the blast radius extends to any sensitive data processed or stored by the gpt-researcher application. The lack of a response from the project increases the risk of exploitation.
This vulnerability was publicly disclosed on 2026-04-06. A proof-of-concept may be available or developed shortly due to the public disclosure. The vulnerability is not currently listed on CISA KEV, but its HIGH severity and public disclosure warrant close monitoring. The project's lack of response to the reported issue raises concerns about the timeliness of a potential fix.
Organizations deploying gpt-researcher in production environments, particularly those with limited security controls or exposed endpoints, are at significant risk. Systems handling sensitive data or integrated with critical infrastructure are especially vulnerable. Shared hosting environments where multiple users share the same server instance also increase the potential attack surface.
• python / server:
import subprocess
# Check for suspicious command executions in server logs
# Example: grep 'eval' /var/log/gpt-researcher/server.log• generic web:
curl -I <gpt-researcher-endpoint> | grep -i 'Content-Type: application/x-python'
# Look for unusual content types that might indicate code executiondisclosure
Exploit-Status
EPSS
0.06% (19% Perzentil)
CISA SSVC
CVSS-Vektor
Due to the absence of a patch from the project, immediate mitigation is crucial. Consider temporarily disabling the affected endpoint (/ws) to prevent exploitation. Implement strict input validation and sanitization on all data received by the extractcommanddata function. Employ a Web Application Firewall (WAF) with rules to detect and block suspicious code injection attempts. Monitor system logs for unusual activity, particularly related to process execution or file modifications. While a direct fix is unavailable, these measures can significantly reduce the attack surface.
Aktualisieren Sie auf eine korrigierte Version von gpt-researcher, die die Code Injection (Code Injection) Schwachstelle in der Funktion extract_command_data behebt. Da das Projekt nicht reagiert hat, wird empfohlen, manuell zu untersuchen und Patches anzuwenden oder aktiv gepflegte Forks mit Korrekturen zu suchen.
Schwachstellenanalysen und kritische Warnungen direkt in deinen Posteingang.
CVE-2026-5631 is a code injection vulnerability affecting gpt-researcher versions 3.4.0–3.4.3. It allows remote attackers to execute arbitrary code via the extractcommanddata function.
You are affected if you are using gpt-researcher versions 3.4.0, 3.4.1, 3.4.2, or 3.4.3 and have not yet applied a patch. Monitor for updates from the project maintainers.
A patch is pending. Implement input validation, WAF rules, and monitor logs as immediate mitigations. Upgrade as soon as a fix is released.
The vulnerability has been publicly disclosed, and a public exploit may be available. Monitor security advisories for confirmed exploitation.
Check the project's GitHub repository and relevant security mailing lists for updates. The project maintainers have not yet responded to the issue report.
Lade deine Abhängigkeitsdatei hoch und erfahre sofort, ob dich diese und andere CVEs treffen.
Lade deine requirements.txt-Datei hoch und wir sagen dir sofort, ob du betroffen bist.