CVE-2026-39890
Published: 08 April 2026
Summary
CVE-2026-39890 is a critical-severity Deserialization of Untrusted Data (CWE-502) vulnerability in Praison Praisonai. Its CVSS base score is 9.8 (Critical).
Operationally, exploitation aligns with the MITRE ATT&CK technique Exploit Public-Facing Application (T1190); ranked in the top 31.7% of CVEs by exploit likelihood; it is not currently listed in the CISA KEV catalog.
The strongest mitigations our analysis identified are NIST 800-53 SI-10 (Information Input Validation) and SI-2 (Flaw Remediation).
Threat & Defense at a Glance
Threat & Defense Details
Mitigating Controls (NIST 800-53 r5)AI
Requires validation of YAML inputs at the API endpoint to reject malicious payloads with dangerous tags like !!js/function, directly preventing unsafe deserialization and RCE.
Mandates timely application of vendor patches, such as upgrading to PraisonAI 4.5.115, which fixes the unsafe YAML parsing vulnerability.
Restricts untrusted file uploads at the API to safe YAML formats or types, limiting the attack surface for malicious agent definitions.
MITRE ATT&CK Enterprise TechniquesAI
Why these techniques?
The deserialization vulnerability (CWE-502) in PraisonAI's public-facing API endpoint enables unauthenticated remote attackers to upload malicious YAML files, resulting in arbitrary JavaScript RCE, directly facilitating T1190: Exploit Public-Facing Application.
NVD Description
PraisonAI is a multi-agent teams system. Prior to 4.5.115, the AgentService.loadAgentFromFile method uses the js-yaml library to parse YAML files without disabling dangerous tags (such as !!js/function and !!js/undefined). This allows an attacker to craft a malicious YAML file that,…
more
when parsed, executes arbitrary JavaScript code. An attacker can exploit this vulnerability by uploading a malicious agent definition file via the API endpoint, leading to remote code execution (RCE) on the server. This vulnerability is fixed in 4.5.115.
Deeper analysisAI
CVE-2026-39890 is a critical deserialization vulnerability (CWE-502) in PraisonAI, an open-source multi-agent teams system. Prior to version 4.5.115, the AgentService.loadAgentFromFile method parses YAML files using the js-yaml library without disabling dangerous tags like !!js/function and !!js/undefined. This flaw enables attackers to execute arbitrary JavaScript code during parsing of untrusted YAML input. The vulnerability carries a CVSS v3.1 base score of 9.8 (AV:N/AC:L/PR:N/UI:N/S:U/C:H/I:H/A:H), indicating high severity due to its network accessibility and lack of prerequisites.
Any unauthenticated remote attacker can exploit this vulnerability by crafting a malicious YAML file containing JavaScript payloads and uploading it as an agent definition via the PraisonAI API endpoint. Successful exploitation leads to remote code execution (RCE) on the server hosting the application, potentially allowing full compromise of the environment, data exfiltration, or further lateral movement.
The vulnerability is addressed in PraisonAI version 4.5.115, where the parsing logic was updated to mitigate the dangerous YAML tags. Security practitioners should upgrade to this version or later. Official details are available in the GitHub Security Advisory at https://github.com/MervinPraison/PraisonAI/security/advisories/GHSA-32vr-5gcf-3pw2.
As PraisonAI supports multi-agent AI workflows, this RCE flaw could impact AI/ML deployments relying on dynamic agent loading from files, though no public evidence of real-world exploitation has been reported as of the CVE publication on 2026-04-08.
Details
- CWE(s)