Key points
- CVE-2026-57118 carries a CVSS score of 9.8, rated critical by the National Vulnerability Database.
- The flaw affects PraisonAI versions prior to 1.6.59, allowing unauthenticated access to agent lists and chat functions.
- MervinPraison released a fix in praisonaiagents 1.6.59 to restore required access-control checks.
Business leaders must ensure their PraisonAI deployments are updated immediately to prevent unauthorized access. A critical vulnerability in the multi-agent system allows remote attackers to enumerate agents and consume resources without authentication. The National Vulnerability Database rates CVE-2026-57118 as critical with a CVSS score of 9.8.
The flaw exists in the AgentTeam.launch function and related aliases. Prior to version 1.6.59, these components register endpoints without proper access-control checks. This omission permits unauthenticated clients to list agents, trigger chat functions, and potentially access private context or trigger workflow side effects.
What we know
The vulnerability affects the praisonaiagents package developed by MervinPraison. According to the NVD record, versions below 1.6.59 are vulnerable. The issue also impacts the main PraisonAI product in versions below 4.6.59. The weaknesses are classified under CWE-306 and CWE-862, indicating missing authentication and improper access control mechanisms.
Attackers can exploit this by sending requests to specific GET and POST endpoints. These endpoints include paths for listing agents and interacting with individual agent IDs. Because no authentication is required, any remote client can interact with the system. This allows for resource consumption and potential data exposure from connected private contexts.
MervinPraison has addressed the issue in praisonaiagents version 1.6.59. This update introduces the necessary access-control checks to the affected endpoints. Organizations using earlier versions are exposed to remote exploitation until they apply the patch. The fix restores security boundaries for agent enumeration and chat interactions.
Defender's checklist
- Verify current PraisonAI and praisonaiagents versions against the vulnerable ranges.
- Identify all exposed endpoints that allow agent listing or chat interactions.
- Check logs for unauthenticated requests to agent-related paths prior to patching.
- Confirm that access-control checks are active after upgrading to 1.6.59.
What to do and how to stay safe: PraisonAI
- Audit your network perimeter to ensure PraisonAI services are not publicly accessible without authentication.
- Monitor server resources for unusual spikes that may indicate unauthorized model usage or tool triggering.
- Review access logs for requests to agent listing endpoints that lack valid user credentials.
- Apply the vendor update to praisonaiagents 1.6.59 as soon as possible to close the gap.
Step-by-step guide: Security Advisories: Small Business Patching Strategy
General security guidance from the Malware Chronicle newsroom. It is not confirmed advice from the organisations named in this story.
Frequently asked questions
What is the CVSS score for CVE-2026-57118?
The National Vulnerability Database assigns a CVSS score of 9.8, which is rated as critical.
Which versions of PraisonAI are affected by this vulnerability?
PraisonAI versions below 4.6.59 and praisonaiagents versions below 1.6.59 are affected.
Is there a fix available for this critical flaw?
Yes, MervinPraison released a fix in praisonaiagents version 1.6.59 that adds access-control checks.
