Arbitrary Command Execution Risk in PraisonAI Multi-Agent System
CVE-2026-57117

8.8HIGH

Key Information:

Vendor

Mervinpraison

Status
Praisonai
Vendor
CVE Published:
23 October 2026

What is CVE-2026-57117?

PraisonAI, a multi-agent team system, is susceptible to a vulnerability that enables arbitrary command execution. This flaw arises when the LocalManagedAgent and SandboxedAgent compute bridging process wraps file operations, converting them into shell command strings that are sent to shell-backed compute providers. If an attacker manipulates a file-tool path argument, they can break out of the intended path to execute unauthorized shell commands in the compute environment. Commands executed with compute set to 'local' run on the host, while those using Docker execute within the container. This vulnerability exists in PraisonAI versions 4.6.10 to just before 4.6.59 and has been patched in version 4.6.59.

Affected Version(s)

PraisonAI >= 4.6.10, < 4.6.59

References

https://github.com/MervinPraison/PraisonAI/security/advis...
x_refsource_CONFIRM

CVSS V3.1

Score:
8.8
Severity:
HIGH
Confidentiality:
High
Integrity:
High
Availability:
High
Attack Vector:
Network
Attack Complexity:
Low
Privileges Required:
Low
User Interaction:
None
Scope:
Unchanged

Timeline

  • Vulnerability published

  • Vulnerability Reserved

.