PT-2026-108714 · Unknown · Praisonaiagents
CVE-2026-62177
·
Published
2026-10-08
·
Updated
2026-10-15
CVSS v4.0
8.7
High
| Vector | AV:N/AC:L/AT:N/PR:N/UI:N/VC:H/VI:N/VA:N/SC:N/SI:N/SA:N |
Name of the Vulnerable Software and Affected Versions
PraisonAI Agents (affected versions not specified)
Description
The Subprocess Sandbox backend fails to enforce security policies defined in the
SecurityPolicy data class. Although the strict() method is intended to create a restrictive environment for untrusted code by setting allow subprocess and allow file write to False and defining blocked paths and blocked commands, these restrictions are not implemented in the praisonai/sandbox/subprocess.py file. The backend invokes commands via asyncio.create subprocess exec() without validating them against the configured security fields. This allows for unauthorized subprocess creation, access to sensitive files such as /etc/passwd, and the execution of destructive commands like rm -rf regardless of the active security policy.Recommendations
Enforce the
blocked commands, blocked imports, blocked paths, allow subprocess, and allow file write restrictions within the Subprocess Sandbox before invoking child processes.
Use native backends such as Landlock or Seatbelt for executing untrusted code instead of the subprocess backend.
Restrict the use of the SecurityPolicy.strict() configuration when using backends that cannot enforce its security fields.Fix
Found an issue in the description? Have something to add? Feel free to write us 👾
Weakness Enumeration
Related Identifiers
Affected Products
Praisonaiagents