Advisory record · GHSA-6vh2-h83c-9294

PraisonAI: Python Sandbox Escape via str Subclass startswith() Override in execute_code

CVSS 10.0 CriticalCriticalEPSS 0.13 · 96th percentile

Critical, because:

  • CVSS score 10.0: 9.0 or more.
  • CVSS score 10.0 (7.0 or more) and EPSS 0.13 (0.1 or more): a high score and a real chance of exploitation in the next 30 days.

How we decide what’s critical

Published by the GitHub Advisory Database at 29 Jun 2026, 11:50 UTC. agentprime has not analysed this yet, so whether it affects any app is not known.

PackageEcosystemAffectedFixed in
praisonaiagentsPyPI<1.5.901.5.90
Identifiers
GHSA-6vh2-h83c-9294 · CVE-2026-34938 · PYSEC-2026-484
Weakness (CWE)
CWE-693
Published
29 Jun 2026
Updated
13 Jul 2026

Machine-written summary

Summary written by Claude Haiku from the source advisory. Not reviewed by a person.

The praisonaiagents package (PyPI) before 1.5.90 is vulnerable. The execute_code() function's sandbox can be bypassed by passing a str subclass with an overridden startswith() method, letting an attacker run arbitrary OS commands as the process user. Upgrade to praisonaiagents 1.5.90 or later; the source describes deployments that auto-approve tool execution by default, which makes indirect prompt injection more dangerous.

Written 6 Nov 2026.

Not analysed yet

We haven’t published our analysis of this Advisory yet. Once a person has reviewed it, this page will say what’s vulnerable, what has to be true for your app to be affected, and how to check.

Severity and exploitation

ScoreRatingCVSSVectorSource
10.0 Critical 3.1 CVSS:3.1/AV:N/AC:L/PR:N/UI:N/S:C/C:H/I:H/A:H CNA (GitHub_M)
  • CISA KEV Not listed.
  • EPSS 0.13, the 96th percentile, on 9 Nov 2026: FIRST’s estimate of the chance of exploitation in the next 30 days.

Timeline

  1. 31 Oct 2026, 15:04 UTC First seen by agentprime
  2. 31 Oct 2026, 15:04 UTC Confirmed by OSV

Sources

Facts on this page come from the GitHub Advisory Database (CC-BY 4.0), OSV and the other sources listed, as published. We don’t reproduce full advisory descriptions, which can contain exploit code.