// Security Advisory Triage
PraisonAI agents package allows arbitrary code execution via tools.py
GitHub advisories — pip · published 2026-10-09 13:19 UTC
WHAT CHANGEDGitHub published CVE-2026-61437 on 8 October 2026, rating it high severity. The praisonaiagents package before version 1.6.78 unsafely loads modules dynamically in tools.py, allowing arbitrary code execution. Version 1.6.78 patches the vulnerability.
WHY IT MATTERSAn attacker who can control input to the tools.py module loading mechanism can execute arbitrary code in your application's context. This is a direct path to full system compromise in any deployment where the agent framework processes untrusted input.
WHAT TO DOCheck if you are running praisonaiagents below 1.6.78. If so, upgrade to 1.6.78 or later this week. If you cannot upgrade immediately, audit all code paths that invoke the agent framework's tool loading and ensure no untrusted input reaches them.
Model-reported confidence 90% (not measured accuracy)
Primary source ↗
written by anthropic/claude-sonnet-4.5
Show the reasoning — raw diff, triage verdict, confidence
RAW DIFF
+ghsa_id: GHSA-4gfv-wg42-7jw5 +summary: PraisonAI: Unsafe Dynamic Module Loading Leads to Arbitrary Code Execution via tools.py in AgentFlow +severity: high +published_at: 2026-10-08T16:48:57Z +cve_id: CVE-2026-61437 +vulnerabilities[].package.name: praisonaiagents +vulnerabilities[].first_patched_version: 1.6.78
HOW IT WAS JUDGED
Triage signals below are model-reported assertions, preserved verbatim; they are not independent arithmetic verification.
| triage verdict | MATERIAL at 0.85, needed 0.5 |
|---|---|
| triage reasoning | PraisonAI is an AI agent framework — a dependency likely in services running AI workloads. Arbitrary code execution via unsafe module loading is exploitable. |
| signals | AI agent framework package · arbitrary code execution vulnerability · high severity rating · patched version available (1.6.78) · unsafe dynamic module loading in tools.py |
| suppression | — |
| triage model | anthropic/claude-haiku-4.5 ($0.00000) |
| writer model | anthropic/claude-sonnet-4.5 ($0.00000) |