Connect your agent
AdvisoriesCriticalCheck your lockfilePluginHow it worksPricing Connect your agent

Advisory record · GHSA-29r9-67vg-qj56

PraisonAI: Human-in-the-loop tool approval is cached by tool name and silently reused for all subsequent calls with arbitrary arguments

CVSS 6.9 MediumEPSS 0.0034 · 27th percentile

Published by the GitHub Advisory Database at 8 Oct 2026, 16:43 UTC. agentprime has not analysed this yet, so whether it affects any app is not known.

PackageEcosystemAffectedFixed in
praisonaiagentsPyPI<1.6.781.6.78
Identifiers
GHSA-29r9-67vg-qj56 · CVE-2026-60087 · PYSEC-2026-4265
Weakness (CWE)
CWE-863
Tags
Published
8 Oct 2026
Updated
15 Oct 2026

Machine-written summary

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

The PraisonAI package praisonaiagents, versions before 1.6.78, caches a human approval for a dangerous tool by tool name alone, so a single approval of a tool such as write_file is silently reused for later calls with different arguments for the rest of the run. An attacker who can steer the model with malicious input could then have the agent perform further actions, such as writing to sensitive locations, without a new prompt. Upgrade to praisonaiagents 1.6.78 or later, which fixes the issue.

Written 1 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
6.9 Medium 4.0 CVSS:4.0/AV:L/AC:L/AT:N/PR:N/UI:P/VC:N/VI:H/VA:L/SC:N/SI:N/SA:N CNA (VulnCheck)
  • CISA KEV Not listed.
  • EPSS 0.0034, the 27th percentile, on 8 Nov 2026: FIRST’s estimate of the chance of exploitation in the next 30 days.

Timeline

  1. 31 Oct 2026, 14:08 UTC First seen by agentprime
  2. 31 Oct 2026, 14:08 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.