CVE-2026-78379

CVE-2026-78379

High · published August 25, 2026

CVSS v3.1
8.1
EPSS
1%
Percentile
46.7
In the wild
Unconfirmed
What it is

Improper neutralization of input used for LLM prompting in the python_repl tool in Amazon Strands Agents Tools before 0.8.5 might allow remote actors to execute arbitrary Python code on the agent's host by bypassing the human consent gate, via a crafted prompt that forwards non_interactive_mode as a keyword argument through the batch tool. To remediate this issue, users should upgrade to version 0.8.5 or later.

The record
Technical detail
CVSS v3.1
8.1 · HIGH
Vector
CVSS:3.1/AV:N/AC:H/PR:N/UI:N/S:U/C:H/I:H/A:H
CVSS v4.0
9.2 · CVSS:4.0/AV:N/AC:L/AT:P/PR:N/UI:N/VC:H/VI:H/VA:H/SC:N/SI:N/SA:N
EPSS
0.00575 · 46.7th percentile
Published
2026-08-25T23:16Z
References (2)
https://aws.amazon.com/security/security-bulletins/2026-089-aws/ · ff89ba41-3aa1-4d27-914a-91399e9639e5https://pypi.org/project/strands-agents-tools/0.8.5/ · ff89ba41-3aa1-4d27-914a-91399e9639e5
EPSS history
Timeline
  • 27 AUG 06:46Z
    EPSS moved — → 0%
    epss
  • 25 AUG 19:05Z
    Consent bypass in python_repl tool via batch kwargs forwarding in Amazon Strands Agents Tools
    cvelistv5