CVE-2026-78379
LLM prompt injection in python_repl lets attackers run arbitrary Python code before 0.8.5.
- CVSS 9.2
- CWE-94 (inferred)
- Input Validation and Sanitization
- Remote
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.
- CWE
- CWE-94 (inferred)
- CVSS base score
- 9.2
- Published
- 2026-08-25
- OWASP
- A03 Injection
- Orthogonal defect classification
- Interface
- Code defect classification
- Incorrect Interface
- Category
- Input Validation and Sanitization
- Subcategory
- Command Injection
- Accessibility scope
- Remote
- Impact
- Arbitrary Code Execution
- Affected component
- Python
- Fixed by upgrading
- Yes
Solution
Upgrade the Amazon Strands Agents Tools package to version 0.8.5 or newer.
Vulnerable code sample
import json
from strands_agents_tools import python_repl
def execute_user_prompt(prompt: str, **kwargs):
"""
Executes a user supplied prompt via the python_repl tool.
The **kwargs are forwarded directly, allowing a crafted prompt to
inject `non_interactive_mode=True` and bypass the consent gate.
"""
# VULNERABLE: forwards unvalidated kwargs to python_repl, enabling code injection
result = python_repl.run(prompt, **kwargs)
return json.dumps({"output": result})Patched code sample
import json
from strands_agents_tools import python_repl
def execute_user_prompt(prompt: str, **kwargs):
"""
Executes a user supplied prompt via the python_repl tool.
Only a whitelist of safe keyword arguments is allowed.
"""
# FIX: filter kwargs to allowed keys, discarding non_interactive_mode
allowed_keys = {"timeout", "max_iterations"}
safe_kwargs = {k: v for k, v in kwargs.items() if k in allowed_keys}
result = python_repl.run(prompt, **safe_kwargs)
return json.dumps({"output": result})Cite this entry
@misc{vaitp:cve202678379,
title = {{LLM prompt injection in python_repl lets attackers run arbitrary Python code before 0.8.5.}},
author = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
year = {2026},
note = {VAITP Python Vulnerability Dataset, entry CVE-2026-78379},
howpublished = {\url{https://netpack.pt/vaitp/vulnerability/CVE-2026-78379/}}
}
Introducing the "VAITP dataset": a specialized repository of Python vulnerabilities and patches, meticulously compiled for the use of the security research community. As Python's prominence grows, understanding and addressing potential security vulnerabilities become crucial. Crafted by and for the cybersecurity community, this dataset offers a valuable resource for researchers, analysts, and developers to analyze and mitigate the security risks associated with Python. Through the comprehensive exploration of vulnerabilities and corresponding patches, the VAITP dataset fosters a safer and more resilient Python ecosystem, encouraging collaborative advancements in programming security.
The supreme art of war is to subdue the enemy without fighting.
Sun Tzu – “The Art of War”
:: Shaping the future through research and ingenuity ::
