VAITP Dataset

← Back to the dataset

CVE-2026-12530

AWS Bedrock AgentCore SDK command injection via crafted package names.

  • CVSS 8.4
  • CWE-88
  • Input Validation and Sanitization
  • Remote

Improper neutralization of argument delimiters in the install_packages() method in AWS Bedrock AgentCore Python SDK versions >= 1.1.3 and < 1.6.1 might allow a remote authenticated user to execute arbitrary commands within the Code Interpreter sandbox via crafted package name arguments. To mitigate this issue, users should upgrade to version 1.6.1.

CVSS base score
8.4
Published
2026-06-17
OWASP
A03 Injection
Orthogonal defect classification
Checking
Code defect classification
Missing Check
Category
Input Validation and Sanitization
Subcategory
Command Injection
Accessibility scope
Remote
Impact
Arbitrary Code Execution
Affected component
AWS Bedrock
Fixed by upgrading
Yes

Solution

Upgrade AWS Bedrock AgentCore Python SDK to version 1.6.1.

Vulnerable code sample

import subprocess

class AgentCore:
    def install_packages(self, package_names: list[str]):
        # VULNERABLE: The list of package names is joined into a single string
        # and executed in a shell. This allows a crafted package name like
        # "package; arbitrary_command" to execute arbitrary commands.
        command = f"pip install {' '.join(package_names)}"
        subprocess.run(command, shell=True)

Patched code sample

import subprocess
from typing import List

def install_packages(package_names: List[str]):
    """
    This function represents a patched version of the install_packages method,
    fixing a command injection vulnerability like the one described in the
    hypothetical CVE-2026-12530.

    The fix involves passing command arguments as a list, which prevents the
    shell from interpreting special characters in the package names.
    """
    # The command and its primary argument are defined in a list.
    # Using 'python -m pip' is often more robust than just 'pip'.
    command = ["python", "-m", "pip", "install"]

    # The untrusted package names are appended to the command list.
    # Each item is treated as a single, distinct argument. Shell metacharacters
    # like ';' or '&&' within a package name are not executed. This is the
    # core of the fix that neutralizes the vulnerability.
    command.extend(package_names)

    # By using subprocess.run with a list of arguments (the default behavior,
    # where shell=False), we avoid passing the command through a shell.
    # This prevents a crafted package name like "some-package; id" from
    # executing the 'id' command. The system will instead safely look for
    # a non-existent package named "some-package; id" and fail gracefully.
    subprocess.run(command, check=True)

Payload

nonexistent-package; id

Cite this entry

@misc{vaitp:cve202612530,
  title        = {{AWS Bedrock AgentCore SDK command injection via crafted package names.}},
  author       = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
  year         = {2026},
  note         = {VAITP Python Vulnerability Dataset, entry CVE-2026-12530},
  howpublished = {\url{https://netpack.pt/vaitp/vulnerability/CVE-2026-12530/}}
}
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 ::