Eval injection enables arbitrary Python code execution in Open Source Robotics Foundation Robot Operating System (ROS) rosbag

Note: This data is for reference and cybersecurity research purposes only.CyStack advises users not to use this information for unlawful purposes.

What is CVE-2025-3753?

CVE-2025-3753 is a vulnerability classified as Improper Neutralization of Directives in Dynamically Evaluated Code and Improper Control of Generation of Code, affecting Robot Operating System (ROS) (affected versions: Noetic Ninjemys, Melodic Morenia, and other affected versions). This vulnerability is rated High, with a CVSS score of 7.8. Current sources do not report this vulnerability as exploited.

Overview

Original source data

A code execution vulnerability has been identified in the Robot Operating System (ROS) 'rosbag' tool, affecting ROS distributions Noetic Ninjemys and earlier. The vulnerability arises from the use of the eval() function to process unsanitized, user-supplied input in the 'rosbag filter' command. This flaw enables attackers to craft and execute arbitrary Python code.

Affected products and scope

  • Open Source Robotics Foundation Robot Operating System (ROS), package rosbag: Noetic Ninjemys, Melodic Morenia, Kinetic Kame, and Indigo Igloo are marked affected.
  • The vulnerability description also states Noetic Ninjemys and earlier, while the structured affected data explicitly names the four branches above. Do not extend the affected scope to other branches beyond what the record confirms.
  • The product data sets the default status to unaffected for versions not explicitly listed, but no version-specific fixed release is supplied.
  • No fixed ROS 1 release is identified in the supplied remediation data.

Technical details

The rosbag tool passes unsanitized user-supplied input from the rosbag filter command to Python's eval() function. This is eval injection, classified as CWE-95, and also code injection under CWE-94. The recorded attack path is local, requires no prior privileges, and requires user interaction with the input-processing workflow. Successful exploitation could run arbitrary Python code in the context of the process invoking rosbag; the available evidence does not identify the precise source location, process identity, or whether sandboxing is present.

Exploitability

  • The recorded reachability is local; the available evidence does not establish direct network exploitation.
  • No prior privileges are required according to the record, but a user or operator must interact with a workflow that processes attacker-crafted input.
  • The recorded attack complexity is low and user interaction is required.
  • The record does not identify a public exploit or observed exploitation status; the known_exploited field contains no exploitation record.

Technical impact

  • Successful exploitation could execute arbitrary Python code with the privileges of the process running rosbag, potentially affecting the confidentiality, integrity, and availability of local resources.
  • The record states that no prior privileges are required, but user interaction is required and the attack path is local.
  • Impact across a separate security authority is not established; the available evidence also does not say whether the process runs in a sandbox or under a privileged account.
  • Organisational consequences could include corrupted robotics data, disrupted processing pipelines, or the need to isolate legacy ROS systems during migration.

Business impact

  • Arbitrary Python code could read, modify, or delete data accessible to the rosbag process.
  • If the process is used in a robotics data-processing workflow, code execution could corrupt data, alter processing behavior, or disrupt a local service.
  • The practical consequence depends on the account running the tool, the data available on the host, and the degree of environmental isolation.
  • Because the recorded path is local and requires user interaction, the flaw is not described as a direct unauthenticated network entry path.

Remediation

  1. Migrate from all ROS 1 versions to a ROS 2 version, as directed by the supplied solution. The record states that all ROS 1 versions are EOL.
  2. Inventory the four affected branches and prioritize systems that still use rosbag filter with user-supplied or externally sourced input.
  3. Until migration, avoid processing untrusted input with rosbag filter and restrict the workflow to controlled data sources. This is a precaution, not a vendor-confirmed workaround.
  4. Do not treat moving to another ROS 1 branch as a confirmed fix. Although the record marks a patch as available, the supplied data does not name a fixed ROS 1 release.

Detection

  • Inventory systems using Robot Operating System (ROS), the rosbag package, and tools or jobs that invoke rosbag filter.
  • Identify the deployed distribution branches and determine whether rosbag filter receives expressions from files, pipelines, automation, or external sources.
  • Review scripts, job definitions, and process history for rosbag filter invocations, then verify where the filter expression originated.
  • The record provides no specific log event or IOC. Monitoring for abnormal behavior from the process invoking rosbag is a precautionary measure, not a confirmed indicator.
Sources (24)
Learn more

Run an in-depth assessment with complete web risk management

CyStack VulnScan continuously discovers assets, validates vulnerabilities, and helps security teams prioritize remediation across the organization.