tron2-mujoco-sim is a Python MuJoCo-based simulator for the TRON2 robot family (SF, WF, DA, DACH, and DASF variants). Its entry point, simulator.py, together with the tron2_sim/ variant registry, bridges the LimX low-level SDK simulator-side API to a MuJoCo simulation: it accepts RobotCmd messages (q, dq, tau, Kp, Kd) over the SDK's *ForSim transport and applies the resulting joint torques inside mujoco.MjData, publishing RobotState, ImuData, and (for gripper-equipped variants) gripper state back on the same bus.

The wire format and command semantics used by this simulator are identical to those used to drive the real TRON2 robot family. That means:

• Any process that can reach the SDK's simulator-side endpoint (loopback 127.0.0.1; the *ForSim API does not expose a configurable bind address) can push q / dq / tau / Kp / Kd targets. On a real robot, the equivalent SDK API and those same targets move physical actuators.
• A controller that misbehaves against the simulator will misbehave against the physical robot the moment it is pointed at the hardware-side SDK endpoint instead. That switch is the entire safety boundary between sim and real.
• Bugs in simulator.py / tron2_sim/ that cause it to accept malformed or out-of-range commands mask the same bugs in a downstream controller and can lead to unsafe real-robot behavior.
• The simulator applies the commanded PD/torque law as given and does not clamp torque beyond the actuator ctrlrange declared in the model. Treat it as a faithful mirror of the hardware interface, not as a safety net.

For those reasons this repository intentionally does not ship: control policies, calibration values, or a mode that silently forwards to a real robot. See THIRD_PARTY_NOTICES.md for the exclusion list.

Control-path vulnerabilities in the deployment stack itself belong to tron2-rl-deploy-python / tron2-rl-deploy-ros, not here.

<robot-ip> in this repository's Markdown / YAML command examples is a placeholder token, not a real address. Nothing in this repository (including simulator.py, tron2_sim/, CI, or configuration) hard-codes a private IP, and CI fails the build if one appears.

The internal-only sibling repository tron2-rl-deploy-ros retains a documentation-example literal 10.192.1.2 in its source / launch files (e.g. Tron2HW.cpp, tron2_hw_node.cpp, tron2_hw.launch), kept per owner decision and declared in that repository's SECURITY.md. That literal is documented there and is not mirrored into this repository.

Only the tip of the default branch and the most recent tagged release receive security fixes. Older tags are provided as-is.

Do not open a public issue for security reports.

Email: contact@limxdynamics.com Subject prefix: [tron2-mujoco-sim]

Please include:

• Affected file(s) and commit / tag (and, if relevant, submodule pin recorded by git submodule status).
• A minimal reproducer or proof of concept.
• Impact assessment (e.g., "simulator accepts NaN torque without clamping", "malformed RobotCmd crashes the physics thread").
• Whether the issue could affect a real robot if the endpoint were pointed at a physical unit.
• Your preferred disclosure timeline and contact.

We aim to acknowledge reports within 3 business days and provide a remediation plan or an initial mitigation within 14 calendar days. We support coordinated disclosure; please do not publish details until a fix or advisory is available.

• Bugs in third-party parsers or runtimes (MuJoCo, PyYAML) — report those upstream.
• Bugs inside the git submodules (robot-description, robot-joystick, limxsdk-lowlevel) — report those to the respective repositories.
• Physical safety of the robot itself — report to the deployment repositories or to LimX product support.
• Requests to publish calibration data, control policies, or firmware — this repository intentionally excludes those.

Good-faith security research that follows this policy will not be pursued legally by LimX Dynamics. Please respect user privacy, avoid service disruption, and do not access data beyond what is necessary to demonstrate the issue. Never point a proof-of-concept at a real robot's control endpoint — use the local simulator target.