matlab-plan-robot-motion

Pass

Audited by Gen Agent Trust Hub on Aug 2, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it processes untrusted external data such as robot configurations and environment geometry.
  • Ingestion points: The skill ingests user-provided robot configurations (startConfig, goalConfig), path waypoints, and workspace geometry (e.g., collision objects, point clouds) defined in the MATLAB workspace.
  • Boundary markers: The skill establishes boundaries by mandating the use of a separate kinematics validation skill (/matlab-model-robot-kinematics) and including explicit assert statements to verify the robot's integrity and data format before planning.
  • Capability inventory: The skill leverages the evaluate_matlab_code tool to execute complex motion planning algorithms (manipulatorRRT, manipulatorCHOMP) and generate trajectories.
  • Sanitization: The skill uses MATLAB's built-in assertion mechanisms and provides a workflow for post-calculation collision verification of the entire trajectory to ensure safety before execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 2, 2026, 12:12 PM
Security Audit — agent-trust-hub — matlab-plan-robot-motion