matlab-simulink-agent

Pass

Audited by Gen Agent Trust Hub on Jun 24, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTIONDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill facilitates the execution of MATLAB code and shell scripts. It includes a batch runner (scripts/run_matlab_batch.sh) that executes generated .m scripts within the local MATLAB environment. The shell script includes logic to sanitize paths and prevent escaping string literals in the MATLAB command string.
  • [EXTERNAL_DOWNLOADS]: The skill provides instructions for the agent to look up official MATLAB documentation and community forums (e.g., MathWorks Answers, Stack Overflow) to resolve errors. This is handled as a documentation retrieval task for technical troubleshooting.
  • [PROMPT_INJECTION]: As the skill ingests data from external troubleshooting sources and documentation, there is a theoretical risk of indirect prompt injection. However, the skill includes explicit 'Trust filters' in references/community-troubleshooting.md to reject advice that bypasses security or licensing.
  • [DATA_EXFILTRATION]: The skill collects environment metadata (MATLAB version, products, etc.) via scripts/matlab_capability_scan.m and scripts/simulink_inventory.m. This information is written to local JSON files for use in the automation workflow and is not exfiltrated to external servers.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 24, 2026, 06:46 AM
Security Audit — agent-trust-hub — matlab-simulink-agent