check-authoring
Pass
Audited by Gen Agent Trust Hub on Jun 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill provides legitimate templates and guidelines for developing Simulink Model Advisor checks. It follows platform best practices by enforcing the use of modern APIs such as DetailStyle and ResultDetail while explicitly advising against deprecated and unsafe patterns like StyleOne or StyleTwo.
- [COMMAND_EXECUTION]: The skill references the use of evalin and mtree in references/api-data-resolution.md and references/api-code-analysis.md. These are standard MATLAB tools for variable resolution and static code analysis within the model development environment. The skill correctly mandates a warning comment when using the experimental mtree API, mitigating risks associated with its usage.
- [PROMPT_INJECTION]: As a code-generation skill, it processes user requirements to produce MATLAB check files, representing an indirect prompt injection surface.
- Ingestion points: User intent and requirements for check logic are processed via instructions in SKILL.md.
- Boundary markers: None explicitly defined in the provided code templates for user input interpolation.
- Capability inventory: The skill generates code that uses platform-specific commands like set_param, get_param, and find_system to interact with model structures.
- Sanitization: The development workflow includes a mandatory validation step using check_matlab_code to verify all generated scripts before use.
Audit Metadata