The Agent Skills Directory

[SAFE]: The skill consists of extensive documentation and code examples for migrating deprecated MATLAB functions to modern equivalents across multiple domains (Signal Processing, Deep Learning, Image Processing, etc.).
[SAFE]: Explicitly warns against the use of eval, evalc, evalin, and str2num due to security risks and code injection vulnerabilities, recommending str2double and dynamic field names as secure replacements.
[SAFE]: Provides best practices for performance and memory management, such as array preallocation and vectorization, while discouraging the use of global variables and runtime path modifications.
[PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection as it is designed to process and analyze external MATLAB source code.
Ingestion points: MATLAB source files analyzed via the check_matlab_code tool.
Boundary markers: None explicitly defined in the instructions to separate code from instructions.
Capability inventory: The skill has access to evaluate_matlab_code, which could execute code derived from the modernized input.
Sanitization: The skill relies on standard tool execution boundaries and the agent's internal safety filters.
[SAFE]: Analysis of the 13 reference files and metadata found no evidence of obfuscation, hidden URLs, hardcoded credentials, or malicious dependencies.

matlab-modernize-code