The Agent Skills Directory

[SAFE]: The skill performs expected project analysis and code generation tasks. It identifies project components like source directories, test suites, and MEX source files to populate a build configuration template.
[SAFE]: Data processing is localized to the project environment. It reads configuration from standard project files such as toolboxPackaging.prj and toolboxSpecification.m to ensure the build pipeline reflects the project's identity and versioning.
[SAFE]: The skill utilizes built-in MATLAB tasks (CodeIssuesTask, CleanTask, TestTask, MexTask) for core functionality, which is a recommended security and stability practice.
[DATA_EXPOSURE]: As an architectural property, the skill scans the local file system to determine project structure. However, this access is restricted to the project root and is used solely for generating build automation instructions.
[INDIRECT_PROMPT_INJECTION]: The skill has an ingestion surface as it processes untrusted data from the local project structure (folder names, file contents) to interpolate into the generated buildfile.m. While this is a standard behavior for a generator, it represents the primary attack surface. The skill includes basic sanitization, such as replacing spaces with underscores in output filenames.

matlab-create-buildfile