The Agent Skills Directory

[SAFE]: The skill utilizes official internal APIs (library.LibraryConfig, library.BlockPolicy) for all configuration operations, ensuring that file writes and reads follow expected schemas.
[SAFE]: File system operations such as copyfile and rmdir are restricted to the .satk subdirectory within the user's preference directory or project root, representing standard configuration management behavior.
[SAFE]: Explicit guardrails are provided in the instructions, specifically prohibiting the use of introspection commands like find_system or load_system on library files, which mitigates risks associated with processing potentially malicious Simulink models.
[SAFE]: The skill enforces a human-in-the-loop workflow, requiring user confirmation for every configuration gate (declaration, policy, and indexing) before proceeding to the next step.

setup-custom-libraries