The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the unity command eval interface to execute C# code directly within the user's running Unity Editor. This is a high-capability operation necessary for modifying scene data that is not accessible via standard file edits.
[DYNAMIC_EXECUTION]: The agent generates and executes C# statement blocks at runtime. The skill provides specific templates in references/api.md and enforces constraints (such as qualifying types and avoiding using directives) to ensure these snippets remain valid and within the scope of the project's Assets directory.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the Unity project, including GameObject names and Audio Clip asset names, which are used to classify sounds and generate routing reports.
Ingestion points: UnityEngine.Object.FindObjectsByType and UnityEditor.AssetDatabase.FindAssets are used in references/api.md to collect names from the project scene and assets.
Boundary markers: The skill does not explicitly use boundary markers when processing these names, but treats them as string data within C# dictionaries.
Capability inventory: The skill has the capability to read project metadata and modify AudioSource components in the active scene via the eval command.
Sanitization: No explicit sanitization of object or asset names is performed. While names are used for classification logic, they are processed as string literals in the provided C# templates.

audio-setup-mixers