The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from the Unity environment which could contain malicious instructions.
Ingestion points: The skill reads object names and component properties from the active Unity scene using UnityEngine.Object.FindObjectsByType in resources/audio-import-api.md and Section 2 of SKILL.md.
Boundary markers: None identified. The skill does not use specific delimiters or instructions to ignore potential commands embedded in object names.
Capability inventory: The skill uses unity command eval to execute C# snippets that can modify project assets, reimport files, and change editor settings (SKILL.md Section 0 and 4).
Sanitization: No sanitization or validation of scene object names is performed before they are used in string-concatenated C# snippets.
[DYNAMIC_EXECUTION]: The skill generates and executes C# code at runtime to interact with the Unity Editor API.
Evidence: SKILL.md describes the execution path using unity command eval --code '<snippet>'. The snippets are constructed from templates in resources/audio-import-api.md and executed in the Unity Editor process.
[COMMAND_EXECUTION]: The skill uses the unity-cli tool to execute commands in the local environment.
Evidence: SKILL.md specifies the use of unity command eval and unity command --format json to perform its primary functions.

optimize-audio