The Agent Skills Directory

[COMMAND_EXECUTION]: The skill recommends using ffprobe to verify the duration and technical metadata of generated MP3 files. This is a standard and benign use of a common media analysis utility for output validation.
[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow for creating music generation prompts from user-provided details such as purpose, genre, and instrumentation. This creates an attack surface for indirect prompt injection if user inputs are not properly handled.
Ingestion points: User input fields for song purpose, duration, genre, and structure in SKILL.md and references/api-and-prompting.md.
Boundary markers: The skill does not explicitly define delimiters for user-supplied content within the generated prompts.
Capability inventory: The agent is authorized to perform audio generation via the Google Lyria API and execute local file analysis via ffprobe.
Sanitization: No specific filtering or escaping rules are defined for the user-provided music parameters.
[SAFE]: The skill contains multiple references to official technical documentation, pricing guides, and blog posts from well-known and trusted services, specifically within the google.com, deepmind.google, and google.dev domains.

lyria