The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes the curl utility to download audio files from dynamic remote URLs generated by the music creation tool.\n- [CREDENTIALS_UNSAFE]: The skill handles a sensitive authentication token ($COGMUSIC_TOKEN) within a shell command. While using environment variables for secret management is a best practice, passing them via command-line arguments in shell instructions can lead to potential exposure in system process logs.\n- [INDIRECT_PROMPT_INJECTION]: The skill's workflow presents a surface for indirect prompt injection by processing external data and using it in shell-based operations.\n
Ingestion points: User-provided creative briefs used as instructions for the music generation tool.\n
Boundary markers: Absent; there are no clear delimiters or instructions to the model to ignore potential injection content within the user brief or the tool's response.\n
Capability inventory: Shell execution (curl), external tool invocation (create_production), and persistent data storage (memory_store).\n
Sanitization: No explicit validation or escaping of the audio_url variable is defined before it is used in a shell context.

music-generate