The Agent Skills Directory

[SAFE]: No malicious patterns such as prompt injection, persistence mechanisms, or unauthorized privilege escalation were detected.
[COMMAND_EXECUTION]: The skill utilizes a bash script (scripts/acestep.sh) to interact with the music API. This script employs standard command-line tools including curl for network requests, jq for robust JSON processing, and python3 for base64 encoding/decoding of audio data.
[EXTERNAL_DOWNLOADS]: The skill connects to the ACE-Step API endpoints (defaulting to http://localhost:8001 or the official vendor domain https://api.acemusic.ai) to transmit generation tasks and download synthesized audio files. These operations are restricted to the music generation functionality described in the skill's purpose.
[CREDENTIALS_UNSAFE]: The skill manages authentication using the ACE_MUSIC_API_KEY environment variable. The implementation includes explicit logic to prevent the accidental printing or logging of this secret, adhering to secure credential management practices.
[DATA_EXFILTRATION]: No evidence of unauthorized data access or exfiltration was found. File system operations are limited to reading source audio provided by the user and writing generated assets to a dedicated acestep_output directory within the project root.

acestep