The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: Fetches and installs the core binary using the npx command from the vendor's npm package @mvanhorn/printing-press during the setup phase.
[DATA_EXFILTRATION]: Provides a --deliver webhook:<url> flag that allows the agent to POST command output (which may contain user profile information, library details, or listening history) to arbitrary external URLs.
[DATA_EXFILTRATION]: Includes a feedback command capable of transmitting data to a remote endpoint if the SPOTIFY_FEEDBACK_ENDPOINT environment variable is configured.
[CREDENTIALS_UNSAFE]: Provides instructions for the manual configuration and export of sensitive SPOTIFY_CLIENT_ID and SPOTIFY_SECRET credentials and persists authentication tokens in the local file ~/.config/spotify-pp-cli/token.json.
[PROMPT_INJECTION]: Indirect Prompt Injection Surface: The skill ingests untrusted data from the Spotify Web API (such as playlist names, track metadata, and artist descriptions) which could contain instructions designed to influence the agent's behavior.
Ingestion points: Processes external content via commands like me get-users-top-tracks, search, and playlists get.
Boundary markers: Absent. No instructions are provided to the agent to distinguish between music metadata and operational instructions.
Capability inventory: The toolset includes file system writes (--deliver file) and network operations (--deliver webhook), which could be abused if an injection is successful.
Sanitization: No sanitization or validation of ingested metadata is mentioned before it is interpolated into subsequent command arguments or agent responses.

pp-spotify