The Agent Skills Directory

PROMPT_INJECTION (MEDIUM): The skill is vulnerable to Indirect Prompt Injection as it ingests untrusted data from the Spotify API which may contain malicious instructions embedded by an attacker (e.g., in a shared playlist name).
Ingestion points: Data enters via scripts/export_data.py and scripts/test_credentials.py through API calls like get_user_playlists, search_artists, and get_current_user.
Boundary markers: No explicit delimiters or boundary markers are used to separate API data from agent instructions.
Capability inventory: The skill has network access and the capability to read/write Spotify data (create/modify playlists).
Sanitization: No sanitization or filtering of external content for natural language instructions is performed.
DATA_EXFILTRATION (LOW): The skill initiates network connections to api.spotify.com. While necessary for its core purpose, this domain is outside the trusted whitelist.
DATA_EXFILTRATION (LOW): The scripts/export_data.py utility saves sensitive user information, including email, country, and private playlist details, to the local filesystem in unencrypted JSON files, increasing the local attack surface.

spotify-api