The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The installation script scripts/install-mcp-server.sh clones a repository from an unverified GitHub account (marcelmarais/spotify-mcp-server) and executes npm install, which involves downloading and running code from various third-party packages.
[COMMAND_EXECUTION]: The skill uses shell commands to build and run the Spotify MCP server. This includes using scripts/apply-patch.js to dynamically modify the server's source code before compilation to inject environment variable support.
[EXTERNAL_DOWNLOADS]: The skill downloads code and dependencies from external repositories on GitHub and the NPM registry during the installation phase.
[CREDENTIALS_UNSAFE]: The skill requires the user to provide a Spotify REFRESH_TOKEN, which grants long-term, broad access to their account. Storing such credentials in an agent's environment or project files increases the risk of exfiltration.
[PROMPT_INJECTION]: The skill identifies a potential indirect prompt injection surface by instructing the agent to scan project files for credential patterns. Ingestion Point: Scanning context and files for .env blocks as described in SKILL.md. Boundary Markers: None identified. Capability Inventory: Network access via the Spotify API, file system access, and subprocess execution (node). Sanitization: No sanitization or validation of the scanned content is described.

controlling-spotify