The Agent Skills Directory

[SAFE]: The skill implements explicit 'Hard rules' in its instructions and code to prevent automated purchases, protecting the user from unintended financial transactions. Specifically, the plan command calculates costs but explicitly stops before any checkout process.
[SAFE]: Authentication is handled with security best practices for AI agents. It explicitly instructs the agent never to handle passwords, instead relying on short-lived bearer tokens or a standard PKCE OAuth flow. Tokens are stored in a dedicated subdirectory with restricted file permissions (0o600).
[SAFE]: Network operations are restricted to established official Beatport API endpoints (api.beatport.com). The skill includes logic to detect and avoid Cloudflare-protected web pages that are not script-friendly, preventing the agent from getting stuck or leaking data to unintended scrapers.
[SAFE]: The code uses only the Python standard library for its operations (urllib, http.server, hashlib, etc.), avoiding supply chain risks from third-party dependencies. Dynamic execution is limited to a regression test script that loads the skill module locally for validation purposes.
[SAFE]: The 'organize' and 'report' features perform local file organization and harmonic matching based on music theory (Camelot wheel). These operations are confined to the music directories specified by the user and do not access sensitive system files.

beatport-skill