The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill downloads the Chart.js library from a well-known CDN (jsdelivr.net) to render the analytics dashboard. This is a standard practice for creating visual reports.
[COMMAND_EXECUTION]: The skill provides scripts for interacting with the Switchy GraphQL and REST APIs. These scripts use standard Python libraries (urllib) for network requests and do not execute arbitrary shell commands or untrusted external code.
[CREDENTIALS_UNSAFE]: The skill follows security best practices for credential management. It instructs users to store API tokens in environment variables or restricted local files (~/.switchy/token). It also uses a dedicated Gmail App Password for email functionality rather than primary account passwords, referencing a specific local path for retrieval.
[DATA_EXFILTRATION]: No evidence of unauthorized data exfiltration. Network communication is limited to the official Switchy API (api.switchy.io, graphql.switchy.io) for link management and analytics, and Google's SMTP servers for sending reports to the user. All data processed is related to the skill's primary purpose of link tracking.
[PROMPT_INJECTION]: The instructions contain clear boundaries and functional definitions without any attempts to override model safety guidelines or bypass constraints.

switchy-engine