Skip to content

Latest commit

 

History

History
66 lines (51 loc) · 2.94 KB

File metadata and controls

66 lines (51 loc) · 2.94 KB

Skill: ianotifier - Desktop Notifications for AIs

This skill allows AI agents and assistants to effectively notify the user upon task completion using visual desktop notifications (GNOME notifications), acoustic beeps, and Text-to-Speech (TTS).

Important

  • Language Preference: Always compose the notification message and speak to the user in their preferred language (e.g., if the user communicates in Spanish, write the notification text and voice message in Spanish).
  • Proactive Trigger: You MUST trigger this notification automatically as soon as you finish a long-running task (such as refactoring, test suites, compilations, or long-running scripts), without waiting for the user to explicitly ask you to notify them. This saves tokens by avoiding unnecessary back-and-forth prompts.

1. When to Trigger This Skill

Automatically trigger this skill when:

  • You complete a long-running, multi-step task.
  • A critical error occurs that requires human intervention.
  • You finish compiling, testing, or running a code generation pipeline.

2. Command Line Interface (CLI) Usage

When operating in an interactive terminal (such as agy or claude-cli), run the ianotifier command in Bash. The command automatically detects the host or container network configuration.

Basic Syntax:

ianotifier [options] "Message to display and speak"

Common Examples:

# 1. Success notification (matches the user's language)
ianotifier --sender "Antigravity" --status "success" "He completado la compilación y pasé las pruebas con éxito."

# 2. Critical error notification (triggers error beep and critical priority popup)
ianotifier --sender "Claude" --status "error" "Error al compilar el proyecto. Revisa los logs de test."

# 3. Simple quick notification
ianotifier "All code analysis tasks completed successfully."

CLI Options:

  • message (Positional, required): The notification message text to display and speak.
  • -s, --sender (Optional): Your AI agent name (e.g., "Claude", "Antigravity").
  • -st, --status (Optional): "success" (default), "error", "warning", or "info". Adjusts beep sound and GNOME notification urgency.
  • -l, --lang (Optional): "auto" (default), "es", "en". Forces TTS pronunciation in the specified language.

3. Model Context Protocol (MCP) Server Usage

If connected via an MCP client (such as Roo-Cline or Cline in VSCode), invoke the notify tool.

notify Tool Parameters:

  • message (string, required): Notification text.
  • sender (string, optional): Your AI agent name (e.g., "Claude", "Antigravity").
  • status (string, optional): "success", "error", "warning", "info".
  • language (string, optional): "auto", "es", "en".

Example Tool Call:

{
  "name": "notify",
  "arguments": {
    "message": "He finalizado el refactor del código y todo funciona perfectamente.",
    "sender": "Antigravity",
    "status": "success"
  }
}