audio-reply-skill_skill

Overview

This skill generates spoken audio replies using a TTS engine. Trigger it with phrases like “read it to me [public URL]” to fetch and read web content aloud, or “talk to me [topic]” to produce a conversational audio response. It supports short, natural-sounding speech and auto-cleans temporary files after playback.

How this skill works

The skill validates and (when requested) fetches only public http/https URLs, extracts readable text, and summarizes long pages before conversion. For conversational prompts it generates concise spoken text, then calls a TTS command to produce and play audio files. Temporary audio files are deleted immediately after playback and safety checks block private or credentialed links.

When to use it

• You want a quick spoken summary of a public article or page.
• You prefer spoken conversational answers instead of text replies.
• Hands-free listening while multitasking or driving (short pieces).
• Testing TTS voice, speed, or expressiveness for demos.
• Converting small site excerpts into audio for accessibility checks.

Best practices

• Only provide public http/https links; do not include private or intranet URLs.
• Keep requests concise—TTS works best with shorter segments (under ~200 words for conversation).
• If a page is long, ask for a specific section or allow the skill to summarize aggressively.
• Avoid sending links with credentials, API keys, or signed tokens; redact sensitive data first.
• Confirm playback environment (volume, speakers) since --play uses system audio; always clean up temp files.

Example use cases

• User: “read it to me https://example.com/article” — skill fetches, extracts main text, summarizes if needed, and plays audio.
• User: “talk to me about healthy breakfast ideas” — skill generates a short conversational reply and plays it.
• Developer testing TTS parameters like speed or exaggeration by requesting sample phrases to be spoken.
• Accessibility check: convert blog posts into brief audio segments for review.