# The prompt

First give your agent the hyviral-studio skill. It works inside an edit made through the HYVIRAL connector, so add
the connector too: in Claude, Settings → Connectors → Add custom connector → `https://api.hyviral.com/mcp`. Then
replace everything in [BRACKETS] below and paste it in. Anything you leave in brackets is your agent's call.

---

Use the hyviral-studio skill to design the motion graphics and sound effects for my edit, and place them in it
through the HYVIRAL connector.

- **The video:** [VIDEO - e.g. "my 40-minute interview, already uploaded", or "the edit we're working on"]
- **What it's about:** [TOPIC - e.g. "sourdough for beginners"]
- **Pieces I want:** [PIECES - e.g. "a chapter opener per chapter, a stat for each number I say, a name for each guest", or "your call"]
- **Look:** [LOOK - e.g. "dark with one lilac accent", or my brand colors as hex codes]
- **Fonts:** [FONTS - a font file I'm licensed to use in video, or "the kit's defaults"]
- **Sound:** [SOUND - e.g. "soft clicks and whooshes, nothing loud", or "your call"]

Work in this order:

1. Run the skill's capability check: fetch the kit, render the two-second proof in my colors and complete one
   upload. If a step fails, tell me what is missing and which route you'll take instead.
2. Pick one look for the whole edit from the video's topic, and show me the plan: every piece, the word it lands
   on and its sound. Wait for my OK.
3. Build each piece as code, render it, check its frames, upload it and place it on its word. Make the sound
   effects too, and place one under each entrance.
4. Whenever the connector asks for my yes (it does before anything generated and the export), show me what it is
   and wait.
5. Preview the parts you touched, then tell me what you made, what HYVIRAL rendered, and what I can still change.
