Update to Visual Sandbox

{
  "description": "Make a creator-style UGC video ad for a product with the Visual Sandbox MCP tools: a vertical 9:16 clip of one person who shows, uses or talks about the product, for TikTok, Instagram Reels or YouTube Shorts. Trigger when the person asks for a \"UGC ad\", \"creator ad\", \"testimonial video\", \"unboxing video\", \"talking head ad\", \"TikTok ad for my product\" or \"a person holding my product\". Plans a script, approves the first frame as an image, then animates it with speech. Uses the MCP tools only.\n",
  "included_files": [
    {
      "relative_path": "agents/openai.yaml",
      "size_in_bytes": 280
    }
  ],
  "name": "vsb-ugc-ad",
  "skill_md_contents": "---\nname: vsb-ugc-ad\ndescription: >\n  Make a creator-style UGC video ad for a product with the Visual Sandbox MCP\n  tools: a vertical 9:16 clip of one person who shows, uses or talks about\n  the product, for TikTok, Instagram Reels or YouTube Shorts. Trigger when\n  the person asks for a \"UGC ad\", \"creator ad\", \"testimonial video\",\n  \"unboxing video\", \"talking head ad\", \"TikTok ad for my product\" or \"a\n  person holding my product\". Plans a script, approves the first frame as an\n  image, then animates it with speech. Uses the MCP tools only.\nlicense: Proprietary. See LICENSE for the terms.\n---\n\n# A UGC video ad\n\nA UGC ad looks like a real person filmed it on a phone. One person, one\nproduct, one message, 9:16, 5 to 15 seconds. The person talks to the camera,\nor shows the product while a voice talks.\n\nUse the MCP tools only. Video is the expensive part, so this skill makes the\nlook right as a still image first, and spends on video once.\n\n## 1. Get the brief\n\nAsk only for what is missing. Never ask more than three questions.\n\n- **The product.** A photo is best. Pass a public `https` URL. Give a photo\n  attached in the chat, or a local file, to `upload_media` first.\n- **The hook.** The one thing the ad must say, such as \"it fixed my skin in\n  a week\".\n- **The format.** Pick one: a talking head, a selfie walk-and-talk, an\n  unboxing, or a before-and-after. When the person does not say, use a\n  talking head.\n\n## 2. Write the script\n\nWrite 2 to 4 short lines that one person can say in the clip length. About\n2.5 words fit in one second.\n\n- Line 1 is the hook. It names the problem or the result.\n- The middle line shows the product in use.\n- The last line tells the viewer what to do.\n\nDo not use dashes in the spoken lines. A speech model reads a dash badly.\nWrite numbers as words. Show the script to the person before you spend.\n\n## 3. Make the first frame as an image\n\n1. Read [`vsb-ugc-people`](../vsb-ugc-people/SKILL.md) with `get_skill`. It\n   has the two UGC formats, the slot order and the words that stop a face\n   from looking plastic.\n2. Call `list_models` with `modality: \"image\"` and pick the model that pack\n   recommends. Call `get_model` for its fields.\n3. Call `generate` at 9:16 with the product photo as a reference. Name the\n   work with `sandbox_name`, such as \"Serum UGC ad\", and send that same name\n   on every later call.\n4. Call `get_job` after `eta_seconds` and look at the frame. Check the hands,\n   the label on the product and the face.\n5. Show the frame. Change it until the person approves it.\n\n## 4. Animate it with speech\n\n1. Read [`vsb-video`](../vsb-video/SKILL.md) with `get_skill`, at the section\n   \"People who talk\". It says which model speaks in one pass and which needs\n   a voice track first.\n2. Call `list_models` with `modality: \"video\"`. Pick an image-to-video model\n   that speaks, from that section.\n3. Call `get_model`, then `estimate_cost`. Tell the person the price and wait\n   for a yes.\n4. Call `generate` with the approved frame's URL as the start image and the\n   script as the spoken line.\n5. Call `get_job` after `eta_seconds`. A video takes one to three minutes.\n   `get_job` returns four frames, so check them before you report.\n\nWhen the person wants one voice across several clips, make the voiceover\nfirst with [`vsb-audio`](../vsb-audio/SKILL.md) and follow the lip-sync path\nin `vsb-video`.\n\n## 5. Hand it over\n\nGive the person the `share_url` of the video and the `sandbox_url` of the\ncanvas. Offer one next step: a second hook on the same frame, which costs one\nvideo run and no new image.\n\n## Rules\n\n- One approved frame before any video run. A video costs many times more\n  than an image.\n- Never run a video twice to check it. Read the four frames `get_job` sends.\n- Keep the product label sharp. If the model changes the label, say so and\n  offer a new frame, not a new video.\n- When a run is refused for want of credit, say the balance is too low, and\n  follow what the server instructions say about credit.\n"
}