菜单
Burn TikTok / Hormozi-style captions onto any vNext video (R2-hosted). Auto-transcribes via Whisper when transcript is omitted. Styles: hormozi (default), tiktok, minimal.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/gitroomhq/agent-media-app --skill make-subtitles命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Silent 5-10s extreme close-up reaction clip with a static caption overlay ("the crazy look"). One recurring character opens on an exaggerated look and keeps morphing through randomized silent beats — no speech, no subtitles, ambient room tone only. A SERIES MUST START WITH A CHARACTER (create_character first; the saved sheet + pinned seed keeps the SAME face on every clip). Volume workflow: same caption + same character, N calls, N different performances.
Playbook for Agent-Media UGC Video — the one tool for UGC video on agent-media. Always call the single make_ugc skill: give it a `script` (any length) and optionally a person/image/character; it returns the finished captioned vertical video. Short script → one clip, long monologue → full multi-take (never trimmed), `broll_url` → narrated overlay. You never pick a sub-skill.
AI UGC video production via the agent-media CLI (selfie, character, subs). BEFORE running ANY agent-media command you MUST first Read reference/conversation-flow.md and walk the user through the 4 gates IN ORDER — (1) confirm duration first with script-pacing guidance (~2.5 words/sec), (2) confirm script OR scene_action, and if there is no speech also propose background_music, (3) RUN `agent-media character list --json` YOURSELF (do not ask the user, do not mention char_xxx ids — present saved characters BY NAME if any, otherwise confirm the new description), (4) propose a director brief with setting, lighting, wardrobe, props/product, and action, passing non-default motion/product handling through --scene-action. AUTOPILOT exception: if the user explicitly says "autopilot", "just do it", or "run with defaults", proceed end-to-end without re-asking every gate, infer missing details reasonably, keep safety constraints, and report assumptions. While jobs run, poll status and open portrait, character sheet, wire
基于 SOC 职业分类
正在显示 SKILL.md
| name | Make Subtitles |
| description | Burn TikTok / Hormozi-style captions onto any vNext video (R2-hosted). Auto-transcribes via Whisper when transcript is omitted. Styles: hormozi (default), tiktok, minimal. |
| allowed-tools | ["mcp__agent-media__make_subtitles"] |
| x-skill-slug | make_subtitles |
| x-skill-version | 1.0.0 |
| x-primitive | subtitles_v2 |
| x-mcp-tool | mcp__agent-media__make_subtitles |
Burn TikTok / Hormozi-style captions onto any vNext video (R2-hosted). Auto-transcribes via Whisper when transcript is omitted. Styles: hormozi (default), tiktok, minimal.
Call this skill when the user asks for the outcome described above. It runs on the agent-media vNext primitive runtime via the mcp__agent-media__make_subtitles MCP tool. Authentication is the user's existing agent-media Bearer token (issued by agent-media login).
Preferred path: MCP tool mcp__agent-media__make_subtitles. Schema is auto-published via tools/list against the same MCP server, so don't restate the schema here — trust the server's response.
Fallback path: REST.
POST https://api.agent-media.ai/v1/skills/make_subtitles/run
Authorization: Bearer $AGENT_MEDIA_API_KEY
Content-Type: application/json
Idempotency-Key: <any unique string per intent>
{
"video_url": "https://pub-...r2.dev/vnext/primitive-runs/<id>/simple-selfie.mp4",
"style": "hormozi"
}
1520sGET https://api.agent-media.ai/v1/primitives/runs/<run_id>
Authorization: Bearer $AGENT_MEDIA_API_KEY
agent-media login.This file is auto-generated by scripts/generate-public-skill.ts from the registry at services/api-v2/src/skills/registry.ts. Do not hand-edit; CI rejects drift.