Menú
Build enterprise AI ad asset systems and batch AI UGC video production briefs. Use when a user asks for AI UGC ads, product demo videos, ad asset factories, prompt packs, or reusable video production workflows.
Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.
Basado en la clasificación ocupacional SOC
| name | ai-ugc-factory |
| description | Build enterprise AI ad asset systems and batch AI UGC video production briefs. Use when a user asks for AI UGC ads, product demo videos, ad asset factories, prompt packs, or reusable video production workflows. |
| category | media |
| symbolName | video.badge.waveform |
| seedVersion | 2 |
Use this skill to turn a product or service into a repeatable AI UGC ad production system, not a one-off prompt.
使用这个 skill，把一个产品或服务转成可重复运行的 AI UGC 广告生产系统，而不是一次性 prompt。
Produce a compact package:
输出一个轻量生产包：
Read the product website, source files, screenshots, and user notes. Separate verified facts from assumptions. Do not invent prices, discounts, certifications, bonuses, or performance claims.
读取产品官网、源文件、截图和用户说明。区分已验证事实和推测。不要编造价格、折扣、认证、赠品或效果承诺。
Rename or label assets by visual function:
按视觉功能重命名或标记素材：
hero-product-image｜主产品图packaging-detail｜包装/细节lifestyle-context｜生活场景feature-proof｜功能证明website-cta｜网页/CTAcreator-reference｜真人/创作者参考style-reference｜风格参考contact-asset｜联系/转化资产The manifest should say which assets must be actual model inputs. A path in a prompt is not proof that the model read the image.
素材清单必须说明哪些素材需要作为真实模型输入。只在 prompt 里写路径，不等于模型真的读取了图片。
Pick 1-3 lanes that fit the product:
根据产品选择 1–3 条视频生产路线：
Every 1-2 seconds should carry a job: hook, product proof, pain point, benefit, objection, credibility cue, visual reset, CTA.
每 1–2 秒画面都应承担一个明确任务：钩子、产品证明、痛点、利益点、异议、可信度、视觉重置或 CTA。
For every segment, state:
每个片段都要写清：
Before final video generation, confirm any user-owned style decisions: mascot/character identity, voiceover tone, rhythm, CTA/contact asset, and whether exact captions should be post-composited.
最终生成前，确认所有用户拥有的风格决策：角色/吉祥物、口播语气、节奏、CTA/联系资产，以及精确字幕是否需要后期合成。
Require per segment:
每个片段必须检查：
MODEL_INPUT_IMAGES_ACTUALLY_ATTACHEDPRODUCT_REFERENCE_WAS_IMAGE_INPUTPROMPT_ONLY_PRODUCT_REFERENCE_USEDPRODUCT_VISIBLE_AND_MATCHES_REAL_ASSETIf product fidelity or text quality fails in client-facing shots, request a narrow correction pass instead of accepting the render.
如果客户可见镜头中的产品一致性或文字质量失败，应要求窄范围返修，而不是直接接受成片。
User: “帮我为这个产品批量生产 AI UGC 广告视频，并沉淀成可复用 workflow。”
Response shape｜响应结构：