Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add cclank/lanshu-awesome-ai-video-kit --skill prompt-translatorgit clone --depth 1 https://github.com/cclank/lanshu-awesome-ai-video-kitWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/cclank/lanshu-awesome-ai-video-kit/prompt-translator)<a href="https://agentmods.dev/skills/cclank/lanshu-awesome-ai-video-kit/prompt-translator"><img src="https://agentmods.dev/badge/skills/cclank/lanshu-awesome-ai-video-kit/prompt-translator/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/cclank/lanshu-awesome-ai-video-kit/prompt-translator"><img src="https://agentmods.dev/badge/skills/cclank/lanshu-awesome-ai-video-kit/prompt-translator.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector pass
What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00208 | $0.03679 |
| Opus 5 | $0.00104 | $0.01840 |
| Sonnet 5 | $0.00042 | $0.00736 |
| Haiku 4.5 | $0.00021 | $0.00368 |
Grade A, and why
prompt-translator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 12d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 247 lines — stays where its author put it; the contents beside it link to each section on GitHub.
prompt-translator
跨模型提示词转换器。关键差异:不是凭 AI 直觉重写,而是查 110 条对照基准做 in-context learning。
何时不用此 skill
- 用户从零开始写一条新提示词(不是转换) → 用
seedance-prompter/kling-prompter/happyhorse-prompter或model-selector - 用户问"用哪个模型好" → 用
model-selector - 已有提示词出问题但不换模型 → 用
seedance-debugger
核心数据资产
prompts/data/cross-model-matrix.json — 这就是 translator 的"训练数据":
- 10 个核心场景:产品广告 / 情感重逢 / 滑板动作 / 图生视频 / 多人会议 / 恐怖悬疑 / 自然延时 / 抽象艺术 / 武侠决斗 / 萌宠爆款
- 每个场景 × 11 模型 = 110 条对照 prompt,每条严格遵循对应模型的官方公式
- 这构成"同一场景在 11 模型上的最佳写法对照",就是 translator 的查找表
工作流程
步骤 1:接收输入
最少需要:
- 源模型 (如 Sora 2 / Kling 3.0 / Wan 2.7)
- 源 prompt (用户的现有提示词)
- 目标模型 (用户想转到哪个)
可选:
- 转换偏好(更简洁 / 更详细 / 保留中文)
步骤 2:分析源 prompt 的语义内容
提取核心场景元素(与具体写法风格无关的):
- 主体(subject):是谁/什么
- 场景(scene):在哪/什么环境
- 动作(motion):发生了什么时序事件
- 情绪(mood):整体氛围
- 镜头(camera):怎么拍
- 音频(audio):需要什么声音
- 对白(dialogue):有无台词
- 风格(style):视觉锚点
这一步是剥离风格,提取语义。把 Sora 的 Style: → Cinematography: → Actions: 分层结构里的实际内容,抽象成"核心场景描述"。
步骤 3:查 110 条基准对照表找最相似场景
读取 prompts/data/cross-model-matrix.json,在 10 个场景里找与用户输入最相似的 1-2 个场景:
| 用户输入像... | 参考场景 |
|---|---|
| 产品旋转 / 静态主体特写 | scene-1-perfume |
| 双人对白 / 情感叙事 | scene-2-reunion |
| 户外动作 / 物理运动 | scene-3-kickflip |
| 图生视频(有参考图) | scene-4-i2v-cafe |
| 多人对话 / 室内会议 | scene-5-meeting |
| 恐怖悬疑 / 慢推进氛围 | scene-6-horror-balloon |
| 自然延时 / 无人景观 | scene-7-mountain-sunrise |
| 抽象艺术 / 流体特效 | scene-8-liquid-metal |
| 武侠 / 中式打斗 | scene-9-wuxia-duel |
| 萌宠 / 病毒短视频 | scene-10-surfing-dog |
步骤 4:基于相似场景的对照模式,做转换
在 prompt 里给 Claude 这样的 few-shot 模板:
我要把这条 [源模型] 的 prompt 转换成 [目标模型] 的最佳写法。
【参考对照】下面是一个相似场景在两个模型上的对照写法:
[源模型 in scene-N]:
{基准数据中该场景在源模型上的 prompt}
[目标模型 in scene-N]:
{基准数据中该场景在目标模型上的 prompt}
注意观察:
- 字段标签的变化 (e.g. "Cinematography:" → "Camera:" → "镜头:")
- 段落结构的变化 (e.g. 分层 → 5 层 → Entity+Scene+Motion+Sound)
- 措辞密度的变化 (e.g. 100 词 → 30 词 → 中文短句)
- 音频处理的变化 (e.g. "Background Sound:" → "Audio:" → "Sound:")
【用户的源 prompt】
{源 prompt}
【请输出】基于上面对照模式,将用户 prompt 转换为目标模型最佳写法。保留所有语义内容,仅调整结构/标签/措辞。
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 12d ago First seen · 247 lines · 208 tokens per session scan A 93f461a72823
prompt-translator is a skill published in the GitHub repository cclank/lanshu-awesome-ai-video-kit (392 stars, last pushed 3mo ago), licensed MIT. It adds 208 tokens to every session and 3,679 once invoked, about $0.0010 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
video-shotcraft
Create cinematic product videos from shot recipe cards, a validated template, and code/audio assets (Remotion + real page screenshots + 2.5D camera moves + beat-synced cuts + sound design). Use when the user asks to turn a frontend project or webpage into a product video, says "use video-shotcraft to make a…
cassette-video-edit
Edit, trim, cut, caption, subtitle, reframe, combine, add background music to, or export video, audio, and image files through Cassette. Use this skill whenever the user asks to change, preview, or render a media file in the project — even if they never say "Cassette" or name a tool — for example "trim the intro off…
seo-image-gen
AI image generation for SEO assets: OG/social preview images, blog hero images, schema images, product photography, infographics. Powered by Gemini via nanobanana-mcp. Requires banana extension installed. Use when user says "generate image", "OG image", "social preview", "hero image", "blog image", "product photo"…
vidseeds-setup
Use when connecting to VidSeeds.ai, getting AUTHREQUIRED or SUBSCRIPTIONREQUIRED from the vidseeds MCP server, or setting up VIDSEEDSPAT / OAuth for the connector. Not for workflow recipes - use vidseeds-efficiency and domain skills after connect.
vidseeds-local-video
Use when the user's video is on disk - VidSeeds MCP probe, frame/clip extraction, ffmpeg recipes, precision trim, and midroll analysis. Hosted tools return commands/recipes to run locally; media stays on the user's machine.
vidseeds-projects
Use for VidSeeds MCP project workflows - create a project from YouTube, list/get/snapshot, regenerate metadata (sync or async), translate metadata, apply thumbnails, and update per-platform config before publish.