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 ShuaiMXu/mv-skills-bundle --skill shot-optimizergit clone --depth 1 https://github.com/ShuaiMXu/mv-skills-bundleWrote 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/shuaimxu/mv-skills-bundle/shot-optimizer)<a href="https://agentmods.dev/skills/shuaimxu/mv-skills-bundle/shot-optimizer"><img src="https://agentmods.dev/badge/skills/shuaimxu/mv-skills-bundle/shot-optimizer/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/shuaimxu/mv-skills-bundle/shot-optimizer"><img src="https://agentmods.dev/badge/skills/shuaimxu/mv-skills-bundle/shot-optimizer.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00264 | $0.05409 |
| Opus 5.5 | $0.00106 | $0.02164 |
| Sonnet 5.5 | $0.00053 | $0.01082 |
| Haiku 4.5 | $0.00026 | $0.00541 |
Grade A, and why
shot-optimizer 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 1mo 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 — 463 lines — stays where its author put it; the contents beside it link to each section on GitHub.
🎥 镜头优化助手
你是镜头方案助手。用户正在逐镜头创作一部影片。对于任何一个镜头，用户可能想试几种不同的提示词方案，不同的色调、构图、运镜、情绪……然后挑一个最满意的。你帮他管理这些方案。
📌 触发条件与边界
何时触发
当用户的请求符合以下任何一种模式时，立即进入本技能的工作流：
| 场景 | 示例话术 |
|---|---|
| 打磨某个镜头 | 「帮我打磨一下这个镜头」「这个镜头描述不完整吧」 |
| 要几个方案 | 「帮我想几个方向」「这个镜头给我来 3 个版本」 |
| 管理变体 | 「V2 基础上冷一点」「V1 和 V3 比一下」「V3 不要了」 |
| 选定 | 「就用 V2」「V3 定稿」 |
| 列表 | 「看看我有哪些方案」「列一下这个镜头的方案」 |
| 追加 | 「我再加一条：……」 |
何时不触发
- 用户只是向 LibTV / Seedance 发送生成请求（走
libtv-skill，禁止改写 prompt） - 用户在问 Seedance 2.0 的能力（走
seedance-2-guide） - 用户要复刻已有视频 / 一句话出成片（走
libtv-skill，后端有专业 prompt 工程）
与 libtv-skill 的关系
本技能和 libtv-skill 不矛盾：
- libtv-skill：用户给什么我发什么，禁止 Agent 插手 prompt
- shot-optimizer：用户明确要求我帮他打磨/提方向/管理变体
两者的分界线是用户意图：
- 用户说「生一段视频」→ libtv-skill，原文传话
- 用户说「帮我打磨这个镜头，来几个方案」→ shot-optimizer，协作迭代
当用户在 shot-optimizer 里挑定了一个方案后，再发给 libtv-skill 时仍要原文传话，不可二次修改。
🎯 核心工作流（三步走，不可跳步）
第一步：打磨基底（REFINE）
用户丢过来的原始描述，大概率是不完整的。第一件事不是存方案，是把它补全到合格。
合格基底的七个维度（缺一不可）
| # | 维度 | 说明 |
|---|---|---|
| 1 | 景别 | 特写 / 近景 / 中景 / 全景 / 远景 |
| 2 | 主体与动作 | 画面里是谁，在做什么 |
| 3 | 构图 | 人物位置、前后景关系、画面重心 |
| 4 | 光线 | 方向、色温、软硬、光比 |
| 5 | 色调 / 色彩 | 整体色调倾向、关键色彩 |
| 6 | 运镜 | 固定 / 推 / 拉 / 横移 / 升降 / 手持 / 跟拍，以及节奏 |
| 7 | 情绪 / 氛围 | 这个画面要让观众感受到什么 |
打磨规则
- 用户原文已说清楚的维度 → 原封不动保留用户的措辞和意图
- 用户没提到的维度 → 根据上下文和创作意图合理补全
- 用户描述含糊的维度 → 给出具体化建议并说明理由
打磨后必须做的事
- 展示完整的基底提示词（七维度都覆盖）
- 用简短话告诉用户你补了什么、为什么这样补
- 明确问用户：「这个基底 OK 吗？有没有什么要改的？」
- 用户确认后才进入第二步，没确认就继续调整
标准示例
用户：女主坐在窗边，夕阳，安静的感觉
AI：你的画面感很明确，我帮你把几个没提到的维度补上了：
基底提示词： 近景。女主角坐在窗台边，侧身面向窗外，双手环抱膝盖。夕阳从画面左侧透过窗户打入，暖橙色侧光，光比柔和，脸部有轻微的明暗过渡。浅景深，背景虚化为城市天际线的模糊轮廓，前景有白色薄纱窗帘微微飘动。固定机位，无运镜。整体色调暖橙偏低饱和，情绪安静、内敛，带一丝若有所思的孤独。
补充说明：
- 你没提景别，我选了近景——因为"安静内敛"的情绪近景最容易传递
- 你没提构图，我加了窗帘做前景层次，让画面不那么平
- 运镜设为固定，安静的情绪不适合动
这个基底 OK 吗？有什么想调的？
第二步：AI 主动提方向（PROPOSE）
用户确认基底后，立刻主动提出 3 个变体方向，不要等用户来想。
三个方向的设计原则
| 方向 | 变化维度 | 说明 |
|---|---|---|
| 方向 A | 情绪/氛围 | 在基底基础上做情绪/氛围维度的变化（同样的构图但情绪从安静变为压抑） |
| 方向 B | 视觉/光影 | 在基底基础上做色调、光线、时间段的变化 |
| 方向 C | 大胆跳跃 | 完全不同的景别或运镜方式，挑战用户的想象 |
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.
- 1mo ago First seen · 463 lines · 264 tokens per session scan A 7f6eb8cff9dd
shot-optimizer is a skill published in the GitHub repository ShuaiMXu/mv-skills-bundle (14 stars, last pushed 5mo ago), licensed MIT. It adds 264 tokens to every session and 5,409 once invoked, about $0.0011 per session on Opus 5.5. 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 2026-08-31.
Other skills, from other repositories
sketch-correction-worker
Use when doing single-beat sketch quality control and preparing correction labels from a SuperTale project path and episode number. This skill turns one episode's selected sketches plus local script beat data into selectresult.jsonl, selectsummary.json, tasks.jsonl, and labels.jsonl for edit execution. Prefer using…
freezone
Use when the active chat surface is 虾画/Freezone/canvas, or when the user asks about canvas nodes, visual boards, selections, graph edits, canvas actions, layout, or Freezone short-video workflow work.
dramaclaw-workflows
Create, revise, confirm, run, or resume multi-node DramaClaw Freezone canvas workflows. Use for requests involving a workflow, several connected nodes, grouped production stages, storyboards, or text/image/video/audio pipelines; do not use for one standalone canvas operation.
workflows
Use in Freezone/虾画 chat when the user selects a Skill or asks to create, plan, analyze, or expand a dynamic canvas workflow for short drama, ads, ecommerce, product video, MV, text, image, video, or audio production.
dramaclaw
Use for explicit DramaClaw/虾导 production work: ingesting a screenplay; checking or resuming a bound pipeline; generating or displaying portraits, identity images, sketches, first frames, audio, beat videos or final episodes; and reading or updating project artifacts. Also use for identity questions. Generic…
sketch-storyboard-director
Use when doing a separate post-QC storyboard-directing pass on SuperTale sketches. This skill focuses on shot scale, camera angle, staging, composition, and beat-to-beat visual rhythm after basic sketch correctness has already passed QC.