HyperFrames Video Entry Point
heygen-com/hyperframes
Entry point for making, editing and rendering videos from HTML compositions with HyperFrames, routing each request to the right workflow.
拉片:把一条成片拆成逐镜头的分析表——每个镜头的时长、景别、类别、运镜、画面. An agent skill from eternityspring/reelbench-skills.
$ npx skills add eternityspring/reelbench-skills --skill video-shots -a claude-codeProject install by default; add -g for ~/.claude/skills/.
$ gh skill install eternityspring/reelbench-skills video-shots --agent claude-codeProject scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).
$ git clone --depth 1 https://github.com/eternityspring/reelbench-skills.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/video-shots .claude/skills/video-shots && rm -rf skills-srcUse ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.
Claude Code skills documentation · loads skills from .claude/skills/
Install the "video-shots" agent skill from https://github.com/eternityspring/reelbench-skills/tree/main/skills/video-shots into .claude/skills/video-shots/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "video-shots", then confirm the skill loads.Claude Code copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$skill-installer install https://github.com/eternityspring/reelbench-skills/tree/main/skills/video-shotsType this inside Codex. $skill-installer <name> installs a curated skill from openai/skills. The installer writes to $CODEX_HOME/skills (default ~/.codex/skills). Restart Codex if the skill does not show up.
$ npx skills add eternityspring/reelbench-skills --skill video-shots -a codexProject install goes to .agents/skills/; add -g for ~/.codex/skills/.
$ gh skill install eternityspring/reelbench-skills video-shots --agent codexProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/eternityspring/reelbench-skills.git skills-src && mkdir -p .agents/skills && cp -r skills-src/skills/video-shots .agents/skills/video-shots && rm -rf skills-srcUse ~/.agents/skills/ instead of .agents/skills for a personal install.
Codex skills documentation · loads skills from .agents/skills/
Install the "video-shots" agent skill from https://github.com/eternityspring/reelbench-skills/tree/main/skills/video-shots into .agents/skills/video-shots/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "video-shots", then confirm the skill loads.Codex copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add eternityspring/reelbench-skills --skill video-shots -a cursorProject install goes to .agents/skills/; add -g for ~/.cursor/skills/.
$ gh skill install eternityspring/reelbench-skills video-shots --agent cursorProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/eternityspring/reelbench-skills.git skills-src && mkdir -p .cursor/skills && cp -r skills-src/skills/video-shots .cursor/skills/video-shots && rm -rf skills-srcUse ~/.cursor/skills/ instead of .cursor/skills for a personal install.
Cursor skills documentation · loads skills from .cursor/skills/, .agents/skills/, .claude/skills/, .codex/skills/
Install the "video-shots" agent skill from https://github.com/eternityspring/reelbench-skills/tree/main/skills/video-shots into .cursor/skills/video-shots/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "video-shots", then confirm the skill loads.Cursor copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gemini skills install https://github.com/eternityspring/reelbench-skills.git --path skills/video-shots--scope user (default) or --scope workspace; --path is the subfolder of the repo that holds the skill; --consent skips the security confirmation prompt.
$ npx skills add eternityspring/reelbench-skills --skill video-shots -a gemini-cliProject install goes to .agents/skills/; add -g for ~/.gemini/skills/.
$ gh skill install eternityspring/reelbench-skills video-shots --agent gemini-cliProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/eternityspring/reelbench-skills.git skills-src && mkdir -p .gemini/skills && cp -r skills-src/skills/video-shots .gemini/skills/video-shots && rm -rf skills-srcUse ~/.gemini/skills/ instead of .gemini/skills for a personal install, then run /skills reload.
Gemini CLI skills documentation · loads skills from .gemini/skills/, .agents/skills/
Install the "video-shots" agent skill from https://github.com/eternityspring/reelbench-skills/tree/main/skills/video-shots into .gemini/skills/video-shots/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "video-shots", then confirm the skill loads.Gemini CLI copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ gh skill install eternityspring/reelbench-skills video-shotsInstalls for Copilot at project scope by default; add --scope user for a personal install. Preview a skill first with gh skill preview. Needs GitHub CLI 2.90.0 or later (public preview).
$ npx skills add eternityspring/reelbench-skills --skill video-shots -a github-copilotProject install goes to .agents/skills/; add -g for ~/.copilot/skills/.
$ git clone --depth 1 https://github.com/eternityspring/reelbench-skills.git skills-src && mkdir -p .github/skills && cp -r skills-src/skills/video-shots .github/skills/video-shots && rm -rf skills-srcUse ~/.copilot/skills/ instead of .github/skills for a personal install. Commit .github/skills so cloud agent and code review can use it.
GitHub Copilot skills documentation · loads skills from .github/skills/, .claude/skills/, .agents/skills/
Install the "video-shots" agent skill from https://github.com/eternityspring/reelbench-skills/tree/main/skills/video-shots into .github/skills/video-shots/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "video-shots", then confirm the skill loads.GitHub Copilot copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
$ npx skills add eternityspring/reelbench-skills --skill video-shots -a opencodeOpenCode documents no install command of its own. Project install goes to .agents/skills/; add -g for ~/.config/opencode/skills/.
$ gh skill install eternityspring/reelbench-skills video-shots --agent opencodeProject scope by default (.agents/skills/); add --scope user for a personal install.
$ git clone --depth 1 https://github.com/eternityspring/reelbench-skills.git skills-src && mkdir -p .opencode/skills && cp -r skills-src/skills/video-shots .opencode/skills/video-shots && rm -rf skills-srcUse ~/.config/opencode/skills/ instead of .opencode/skills for a personal install.
OpenCode skills documentation · loads skills from .opencode/skills/, .claude/skills/, .agents/skills/
Install the "video-shots" agent skill from https://github.com/eternityspring/reelbench-skills/tree/main/skills/video-shots into .opencode/skills/video-shots/ in this project. Copy the whole folder (SKILL.md and every file beside it), keep the folder name "video-shots", then confirm the skill loads.OpenCode copies the folder itself, the same result as the manual copy. Check what it changed before you commit it.
video-shots拉片:把一条成片拆成逐镜头的分析表——每个镜头的时长、景别、类别、运镜、画面. An agent skill from eternityspring/reelbench-skills.
Video Shots is an agent skill from eternityspring/reelbench-skills. 拉片:把一条成片拆成逐镜头的分析表——每个镜头的时长、景别、类别、运镜、画面。 分工刻在骨子里:能量的都由代码量(切点来自 ffmpeg 场景检测,时长是切点相减, 运动量是逐帧差分的中位数),模型只判断它真正该判断的那几件事(景别 / 类别 / 运镜 / 画面 / 节奏), 然后每一条判断都被代码当场对账——声称推拉摇移却实测几乎不动,门直接拦。 看片走联系表(每镜起手帧 + 收尾帧各拼一张大图,一屏二十几个镜头,a/b 对照就是运镜), 不是一张张翻。检测漏刀多刀用 recut 补刀并刀,自动重编号重算时长,手改边界过不了门。 产出 shots.json + Markdown 镜头表 + 单页交互式拉片报告:内嵌播放器(播放时同步高亮镜头、 点镜头跳转)、镜头节奏带、可搜索可筛选可排序的镜头表(列表 / 卡片两种视图、首尾关键帧并排、 点图开大图)、景别类别运镜分布、出场人物、质量门、导出 JSON。单文件零依赖,离线双击能开。 15 道质量门全部由脚本确定性检查。 零依赖、零 API key,只要 node 和 ffmpeg。 Use when asked to…
Its SKILL.md is about 1.8k tokens, which your agent loads only when the skill is triggered. The skill folder holds 18 other files, including scripts, reference files and assets (for example `README.en.md`, `README.md` and `examples/demo-shots.json`).
It sits in Media & Creative, covering Video production and AI video generation. It works with FFmpeg. The repository describes itself as: Learning notes and tooling skills for AI video - AI 视频相关的学习与工具 skill. The licence is Apache-2.0.
7 steps, taken from the step headings in SKILL.md.
Read from SKILL.md and the folder at commit 1b51af8. It shows what the files ask for, not the result of running them.
Pre-approves these tools, so the agent can use them without asking each time:
ReadWriteEditBashGlobFrom allowed-tools in the SKILL.md frontmatter.
Ships 4 files in scripts/ (JavaScript), which the agent can run.
Shell commands in SKILL.md call:
nodeFrom the folder's file list and the shell code blocks in SKILL.md.
No URLs in SKILL.md.
From URLs in SKILL.md, links to its own repository left out.
Names no API keys, tokens, secrets or passwords.
From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.
Video Shots loads about 1.8k tokens when it runs, and up to ~8.9k if it reads all its reference files. Until then it costs about 147 tokens; SKILL.md has 340 words of instructions outside code blocks.
Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.
The automated check noted patterns worth knowing about, such as sudo or a known installer.
allowed-tools: Read, Write, Edit, Bash, GlobAutomated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); the scripts in this folder are not scanned.
The full file from eternityspring/reelbench-skills at commit 1b51af8, republished under its Apache-2.0 licence (© eternityspring). 340 words, ~1,764 tokens.
.claude/skills/video-shots/SKILL.md (or your agent's skills folder). This skill also uses 14 other files; get the full folder from GitHub.给成片拉片——把一条片子拆成逐镜头的分析表:时长、景别、类别、运镜、画面。
前提刻在骨子里:镜头边界是量出来的,不是看出来的。 模型看视频最不可靠的就是报时间—— 「这个镜头大概 3 秒」和「2.97 秒」差的不是精度,是这份表能不能用。所以这里划一条死线:
| 谁来定 | 什么 | 怎么定 |
|---|---|---|
| 代码 | 切点、时长、片长、帧率 | ffmpeg 场景检测 + ffprobe,两位小数 |
| 代码 | 每个镜头的实测运动量 | 逐帧差分曲线的区间中位数(两端剔除,避开切点尖峰) |
| 模型 | 景别、类别、运镜、画面、节奏 | 看关键帧判断——只有这几件事是模型的活 |
| 代码 | 判断对不对 | 15 道质量门,逐条对账 |
最硬的一道门是运镜实测对账:摄影机真动了,像素不可能不变。所以「声称推/拉/摇/移/跟, 实测帧间变化接近 0」这一向直接拦——这是模型拉片最常见的幻觉。反过来(声称固定、实测很动) 不拦,只出提示:固定机位前面有人跳舞,帧间差一样会爆。
{baseDir} = 本文件所在目录。脚本 {baseDir}/scripts/video-shots.mjs,零依赖,node 直接跑。
边界(不做的事):不转写语音(没有 ASR,台词靠画面上烧录的字幕读,读不到就留空并说明)、
不做人脸识别与人物自动归并(cast 是人工编号)、不评价片子好坏(报告只给事实和统计)、
不剪辑不导出片段、不做镜头内的物体检测。
只要一个视频文件。先问清两件事,问不到就按默认走并在汇报里说明:
note 里该多记什么不同——表的结构是一样的。cd <输出目录>
node {baseDir}/scripts/video-shots.mjs seed <video> --track track.json --title "<片名>" > shots.jsonstderr 会报:片长、帧率、分辨率、检测到几个切点、合并后几个镜头。先看这一行再往下走:
--threshold 0.15 重跑(暗戏、慢片、
同机位对话多的片子都要往下调)0.4 调,或留到 Step 4 用 recut --merge 并底稿里 start / end / seconds / motion 已经填好,size / category / camera / frame 是空的——
那四个空格子才是模型的活。
node {baseDir}/scripts/video-shots.mjs frames shots.json --video <video>
node {baseDir}/scripts/video-shots.mjs sheet shots.json --cols 4 --rows 6
node {baseDir}/scripts/video-shots.mjs sheet shots.json --cols 4 --rows 6 --pick b每镜两张:frames/S01a.jpg(起手 15% 处)和 S01b.jpg(收尾 85% 处)。
联系表把它们各拼成一张大图(行优先,S01 在左上),a 表看内容,b 表看运镜——
同一格前后对照,取景变没变一眼就知道。
先看联系表,再看单帧。 一张张翻完整部片是浪费额度:一张联系表 = 二十几个镜头,
只在判不准的那几个镜头上回去看单帧(Read frames/S07a.jpg)。
一批 ≤ 25 个镜头(正好一张联系表)。每批拿到:
{baseDir}/references/taxonomy.md(四张词表 + 判据,照着填)和 {baseDir}/references/analysis-pass.md(怎么看、常见病)填的顺序:景别 → 类别 → 运镜 → 画面 → 节奏。运镜看 a/b 取景差 + 实测运动值,
两者打架时信实测。顺带记 subjects / onscreenText / audio——
画面上烧录的对白字幕算台词,进 audio 并带上说话人;片名、字卡、界面文字进 onscreenText。
节奏(rhythm + rhythmNote)回答的是「观众为什么还没划走」:
钩子 / 铺垫 / 递进 / 重音 / 转折 / 兑现 / 换气 / 收口,八个角色选一个,
再用一句话写清观众这一刻看到什么(判据见 taxonomy.md 第五节)。
可选字段:不做节奏分析就一镜都别标;标了就得整片标全,门会拦半张表。
标满之后校验会顺带提示三件事:开篇有没有钩子、兑现前面有没有铺垫、有没有连着六镜节奏发平。
编辑 shots.json 时只动那几个字段:start / end / seconds / motion / seedCuts / meta
是机器字段,改了就是伪造证据,门会点名。
场景检测必然在两个地方出错:叠化和暗场对暗场漏刀,手持晃动和闪光多刀。 a 帧和 b 帧根本是两个场景,就是漏刀的铁证。
node {baseDir}/scripts/video-shots.mjs recut shots.json --track track.json \
--split 63.5 --split 127.37 --merge 45.97 > shots.new.json && mv shots.new.json shots.json自动重编号、重算时长与实测运动,补的刀记进 manualCuts(boundary 门认它)。
边界没动过的镜头标注原样保留;被拆被并的镜头标注清空并在 note 里写明出身——
这两半是不是一回事,得重新看画面,不许把旧描述顺下去。
改完重抽这些镜头的帧(frames 会覆盖整个目录,直接重跑就行),把清空的格子补上。
node {baseDir}/scripts/video-shots.mjs validate shots.json --track track.json --frames frames15 道门全是代码:时间轴连续(按序、首尾相接、从 0 到片尾)、时长自洽(seconds = end − start,
短镜必须带 note)、镜号连号、景别/类别/运镜三张词表、转场枚举、画面描述可核对
(中文 ≥12 字 / 英文 ≥8 词 + 空话词表 + 不许「这个镜头…」「This shot…」开头)、画面描述不重复、主体对账 cast、
类别要有证据(对话必须有台词、字卡必须有画面文字、反应必须写是谁、空镜里不许有人)、
运镜实测对账、边界来自检测(自己加的刀必须在 manualCuts 里声明)、关键帧齐全、
节奏分析可核对(可选字段:整片标或整片不标;标了就得写清为什么,空话照拦)。
有违规逐条修,改完重跑,直到通过。 跳过的门会明说原因(没给 --track、没建 cast、
关键帧目录不存在)——跳过不是通过,汇报时要讲。拉英文片加 --lang en,
门的名字和违规信息都会用英文说。
「提示(不拦)」那一栏不是错误,是需要人判断的地方:固定机位实测偏高, 多半是主体在动,也可能是你把一次缓推看漏了,自己回去看一眼那一镜。
node {baseDir}/scripts/video-shots.mjs render shots.json --md --track track.json > shots.md
node {baseDir}/scripts/video-shots.mjs render shots.json --html --track track.json \
--video <原片相对报告的路径> > shots-report.html--video 给报告里的播放器指原片(默认用 JSON 里的 source;观众也能在页面上现场选本地文件)。
界面语言用 --lang zh|en(默认中文)。render 自动去 frames/ 找关键帧,
先抽帧再 render,缺图明说缺、不摆占位图充数。
报告是单文件交互页(样式与交互来自 {baseDir}/scripts/report.css 和 report.js,
render 时整段内联;这三个文件必须一起拷走):
14 项通过 · 1 条提示),提示里点名的镜号可以点着跳过去汇报一句话说清:多少镜、平均镜长、每分钟切次、景别与运镜的大头、最长和最短的镜头在哪、 报告路径;补了几刀并了几刀、哪几道门跳过了、哪几条提示需要人看,明说。
最终落地:
<输出目录>/
├── shots.json ← 拉片主数据
├── track.json ← 运动曲线(机器证据,别手改)
├── shots.md
├── shots-report.html ← 双击就能开
├── frames/ ← S01a.jpg / S01b.jpg …
└── sheets/ ← sheet-a01.jpg / sheet-b01.jpg …(联系表)audio 大面积留空是正常的,
这时把 dialogue 判成 subject 更诚实——类别证据门会逼你做这个选择params.staticMaxMotion / busyMinMotiontransitionIn 写 dissolve--lang zh|en 贯穿所有命令——门的名字、违规信息、命令行输出、报告界面、
四张词表全跟着切(优先级 --lang > JSON 顶层 lang > 中文)。切的是标签,
画面描述、台词、人物名原样不动——那是内容不是标签。
seed --lang en 会把 lang: "en" 写进底稿,后面的 recut / validate / render
不带 --lang 也跟着走英文(连补刀写的 note 和英文表格里的标点都是半角的)--video 指对路径,或在页面上现场选文件。报告本身不嵌视频数据node {baseDir}/scripts/selftest.mjs449 项断言,不调模型、不花额度、不碰 ffmpeg。15 道门每一道都有击穿用例——证明它真的会拦。 改完脚本先跑这个。
{baseDir}/examples/demo-shots.json + demo-track.json:一条 202.9 秒的 AI 短片完整拉片——
53 个镜头,平均镜长 3.83 秒,每分钟 15.7 切;对话占 58%、固定机位占 55%;
最短 0.33 秒(雪地奔跑的闪切),最长 16.06 秒(结尾光柱下的长镜头)。
seedCuts 之外补了 10 刀(叠化和片尾字卡各占一半,全部记在 manualCuts 里),
15 道门全绿,留着两条提示当范例(运镜实测偏高、片尾连着六镜节奏发平)。它是质量基准,也是自测夹具。
© eternityspring, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file
SKILL.md and 14 other files (scripts, references, assets) in skills/video-shots of eternityspring/reelbench-skills.
Open the folder on GitHubat commit 1b51af8
We found 1 copy of this SKILL.md (exact, near-identical or edited) in other folders, from 1 other GitHub owner. This page covers the copy in eternityspring/reelbench-skills, which our catalogue first saw on October 7, 2026.
Video Shots next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.
| Skill | Stars | Used in | Tokens | Auto-check | Licence | Repo updated |
|---|---|---|---|---|---|---|
| Video Shots this skilleternityspring/reelbench-skills | 872 | 1 repos | ~1.8k | Automated safety check: Notes | Apache-2.0 | |
| HyperFrames Video Entry Pointheygen-com/hyperframes | 59k | 3 repos | ~5.2k | Automated safety check: Pass | Apache-2.0 | |
| Stage EditOrkas-AI/Orkas-VideoStudio | 498 | — | ~2.4k | Automated safety check: Pass | MIT | |
| Re Walkthrough Procharlesdove977/re-walkthrough-pro | 164 | — | ~1.1k | Automated safety check: Notes | MIT | |
| Video Productionffroliva/gflow-cli | 266 | — | ~7k | Automated safety check: Pass | MIT | |
| Image To Videonotque/vexjoy-agent | 439 | — | ~817 | Automated safety check: Notes | MIT |
heygen-com/hyperframes
Entry point for making, editing and rendering videos from HTML compositions with HyperFrames, routing each request to the right workflow.
Orkas-AI/Orkas-VideoStudio
Intelligent editing of real user-supplied footage—understand it with transcript/inspected-frame/scene/silence/quality evidence, then choose deterministic timeline operations or a constrained…
charlesdove977/re-walkthrough-pro
Turn a Zillow listing into a cinematic room-by-room walkthrough video (Apify scrape → Higgsfield image-to-video → ffmpeg stitch) to sell to real estate agents
ffroliva/gflow-cli
A skill your agent uses when the user wants a finished video out of gflow rather than a single clip — a scripted scene, a talking-head or dialogue piece, an explainer, a product montage, a story…
notque/vexjoy-agent
FFmpeg-based video creation from image and audio. An agent skill from notque/vexjoy-agent.
Orkas-AI/Orkas-VideoStudio
Deterministically assemble an approved cross-modal EDL (plan.json) into a finished video — produce each segment (edit/compose/generate/provided, delegated to its line), then assemble in ffmpeg…
eternityspring/reelbench-skills
把拉片数据和原片合成一条能直接看的视频:一边是画面,一边是这一镜的分镜信息 (镜号、起止、时长、景别、类别、运镜、画面描述、台词),镜头切了信息跟着切, 镜头表自动滚动并高亮当前这一镜。
eternityspring/reelbench-skills
把一条视频重建成「只有画面和声音」的干净文件——源片的元数据一概不搬: GPS、设备型号、账号 ID、创建时间、章节、GoPro 的遥测轨,全部留在原地。
Works with
Categories
拉片:把一条成片拆成逐镜头的分析表——每个镜头的时长、景别、类别、运镜、画面. An agent skill from eternityspring/reelbench-skills. Video Shots is an agent skill from eternityspring/reelbench-skills.
Video Shots fits situations like: asked to 拉片、拆镜头、分析视频镜头、镜头时长、景别、运镜、镜头表、 video shot breakdown、shot list from video; tasks that involve Video production; tasks that involve AI video generation.
Run `npx skills add eternityspring/reelbench-skills --skill video-shots -a claude-code`. Or copy the skill folder (skills/video-shots in eternityspring/reelbench-skills) into .claude/skills/video-shots in your project. Claude Code loads it when a task matches its description.
Run `npx skills add eternityspring/reelbench-skills --skill video-shots -a codex`. Or copy the skill folder (skills/video-shots in eternityspring/reelbench-skills) into .agents/skills/video-shots in your project. Codex loads it when a task matches its description.
Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add eternityspring/reelbench-skills --skill video-shots -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/video-shots, .gemini/skills/video-shots, .github/skills/video-shots and .opencode/skills/video-shots in your project.
Going by SKILL.md and its folder, Video Shots needs JavaScript for the scripts in its folder and the command-line tools its instructions call (node). Our summary lists: Node.js. Its frontmatter pre-approves these tools: Read, Write, Edit, Bash, Glob.
SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.
Our automated static check of SKILL.md found notes only (pre-approves every shell command (allowed-tools: bash)), nothing it rates as a warning. It is not a guarantee. The check reads SKILL.md only: the scripts in the folder are not scanned, so read them before running anything.
Video Shots is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.
About 1.8k tokens (SKILL.md is roughly 7.1k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full. Its references folder adds about 7.1k tokens, read only when the agent opens those files.
Skills that share tags, products or a category with Video Shots: HyperFrames Video Entry Point (heygen-com/hyperframes, 59k stars), Stage Edit (Orkas-AI/Orkas-VideoStudio, 498 stars), Re Walkthrough Pro (charlesdove977/re-walkthrough-pro, 164 stars) and Video Production (ffroliva/gflow-cli, 266 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.
eternityspring (a GitHub user) maintains it in eternityspring/reelbench-skills, which has 872 GitHub stars. The repository holds 3 skills in this directory. The repository was last updated on September 21, 2026.
Source: eternityspring/reelbench-skills on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.