Agent skill

Capcut Theme Video

by OPPO-Mente-Lab in OPPO-Mente-Lab/X-OmniClaw

CapCut (剪映) one-tap video creation on Android. An agent skill from OPPO-Mente-Lab/X-OmniClaw.

Apache-2.0Auto-check passedProductivity & Automation

Install Capcut Theme Video

skills CLI
$ npx skills add OPPO-Mente-Lab/X-OmniClaw --skill capcut-theme-video -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install OPPO-Mente-Lab/X-OmniClaw capcut-theme-video --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/OPPO-Mente-Lab/X-OmniClaw.git skills-src && mkdir -p .claude/skills && cp -r skills-src/app/src/main/assets/skills/capcut-theme-video .claude/skills/capcut-theme-video && rm -rf skills-src

Use ~/.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/

Facts

Skill name
capcut-theme-video
GitHub stars
265
Token cost
~1.3k tokens
SKILL.md length
391 words
Files
1
Skills in repo
14
Repo updated
First seen
Licence
Apache-2.0

At a glance

CapCut (剪映) one-tap video creation on Android. An agent skill from OPPO-Mente-Lab/X-OmniClaw.

  • Works in 4 steps: Deep Link 直达一键成片相关页(优先) → 选素材页:相册与分类切换 → 勾选素材(点「圈」不点「大图」)(批量点击) → …
  • Productivity & Automation work in your project
  • SKILL.md covers 何时使用, 相册名 A_latest(硬规则), 核心策略:如何获取照片与目标张数 K and 剪映操作执行步骤, plus 1 more section
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

Capcut Theme Video is an agent skill from OPPO-Mente-Lab/X-OmniClaw. CapCut (剪映) one-tap video creation on Android. Always clear Alatest first (copyimagestoalbum clear-only), then retrieve by time (listgalleryimages) or content (image-memories). Copy to Alatest (clearalbumfirst), CapCut 照片视频 → Alatest → 全部, multitap, N=K.

Its SKILL.md is about 1.3k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Productivity & Automation. It works with Android. The repository describes itself as: An edge-native Multimodal Android Agent that integrates multimodal perception, memory and action. The licence is Apache-2.0.

When your agent uses it

  • Productivity & Automation work in your project

Example prompts

  • “/capcut-theme-video”

Workflow steps

4 steps, taken from the step headings in SKILL.md.

  1. Deep Link 直达一键成片相关页(优先)
  2. 选素材页:相册与分类切换
  3. 勾选素材(点「圈」不点「大图」)(批量点击)
  4. 等待与下一步

What it can do on your machine

Read from SKILL.md and the folder at commit 8614c93. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    No scripts in the folder and no shell commands in SKILL.md.

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    No URLs in SKILL.md.

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Capcut Theme Video loads about 1.3k tokens when it runs. Until then it costs about 71 tokens; SKILL.md has 391 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~71
When it runs · the whole SKILL.md, loaded when a task matches
~1.3k

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.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated 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); files beside SKILL.md are not scanned.

SKILL.md

The full file from OPPO-Mente-Lab/X-OmniClaw at commit 8614c93, republished under its Apache-2.0 licence (© OPPO-Mente-Lab). 391 words, ~1,336 tokens.

Download SKILL.mdSave it as .claude/skills/capcut-theme-video/SKILL.md (or your agent's skills folder).
name
capcut-theme-video
description
CapCut (剪映) one-tap video creation on Android. Always clear A_latest first (copy_images_to_album clear-only), then retrieve by time (list_gallery_images) or content (image-memories). Copy to A_latest (clear_album_first), CapCut 照片视频 → A_latest → 全部, multi_tap, N=K.

剪映(CapCut)一键成片与相册选图

本 Skill 指导你如何获取目标照片、统一存入 A_latest 相册,并在剪映中完成「一键成片」。

何时使用

用户要:打开剪映、一键成片、从相册选多张图(按时间或按主题)、去成片 / 导出视频 等多步 UI 操作时加载本 Skill。

包名:com.lemon.lv

相册名 A_latest(硬规则)

  • 本流程使用的相册名有且只有一个:A_latest(与 Pictures/A_latest/ 及剪映「照片视频」下拉显示一致)。
  • 禁止变体:A_latest_风景、主题_A_latest、latest、带空格等;主题/时间只体现在检索,不要写进相册名。
  • 入库前必须先清相册:最终写入时 copy_images_to_album 须 clear_album_first: true。
  • 任意检索前也须先清空 A_latest(路径 A、路径 B 相同):见下「步骤 0」。避免旧副本混入按时间的 list_gallery_images 结果,也避免按内容选图时相册里仍残留上一轮素材、与本次拷贝混淆。

核心策略:如何获取照片与目标张数 K

在进入剪映前,必须先确定要选哪些照片以及目标张数 K。不论是按时间还是按主题,所有选出的照片都必须先复制到 A_latest 相册中,然后在剪映里统一从 A_latest 相册中选取。

步骤 0:先清空 A_latest(路径 A 与路径 B 必做,先于一切检索)

在任何 list_gallery_images(路径 A)或 memory_search / memory_get(路径 B)之前,先调用 copy_images_to_album:

  • album_name: "A_latest"
  • clear_album_first: true
  • 不要传 content_uris / media_ids → 工具只清空该相册(空相册则删除 0 条)。

原因简述:

  • 路径 A:A_latest 里旧副本仍在 MediaStore 时,date_preset 等未按相册排除的检索会把它们和真实目标照片混在一起。
  • 路径 B:先清空可保证后续写入 A_latest 时相册状态干净,不与上一轮成片残留混淆。

可选探针(与直接先清空等价):先 list_gallery_images,album_filter: "A_latest",limit: 1;仅当 count ≥ 1 时再执行上述仅清空调用。

步骤 1:获取目标照片的 content_uris

完成步骤 0 后,按用户需求二选一:

路径 A:基于时间或相册名(非主题语义)

适用场景:用户明确要求「今天拍的照片」「昨天的照片」「最近 24 小时」,或者指定了具体的系统相册名。 操作步骤:

  1. 调用 list_gallery_images,使用 date_preset (today | yesterday | last_24h) 或 album_filter(必要时提高 limit,最大 100)。
  2. 从返回的图片列表中提取所有的 content_uri。
路径 B:基于特定主题、事件或图像内容

适用场景:用户要求按内容找图,如「风景」「海边」「猫」「去北京旅行」。 操作步骤:

  1. 提取主题词:从用户话里抽出检索用词。
  2. 检索 image-memories(按内容找图时不要用 list_gallery_images 的 name_filter 当主题检索:文件名多为 IMG…,不含「风景」「猫」等语义,易得到 0 条):
    • 调用 memory_search,query 填主题词;可适当增大 maxResults(建议 ≥ 10)、略降 minScore(建议 ≤ 0.4)以多召回。
    • 遍历返回的 citations 列表,逐条 memory_get,提取 content_uri 或 media_id;不要只读一段就停,避免漏图。
步骤 2:写入固定相册 A_latest(硬规则)

无论走的是路径 A 还是路径 B,拿到 content_uris 或 media_ids 后:

  1. 调用 copy_images_to_album:album_name: "A_latest",传入收集到的 content_uris 或 media_ids,clear_album_first: true(再次清空后拷贝;与步骤 0 衔接,避免遗漏或并发下残留)。
  2. 拷贝后调用 list_gallery_images,album_filter: "A_latest",记下返回的 count,这就是目标张数 K。
  3. 在剪映中,必须先点「照片视频」下拉选择 A_latest 相册,再进行选图。

剪映操作执行步骤

无需在剪映首页再点「一键成片」时,用 device action=open 传 uri + package_name:

text
device(action="open", package_name="com.lemon.lv", uri="videocut://template/smart_recommend?enter_from=intelligent_edit&tab_name=edit")

随后 device(action="snapshot") 确认是否已进入选素材界面。若出现「是否允许打开剪映」等系统权限弹窗(如 com.oplus.securitypermission),先点允许再 snapshot。若 Deep Link 失败,请在首页用 ref 点击 text '一键成片'。

2. 选素材页:相册与分类切换

snapshot vs screenshot 硬规则(违反必失败) 全程只能用 device(action="snapshot") 返回带 refs 的无障碍树并用 tap ref 点击。禁止用 screenshot 盲猜坐标。

  1. 先点相册下拉(硬规则):在树里找到 link '照片视频' 这一行,tap 这一行的 ref(带上 snapshot_id)。禁止用同区域附近的无文案 button 代替。
  2. 选择 A_latest 相册:展开列表后再 snapshot,在列表里找 text 'A_latest'(名称精确为 A_latest,无后缀)。
    • 正确:tap 紧挨在 text 'A_latest' 的「上一节点」那个 button 的 ref。
    • 错误:误点后面的 button(那是下一行相册)。
  3. 切换到「全部」Tab(重要):
    • 选中 A_latest 后,剪映默认可能停留在「照片」或「视频」Tab。
    • 必须在顶部 Tab 栏找到 text '全部'(或 button '全部'),点击其对应的 ref。
    • 目的:确保能看到 A_latest 相册下的所有素材(无论图片还是视频)。
    • snapshot 确认 全部 已被选中(selected)。
Show full SKILL.md (138 more words)Show less
3. 勾选素材(点「圈」不点「大图」)(批量点击)
  1. 勾选注意事项 在相册网格中,每个格子常出现一对相邻节点(先大块 button 再小 link 圈):
    • 只点勾选小圈:多为 link(未选中时无文案,选中后常带序号如 link '1' (selected))。
    • 禁止点大块无文案 button:那是整张缩略图区域,点击会进入大图预览。
  2. 使用 multi_tap 勾选全部:
    • 在当前 Tab 网格中,只列各格勾选小圈的 ref(多为 link),勿混入同格大块无文案缩略图 button。
    • multi_tap 单次最多 15 个 ref(与单屏可见格数一致);超出则:multi_tap 一批 → snapshot → 滚动 → 再 multi_tap 下一批。重复 ref 会去重(保序)。
    • 每次 act(含 multi_tap)须带当前最近一次 snapshot 返回的 snapshot_id(严格模式下 ref 会过期)。
  3. 校验:
    • 核对张数:每次点击或 multi_tap 后必须 snapshot,核对底部 下一步 (N) 的 N 是否达到目标张数 K。
    • 易错:界面常出现「选择3个以上素材,成片效果更佳」,这只是推荐,绝不表示选 3 张就可以点下一步。必须选到 N = K。
4. 等待与下一步
  • 下一步 (N) 的 N = K 后,点击「下一步」。
  • 后续生成/导出环节,必须使用 act + wait 进行等待:
    text
    device(action="act", kind="wait", timeMs=10000)
    错误:device(action="wait")。
  • 导出很慢时应用多次短 wait + snapshot,不要假设一次等 10 分钟已生效。

简要检查清单

  1. 步骤 0(两路径相同):copy_images_to_album(A_latest + clear_album_first: true,无 URIs)仅清空。
  2. 检索:路径 A → list_gallery_images;路径 B → memory_search / memory_get → 得到 content_uri 列表。
  3. 统一入库:copy_images_to_album(A_latest + clear_album_first: true + 本次 URIs)→ list_gallery_images album_filter: "A_latest" 得 K。
  4. 进入剪映:优先 Deep Link,失败则首页 ref 点击。
  5. 选相册集:下拉选择 A_latest。
  6. 切全部 Tab:必须点击「全部」Tab,确保素材可见。
  7. 选图:只点勾选小圈(link);multi_tap(≤15/次)配合 snapshot。
  8. 核对与等待:下一步 (N) 直到 N = K;act + wait 再 snapshot。

© OPPO-Mente-Lab, 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

Files

Just SKILL.md in app/src/main/assets/skills/capcut-theme-video of OPPO-Mente-Lab/X-OmniClaw.

Open the folder on GitHubat commit 8614c93

Compare with similar skills

Capcut Theme Video 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.

Capcut Theme Video compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Capcut Theme Video this skillOPPO-Mente-Lab/X-OmniClaw265—~1.3kAutomated safety check: PassApache-2.0
Cherrypick Logkomikku-app/komikku4.8k—~1.9kAutomated safety check: PassApache-2.0
OpenGUI Phone Remote ControlCore-Mate/OpenGUI1.8k—~2.6kAutomated safety check: NotesCustom licence
Phone And Simulator GUI ControlCore-Mate/OpenGUI1.8k—~7.7kAutomated safety check: PassCustom licence
Kane CLI Browser TestingLambdaTest/kane-cli247—~8.4kAutomated safety check: PassApache-2.0
Burner Phone Android Controlsmallnest/goclaw598—~436Automated safety check: PassMIT

Similar skills

  • Cherrypick Log

    komikku-app/komikku

    Update and review cherrypicklog.md, the todo list of upstream commits not yet cherry-picked into the fork's target branch.

    4.8k GitHub stars~1.9k tokensUpdated 2 days ago
    Productivity & AutomationAuto-check passed
  • Lets a coding agent list, run, pause, resume or cancel natural-language tasks on an Android phone through the OpenGUI backend instead of raw ADB commands.

    1.8k GitHub stars~2.6k tokensUpdated yesterday
    Productivity & AutomationAuto-check: notes
  • Completes an already-authorized task on a real Android phone or a macOS iOS simulator one screenshot-driven action at a time.

    1.8k GitHub stars~7.7k tokensUpdated yesterday
    Productivity & AutomationAuto-check passed
  • Kane CLI Browser Testing

    LambdaTest/kane-cli

    Drives a real browser through the kane-cli tool and designs requirement-linked test suites from a PRD or a plain description, with mobile and cloud-grid runs.

    247 GitHub stars~8.4k tokensUpdated 3 days ago
    Testing & QAAuto-check passed
  • Controls an Android device over ADB in a loop of screenshot, vision analysis, tap and verification, so the agent acts on real on-screen coordinates.

    598 GitHub stars~436 tokensUpdated 6 mo ago
    MobileAuto-check passed
  • Autofish Android Control

    felinics/Autofish

    Controls an Android device through the af CLI with observe, act and verify steps, ref-based taps, screenshots and recovery when a step goes wrong.

    118 GitHub stars~1.7k tokensUpdated 1 mo ago
    MobileAuto-check passed

More from OPPO-Mente-Lab/X-OmniClaw

All 14 skills in this repo
  • Feishu Bitable

    OPPO-Mente-Lab/X-OmniClaw

    Feishu multi-dimensional table (Bitable) operations. An agent skill from OPPO-Mente-Lab/X-OmniClaw.

    265 GitHub stars~631 tokensUpdated 4 mo ago
    Auto-check passed
  • Feishu Chat

    OPPO-Mente-Lab/X-OmniClaw

    Feishu chat operations. An agent skill from OPPO-Mente-Lab/X-OmniClaw.

    265 GitHub stars~592 tokensUpdated 4 mo ago
    Auto-check passed
  • Feishu Task

    OPPO-Mente-Lab/X-OmniClaw

    Feishu task management operations. An agent skill from OPPO-Mente-Lab/X-OmniClaw.

    265 GitHub stars~482 tokensUpdated 4 mo ago
    Auto-check passed
  • Feishu Urgent

    OPPO-Mente-Lab/X-OmniClaw

    Feishu urgent notification tool. An agent skill from OPPO-Mente-Lab/X-OmniClaw.

    265 GitHub stars~601 tokensUpdated 4 mo ago
    Auto-check passed
  • App Search

    OPPO-Mente-Lab/X-OmniClaw

    主流APP搜索。当用户想要在拼多多、美团、高德地图、抖音、快手、小红书、哔哩哔哩、知乎、百度、OPPO音乐、OPPO软件商店等APP中搜索时激活。覆盖电商购物、生活服务、短视频内容、知识搜索、音乐搜索、应用商店搜索等场景。

    265 GitHub stars~4.1k tokensUpdated 4 mo ago
    Auto-check passed
  • Clipboard To Shortcut

    OPPO-Mente-Lab/X-OmniClaw

    将剪切板中的 App 跳转链接制作成一个可复用的快捷指令 Skill,下次用户说出 Skill 名称即可一键直达对应 App 页面。

    265 GitHub stars~2.8k tokensUpdated 4 mo ago
    Auto-check passed

Works with

Questions about Capcut Theme Video

What does Capcut Theme Video do?

CapCut (剪映) one-tap video creation on Android. An agent skill from OPPO-Mente-Lab/X-OmniClaw. Capcut Theme Video is an agent skill from OPPO-Mente-Lab/X-OmniClaw. CapCut (剪映) one-tap video creation on Android.

When should I use Capcut Theme Video?

Capcut Theme Video fits situations like: productivity & Automation work in your project.

How do I install Capcut Theme Video in Claude Code?

Run `npx skills add OPPO-Mente-Lab/X-OmniClaw --skill capcut-theme-video -a claude-code`. Or copy the skill folder (app/src/main/assets/skills/capcut-theme-video in OPPO-Mente-Lab/X-OmniClaw) into .claude/skills/capcut-theme-video in your project. Claude Code loads it when a task matches its description.

How do I install Capcut Theme Video in Codex?

Run `npx skills add OPPO-Mente-Lab/X-OmniClaw --skill capcut-theme-video -a codex`. Or copy the skill folder (app/src/main/assets/skills/capcut-theme-video in OPPO-Mente-Lab/X-OmniClaw) into .agents/skills/capcut-theme-video in your project. Codex loads it when a task matches its description.

Can I use Capcut Theme Video in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add OPPO-Mente-Lab/X-OmniClaw --skill capcut-theme-video -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/capcut-theme-video, .gemini/skills/capcut-theme-video, .github/skills/capcut-theme-video and .opencode/skills/capcut-theme-video in your project.

What does Capcut Theme Video need to run?

SKILL.md names no scripts, command-line tools or credentials: Capcut Theme Video is instructions for the agent only.

Does Capcut Theme Video access the network?

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.

Is Capcut Theme Video safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Capcut Theme Video use?

Capcut Theme Video 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.

How many tokens does Capcut Theme Video use?

About 1.3k tokens (SKILL.md is roughly 5.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Capcut Theme Video?

Skills that share tags, products or a category with Capcut Theme Video: Cherrypick Log (komikku-app/komikku, 4.8k stars), OpenGUI Phone Remote Control (Core-Mate/OpenGUI, 1.8k stars), Phone And Simulator GUI Control (Core-Mate/OpenGUI, 1.8k stars) and Kane CLI Browser Testing (LambdaTest/kane-cli, 247 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Capcut Theme Video?

OPPO-Mente-Lab (a GitHub organization) maintains it in OPPO-Mente-Lab/X-OmniClaw, which has 265 GitHub stars. The repository holds 14 skills in this directory. The repository was last updated on May 22, 2026.

Source: OPPO-Mente-Lab/X-OmniClaw on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.