agentleFS
harshil-1411harshil-1411/claude_plugin_video_studio/skills/localize/SKILL.md

localize

Skill0 starsChanged 47 days agoApache-2.0View on GitHub

Make a language version of a planned video-studio project (e.g. Hindi, Japanese, Arabic, German) - translate the voiceover, on-screen text, text in the motion graphics, cover headline and post copy, re-time scenes for the language, validate, then render. Use when the user runs /video-studio:localize or asks for the video "in Japanese", "a Hindi version" or several language versions.

22 other projects use a version of this.

Adopted by

Most recently taken up by

Used with

Projects with it also use soak-test, team-live-ops and 10 more.

Other files people use alongside this one.

Files found in the same public projects as a version of this one, more often than they turn up anywhere else. Each project counts once, large collections count for less, and this file's own repository is left out.
  • lilangMaxSkill

    soak-test

    Generate a soak test protocol for extended play sessions. Defines what to observe, measure, and log during long play sessions to surface slow leaks, fatigue effects, and edge cases that only appear after sustained play. Primarily used in Polish and Release phases.

    Also in 9 projects that use this

  • lilangMaxSkill

    team-live-ops

    Orchestrate the live-ops team for post-launch content planning: coordinates live-ops-designer, economy-designer, analytics-engineer, community-manager, writer, and narrative-director to design and plan a season, event, or live content update.

    Also in 9 projects that use this

  • +10 more from lilangMax

The file

We split this skill into the rules it gives an agent

7 rules. 0 are common, 2 are distinctive, and 5 are very specific to this project.

  • 0

    Common

    Most of the 21 edited versions we read keep these.

  • 2

    Distinctive

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Step 1: call localize {project_dir, language}.

    line 19
  2. Very specific:

    Translate: read the sheet and fill each entry's target:

    line 22
  3. Distinctive:

    Never translate code, commands, URLs or UI labels in screenshots.

    line 31
  4. Very specific:

    Leave target empty only for text that stays the same (names).

    line 42
  5. Very specific:

    Step 2: call localize {project_dir, language, apply: true}.

    line 47
  6. Distinctive:

    If the call is refused because the source spec changed, run step 1 again, translate the new entries, then apply.

    line 55
  7. Very specific:

    Call storyboard_render on the localized folder and show it.

    line 57

“Common” is measured: kept by at least 6 in 10 of the 21 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Localize a video

The source project is never changed. Each language gets its own project folder, <project>/localized/<language>/ by default. One language at a time; repeat for each language the user asked for.

  1. Resolve the source project folder (absolute; it needs project/video-spec.json) and the BCP-47 tag (hi-IN, ja-JP, ar-SA, de-DE, ...). Ask which region when the user gives only a language and it matters (for example pt-BR or pt-PT).
  2. Step 1: call localize {project_dir, language}. It copies the project and writes project/translation.json in the new folder. Every entry has a path, a kind and the source text. Many also have a note.
  3. Translate: read the sheet and fill each entry's target:
    • Translate the meaning, not word by word, in the register of the brief (project/creative-brief.yaml). Keep the same claims: no new facts, no dropped caveats. Numbers, units, product and people's names stay as they are unless the language has an established form. Never make up statistics.
    • Stay within each note's budget (words, or characters for Japanese, Chinese, Korean and Thai). Narration that runs over the budget makes its scene longer when the sheet is applied.
    • Never translate code, commands, URLs or UI labels in screenshots. The sheet leaves most of these out already.
    • Mark the emphasised word with *asterisks* when a note asks for it, for example "nicht *Bedeutung*". The emphasis has to be a word of the translated line.
    • For RTL languages (Arabic, Hebrew, Persian, Urdu), write the text in its natural order. The renderers handle direction. Keep Latin names and numbers as they are.
    • For CJK, do not add spaces between words, and keep lines short.
    • Hashtags: # followed by letters, digits or _, with no spaces. Use the tag people search for in that language, or keep the original.
    • Leave target empty only for text that stays the same (names). An empty target keeps the source text. Save the sheet with Edit. Do not edit the localized project/video-spec.json by hand: step 2 rebuilds it from the source spec and the sheet.
  4. Step 2: call localize {project_dir, language, apply: true}. Report:
    • re-timed scenes and any change to target_duration_sec. If the result runs over the targets' duration limits, shorten the longest translations in the sheet and apply again.
    • emphasis notes: a dropped emphasis means no word was marked.
    • errors: fix them in the sheet (or, for structural errors, in the source spec, then run step 1 again, which keeps the unchanged translations) and apply again. If the call is refused because the source spec changed, run step 1 again, translate the new entries, then apply.
  5. Call storyboard_render on the localized folder and show it. Then offer the render skill for that folder. Captions, fonts and the voice follow spec.language. A voice_id in the source spec is removed, because it speaks the source language. Suggest lint afterwards, because reading speed differs by language.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.