agentleFS
thevarunthevarun/short-film-builder/.claude/skills/story-film/SKILL.md

story-film

Skill0 starsChanged 2 months agoMITView on GitHub

Animate a written story into a narrated film — cast, beats, plates, one continuous voice take, force-aligned timeline, Veo shots cut to sentences, reviewed strips, verified renders. Run at the start of a new story and resume at any stage.

1 other project uses a version of this.

Adopted by

Most recently taken up by

  • Other versions1 project uses a version of this.

    Word for word · 0

    None found.

    Edited · 0

    None found.

    Wrote their own version · 1

    1. badgids/Story-Film-Skills★ 4same name

    How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.

  • In this repository3 other files it gives its agents.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 2 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    gate the user passes explicitly: before the next begins. Every expensive mistake on the first film came from skipping one.

    line 12
  2. Distinctive:

    Use lite tier throughout unless the user asks otherwise.

    line 20
  3. Very specific:

    Read shows/<slug>/ first: which artifacts exist tells you the stage.

    line 40

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “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

Story → Film

Turn $ARGUMENTS (a story file, optionally a show slug) into a finished film under shows/<slug>/, following docs/animated-story-pipeline.md. Twelve stages; each ends at a gate the user passes explicitly before the next begins. Every expensive mistake on the first film came from skipping one.

Stance

  • The author's words are the script. Adapt narration; never cut a verbatim line without asking.
  • Review before spend. Plates are painted only after the shot plan is confirmed; clips are generated only after the plate audit; nothing is rendered "final" until the checks pass on the file.
  • Show evidence, not assertions: strips for clips, verify-sync and verify-picture output for renders, volumedetect numbers for the mix.
  • Use lite tier throughout unless the user asks otherwise; it was enough for a whole film.

Stages

  1. Doctor → npx tsx scripts/doctor.ts prints no ✗. Keys are the user's to paste; tell them the page, never the value. Gate: green.
  2. Read & cast → character-bible.md. Canonical heights; proportion rules by subject class; a locked style anchor. Gate: intent confirmed.
  3. Character sheets → art/characters/. Generate adults fresh (edit mode inherits child anatomy). Gate: each approved.
  4. Beat sheet → beat-sheet.md with verbatim lines, runtime budget, the governing world rule. Gate: runtime and adaptation agreed.
  5. Location bible → art/locations/, one master plate per place. Gate: approved.
  6. Audio script → audio-script-*.md + audio/script-*.json, pre-normalised, with audio tags. Audition new voices with voice-audition.ts. Gate: cast chosen, script read once.
  7. Record & align → record-dialogue → force-align → split-sentences → align-sentences. Gate: the user listens to the take and accepts the runtime.
  8. Shot plan → shot-plan-*.md: shots cut to sentence spans, live vs 2.5D, joins, plates needed. Gate: confirmed.
  9. Plates → art/shots/, in story order, each referencing the last. Then the plate audit: props from a later beat, characters not yet introduced, legibility, nothing that breaks the world rule. Gate: audit clean.
  10. Shot specs → motion/motion-*.json per the video-shot-prompts skill, and the cut in film.json (schema schemas/film.ts; link-show once). Gate: check-retime <slug> passes.
  11. Generate → animate-shots (detach with nohup for more than a few clips). Review every clip as a strip (build-strips) against the audit rules; re-roll what fails, fixing the plate when the plate is the cause. Gate: strips reviewed.
  12. Compose & render → <Id>Arc<X>, then <Id><Part>, then <Id>Film, all from film.json. Gate: verify-sync and verify-picture pass on each render.
  13. Finish → cards and music in film.json (title, Fin, credits), compose-music long enough to land under the credits (check-retime says), mix levels measured. Gate: the user watches the whole film.

When resuming

Read shows/<slug>/ first: which artifacts exist tells you the stage. Re-run the gate for the last completed stage rather than trusting it.

References

  • Process and learnings: docs/animated-story-pipeline.md
  • Prompt craft: the video-shot-prompts skill
  • Worked example, with costs and post-mortems: shows/alice-in-dragonland/retrospective.md

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.