moonlight-lupin/agent-skills / creative/clips-studio
clips-studio
VerifiedUse when the user wants to "make a video/clip", "generate a video from text", "animate this image/photo", "create a marketing reel / social video / teaser", "do a 3D / parallax move", "pan/zoom/orbit a shot", or mentions fal.ai, Kling, Veo or Seedance for video. Staged fal.ai workflow — brainstorm, draft cheaply, produce the final at quality. Three modes: text-to-video, animate (still to motion), camera-move (push-in/pan/orbit). Not for still images (image-studio), slide decks, or charts. Honesty discipline: never depict real identifiable subjects via text-to-video.
| File | Kind | Size |
|---|---|---|
| README.md | markdown | 1.9 KB |
| SKILL.md | skill md | 23.5 KB |
| examples/example-run.md | markdown | 8.1 KB |
| references/fal-video-models.md | markdown | 11.9 KB |
| references/motion-honesty-guide.md | markdown | 11.4 KB |
| references/video-prompting-techniques.md | markdown | 3.4 KB |
| scripts/falvid.py | script | 34.5 KB |
| tests/__init__.py | script | 0 B |
| tests/test_falvid.py | script | 31.3 KB |
references/video-prompting-techniques.md
View on GitHub# Video prompting techniques — transferable across models
Distilled from Black Forest Labs' FLUX 3 prompting guide. These techniques
work across fal.ai video models (Kling, Veo, Seedance, Wan) — they are model-
agnostic craft, not FLUX-3-specific.
## Start from the user's wording
The prompt is read by a reasoning harness, not a tag encoder. Keyword tricks
and word order do nothing. Write plain prose at whatever length the brief
deserves; a single line is a valid prompt. Don't restyle the user's wording
yourself — that just stacks a second rewrite and gives their intent another
chance to drift. Add only: grounding facts for a real subject, and the
behaviours below that the user had no way to know about.
## Grounding
What you specify is preserved; what you leave out is filled in for you. So
when a subject has a real, checkable appearance, research it first and put
what you found into the prompt.
Ground whenever someone who knows the subject could watch the clip and say
"that's not what that is": a named person or place, a landmark, a particular
vehicle or machine, anything technical, anything culturally specific, or a
period setting. Skip it for generics ("a dog on a beach") — there is no fact
to get wrong.
To ground: search for VISUAL references — three good photographs beat any
amount of prose. If you can analyze images, analyze what you find. Then put
only what a camera could see into the prompt: silhouette and proportion,
materials and finish, specific colours, distinctive details, era-correct
context.
## Audio
Audio is generated by default on models that support it (Veo, Kling v3,
Seedance), whether or not you mention it. Leaving it out gets you invented
sound rather than silence. Name the ambient sound, the music and any speech
separately — each lands as its own layer — and say "no music" when you do
not want it. In `falvid.py`, audio is **off by default** (`--audio` opt-in),
so you control this explicitly.
## Multi-shot sequences
Some models (Veo 3.1, Kling 3.0) support multi-shot sequences inside a single
generation. Use explicit shot language: "SHOT ONE ... HARD CUT. SHOT TWO ..."
produces real cuts, and "one continuous unbroken shot" gets an uncut take.
Consecutive shots must contrast in scale, location or colour or the cut will
not read as one — near-identical coverage blends back into a continuous take.
## Quoted lines and speech
A quoted line becomes speech only if a speaker is visible on camera. Without
one it tends to render as burned-in text instead. Quote the line, describe the
speaker, and add "no on-screen text, no subtitles" to prevent burned-in
captions.
## Continuation / chaining
To extend a clip or chain segments, use the continuation workflow: open with
"Continue this video from its final frames:", re-establish the subject and the
moment the clip ended on, then describe what happens next. Plan segments of
15 seconds or less so you can feed outputs back in.
For fal.ai models, this is done by passing the last frame (or the clip
itself) as the `--image` for the next `animate` call with a continuation
prompt. Not all models support this — Veo 3.1 and Kling v3 handle it best.
## What you leave out is the model's call
Only the start frame is pinned; every frame after it is generated. What you
leave out of the prompt is the model's call, not yours. If a detail matters —
a colour, a material, a direction of motion, whether a hand is open or
closed — say so explicitly.Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| canvas-designanthropics/skillsCreate beautiful visual art in .png and .pdf documents using design philosophy. You should use this skill when the user asks to create a poster, piece of art, design, or other static piece. Create original visual designs, never copying existing artists' work to avoid copyright violations. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| docxanthropics/skillsUse this skill whenever the user wants to create, read, edit, or manipulate Word documents (.docx files) or Word templates (.dotx files). Triggers include: any mention of 'Word doc', 'word document', '.docx', '.dotx', or requests to produce professional documents with formatting like tables of contents, headings, page numbers, or letterheads. Also use when extracting or reorganizing content from .docx or .dotx files, inserting or replacing images in documents, performing find-and-replace in Word files, working with tracked changes or comments, or converting content into a polished Word document. If the user asks for a 'report', 'memo', 'letter', 'template', or similar deliverable as a Word or .docx file, use this skill. Do NOT use for PDFs, spreadsheets, Google Docs, or general coding tasks unrelated to document generation. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| pdfanthropics/skillsUse this skill whenever the user wants to do anything with PDF files. This includes reading or extracting text/tables from PDFs, combining or merging multiple PDFs into one, splitting PDFs apart, rotating pages, adding watermarks, creating new PDFs, filling PDF forms, encrypting/decrypting PDFs, extracting images, and OCR on scanned PDFs to make them searchable. If the user mentions a .pdf file or asks to produce one, use this skill. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |
| pptxanthropics/skillsUse this skill any time a .pptx or .potx file is involved in any way — as input, output, or both. This includes: creating slide decks, pitch decks, or presentations; reading, parsing, or extracting text from any .pptx or .potx file (even if the extracted content will be used elsewhere, like in an email or summary); editing, modifying, or updating existing presentations; combining or splitting slide files; working with templates (.potx), layouts, speaker notes, or comments. Trigger whenever the user mentions "deck," "slides," "presentation," or references a .pptx or .potx filename, regardless of what they plan to do with the content afterward. If a .pptx or .potx file needs to be opened, created, or touched, use this skill. | Stars: 198k | Updated | Claude CodeCodex | Verified | Install |