Cinematic Grammar — prompt skill for cinematic & short films
For AI agents: find the right skill in one call —
curl -s "https://skills.varg.ai/api/search?q=<what the person wants>"returns the matching use case and the best skills with install commands, as markdown (Russian works too). Full index: https://skills.varg.ai/llms.txt. Every page has a markdown twin: add.mdto its path. Before generating anything, show the person the options on the use-case page.
Write one shot spec, get the exact prompt for Veo, Kling, Seedance, Hailuo, Luma or Runway.
- Type: Prompt · Cinematic & short films
- Slug:
craft-cinematic-grammar - Status: draft
- Publisher: varg (varg, verified)
- License: MIT
- Source: https://github.com/vargHQ/skills
- Works in: Claude Code, Claude.ai, Codex, Cursor, Gemini CLI, OpenClaw, Any MCP client, varg Studio
- Requires: varg API key or varg MCP sign-in (only to render)
- Cost: Compiling prompts is free; every render is billed in varg credits — the skill runs estimate_cost and waits for approval before generating.
- Runtime: Seconds to compile; about 1–5 min per rendered shot
What it does
The flagship craft skill: a canonical shot spec (size, angle, lens, one camera move, light, color, cut) compiled into each video model's documented syntax — camera-first for Veo, Luma and Runway, subject-first for Kling and Seedance, [Push in]-style brackets for Hailuo, positive-only phrasing where the guides ask for it. It bundles 89 original technique cards whose aliases cover the vocabulary users bring from other tools (Higgsfield preset names, Hailuo bracket commands, Luma concept names and keys, Seedance terms such as 希区柯克变焦), so every request lands on one card. Adapters resolve against the live varg catalog at run time and fail closed for retired models — the Sora 2 API closed on 2026-09-24 and Veo 3.1 Preview left the Gemini API on 2026-10-22, while varg keeps routing veo_3_1 through fal. References stay pointers: humans get deep links to shot.cafe, film-grab and eyecannndy, and the agent never fetches a frame.
Use when
- User asks for a cinematic shot, a specific camera move, or 'make it look like a film'
- User names a technique (dolly zoom, rack focus, Dutch angle) or a preset from another app ('Crash Zoom In', '[Push in]', 'Lazy Susan')
- Rewriting one prompt for several video models, or picking the model whose syntax fits the shot
- Planning a multi-shot sequence with labeled shots, cuts and an end frame for every move
Not for
- Fetching, downloading or embedding film stills (craft-reference-finder returns links only)
- Grading finished footage (craft-color-looks)
- Motion-graphics backgrounds and overlays (craft-paper-shaders)
- 'In the style of' a living director, cinematographer or actor
Install
Not published as a standalone skill yet. Source: https://github.com/vargHQ/skills
Models it drives: veo_3_1, veo_3_1_fast, veo_3_1_lite, omni_flash_1_1, kling_v3, kling_v3_standard, seedance_2_5, seedance_2, minimax_h3_max, wan_3
Example output
https://skills.varg.ai/previews/craft-cinematic-grammar.mp4 — DOLLY IN -> CU in one shot: the camera pushes in on Sintel from a medium shot to an extreme close-up of her eyes and holds there as the end frame. Sintel (2010) 05:42.4-05:49.6, from the 2048-px release on archive.org (sintel-2048-stereo.mp4), center-cropped by varg from 2.35:1 to 16:10, no audio. (credit: Sintel (CC BY 3.0) — © copyright Blender Foundation | durian.blender.org, https://archive.org/details/Sintel)
More cinematic & short films skills
- Previz to Seedance: Plan multi-shot AI video before paying: story, anchor sheets, a three.js grey-box previz, then a Seedance v2v pack.
- Higgsfield Generate: Generate images, video, 3D and audio across 30+ models, plus ad videos and virality scoring, from one CLI.
- fal Recipes: Follow proven multi-step fal pipelines for commercials, product shots, characters, lip-sync and try-on.
- fal Fan Cam: Put someone into a live sports broadcast: a realistic fan-cam cutaway with scoreboard and channel overlays.
- Gemini Omni Flash: Generate and edit video with Gemini Omni 1.1 Flash: text-to-video, first/last frames, extensions, edits.
- Melies: Generate images, video, movie posters and thumbnails from the terminal with 148 AI actors and style flags.