Model Prompt Adapters — prompt skill for prompts by model
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.
Dated syntax adapters for Veo / Gemini Omni, Kling 3, Seedance 2.x, Hailuo 2.3, Luma Ray and Runway Gen-4.5.
- Type: Prompt · Prompts by model
- Slug:
craft-model-adapters - 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
- Cost: Free — text only.
- Runtime: Seconds
What it does
The per-model layer behind Cinematic Grammar, also usable on its own to port a prompt: element order (camera first for Veo, Luma and Runway; subject first for Kling and Seedance), bracket meaning (camera on Hailuo, shot labels on Kling, timestamps on Veo, sound on Seedance), negatives (positive-only for Luma and Runway, no 'no'/'don't' on Veo, constraint sentences on Seedance) and the rewriters to switch off for exact camera work (prompt_expansion_mode: "disabled" on MiniMax H3 Max routes, enable_prompt_expansion: false on Wan 3). Hailuo's bracket commands are documented for Hailuo 2.3 and 02, while varg serves MiniMax H3, whose examples use plain-language camera, so on H3 the compiler falls back to Kling-style prose. Every adapter file is dated and regenerated from the varg catalog: the Sora 2 API closed on 2026-09-24, Veo 3.1 Preview left the Gemini API on 2026-10-22 (varg routes veo_3_1 through fal; Google points to Gemini Omni Flash), Kling 4.0 was switched off on varg on 2026-10-22, and Luma and Runway adapters serve prompts users take to those tools, since neither is in the varg catalog.
Use when
- Porting a prompt from one video model to another
- A model ignores the camera direction — check element order, brackets and rewriter settings
- A model was added to or removed from the varg catalog
Not for
- Inventing syntax a provider has not documented
- Sora 2 (API shut down 2026-09-24)
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, seedance_2_fast_preview, minimax_h3, minimax_h3_max, minimax_h3_max_turbo, wan_3
More prompts by model skills
- fal Prompting: Prompt specific model families well: Kling, GPT Image 2 and Happy Horse guides, loaded only when needed.
- Replicate Prompt Images: Write better image prompts and edits: photographic language, lighting, text rendering, consistent characters.
- Replicate Prompt Videos: Prompt video models like a director: shots, camera motion, four audio layers, dialogue, time-coded beats.
- Video Prompting Guide: Write prompts that video models follow: shot types, camera moves and per-model tips from Veo to Sora.
- FLUX 3 Prompt Doctor: Diagnose a FLUX 3 video brief before spending: resolve missing, conflicting or schema-changing asks.
- FLUX Image Best Practices: Prompt FLUX image models well: structure, JSON prompts, hex colors, typography and multi-reference edits.