Omni-Rewriter: Skill for Cursor

.cursor/skills/omni-rewriter-h3-pe/SKILL.md

omni-rewriter-h3-pe is a skill for Cursor from WayneJin0918/Omni-Rewriter. It costs 59 tokens per session (1,293 once invoked), scanned A, original, Apache-2.0.

A set of rules for extending a framework that expands user intent into structured prompts for video and image tasks. Prompt expansion turns a short request into more detailed, validated instructions; it does not generate media by itself.

In plain words
What is it for?
Use it when adding or repairing prompt profiles, validators, renderers, adapters, and tests for H3 video or Seedream and Qwen-Image image workflows.
Why use it?
It keeps model-specific rules, validation, and output formats separate, reducing unsupported compatibility claims and malformed requests.

Skill for Cursor

Written for Cursor: installed under .cursor/.

This is WayneJin0918/Omni-Rewriter's own configuration. It tells Cursor how to work on Omni-Rewriter itself, so it is not a mod to install elsewhere. Copy it as a starting point and replace the rules that are about this project. Everything Omni-Rewriter configures →

Reuse

Borrowing it

Nothing to install: this file belongs to WayneJin0918/Omni-Rewriter. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.

Copy the file
curl -O https://raw.githubusercontent.com/WayneJin0918/Omni-Rewriter/main/.cursor/skills/omni-rewriter-h3-pe/SKILL.md
Clone the repo
git clone --depth 1 https://github.com/WayneJin0918/Omni-Rewriter

Made for: Cursor.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for omni-rewriter-h3-pe

README.md
[![agentmods](https://agentmods.dev/badge/skills/waynejin0918/omni-rewriter/omni-rewriter-h3-pe/github.svg)](https://agentmods.dev/skills/waynejin0918/omni-rewriter/omni-rewriter-h3-pe)
Your own site
<a href="https://agentmods.dev/skills/waynejin0918/omni-rewriter/omni-rewriter-h3-pe"><img src="https://agentmods.dev/badge/skills/waynejin0918/omni-rewriter/omni-rewriter-h3-pe/github.svg" alt="Measured on agentmods" height="20"></a>

Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.

agentmods 80×15 button for omni-rewriter-h3-pe

Your own site · 80×15
<a href="https://agentmods.dev/skills/waynejin0918/omni-rewriter/omni-rewriter-h3-pe"><img src="https://agentmods.dev/badge/skills/waynejin0918/omni-rewriter/omni-rewriter-h3-pe.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 59 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,293 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe. Third-party audits
  • NVIDIA SkillSpector pass 7 Sept 2026
How audits are shown
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00059 $0.01293
Opus 5 $0.00030 $0.00647
Sonnet 5 $0.00012 $0.00259
Haiku 4.5 $0.00006 $0.00129

Measured 13d ago against content hash d12028a2c11b, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-12, from the pricing page.

Security

Grade A, and why

omni-rewriter-h3-pe scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 13d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

.cursor/skills/omni-rewriter-h3-pe/SKILL.md · 100 lines

How it starts

The opening of the file, as written. The whole thing — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Omni-Rewriter PE skill

Mission

Omni-Rewriter is a model-extensible PE framework: transport-neutral intent → typed profile → deterministic validation/bounded repair → dialect render → optional explicit generation adapter. H3 and the initial image dialects are profiles, not framework limits. Prefer schema, validators, repairs, renderers, and public contracts over claims about private vendor internals.

Framework rules

  1. Expand is not generate. Expansion returns validated text/JSON; adapters and independent runners are opt-in consumers.
  2. A PE profile does not prove generation-runtime compatibility.
  3. Cite public runtime/API evidence, pin tested versions, and label untested paths unverified.
  4. Do not treat stock vLLM, model-specific vLLM forks, and vLLM-Omni as interchangeable.
  5. Preserve transport-neutral public contracts and keep model-specific mapping at profile/adapter boundaries.

When expanding video (Seedance)

  1. Set metadata.video_pe_profile=seedance (default remains H3).
  2. Support t2va / ref2va only for this profile; require duration_seconds.
  3. Emit SeedanceRewrite fields: style, summary, static/dynamic descriptions, subjects, optional reference_roles / stages / preserve / unused_materials, instruction, optional BGM, generate_audio. Follow public Seedance 2.5 habits (typed @Image / @Video / @Audio, role+exclude lines, observable stage end states, {dialogue} / <sfx> / (music) delimiters). Do not claim private vendor internals.
  4. Render via seedance_render=natural|fused|json (default natural = public 2.5 template) and seedance_ref_style=public|omni (default public type-local @Video 1).
  5. Never commit private dump markers or vendor-internal corpus metadata. See docs/dialects/seedance-pe.md and assert_sanitized_seedance_payload. No generation adapter in this profile pass.

When expanding video (LTX-2.5)

  1. Set metadata.video_pe_profile=ltx (default remains H3).
  2. Support t2va / i2va / l2va / fl2va / ref2va; require duration_seconds.
  3. Emit LTXRewrite fields that render to one flowing paragraph: action, movements, appearance, environment, camera, lighting, optional changes, audio when generate_audio is true. Follow the public LTX-2 guide (start with the action, ~200 words, no duration/resolution in the body). Image refs are generate-time --image flags, not prompt tokens.
  4. Render via ltx_render=paragraph|json (default paragraph). Optional LTXVideoRunner maps the paragraph onto python -m ltx_pipelines.distilled. Live generate is unverified. See docs/dialects/ltx-pe.md.

Read the full file on GitHub · 100 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 13d ago First seen · 100 lines · 59 tokens per session scan A d12028a2c11b

Subscribe to this mod's changes

omni-rewriter-h3-pe is a skill published in the GitHub repository WayneJin0918/Omni-Rewriter (85 stars, last pushed 23d ago), licensed Apache-2.0. It adds 59 tokens to every session and 1,293 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.

Related

Other skills, from other repositories

seedance-prompt

This skill should be used when the user asks to write, improve, translate, compress, or debug a Seedance 2.0 video prompt; mentions T2V, I2V, V2V, R2V, camera direction, prompt quality, or provides reference assets for a production-ready prompt.

1370998960-del/https-github.com-Emily2040-seedance-2.0 · 68 tokens

seedance-2-5-prompt-director

A skill for writing and improving prompts for Seedance 2.5, a video-generation system. It covers text-to-video, image-to-video, editing, video extension, references, camera movement, green-screen work, sound, and lip sync.

zhouwei713/seedance-2-5-prompt-director · 159 tokens

model-aware-image-prompt-engineer

Model-aware image prompt engineering for any agent or image generation workflow. Use when writing, improving, translating, debugging, or evaluating prompts for OpenAI image models, Gemini Nano Banana, Midjourney, FLUX, Qwen-Image, Z-Image, Stable Diffusion, SDXL, Pony, Illustrious, NoobAI, Animagine, HunyuanImage…

Emily2040/nano-banana-image-skill · 210 tokens

nano-banana-image-skill

Model-aware image prompting for Nano Banana Pro and Nano Banana 2. Covers generation, editing, continuity, text-in-image, style routing, and structured JSON output.

Emily2040/nano-banana-image-skill · 40 tokens

awesome-gpt-image-2

A structured library of reusable prompts for GPT-Image2, an image-generation model. It breaks prompts into parts such as subject, lighting, materials, composition, and visual details, with more than 20 templates.

AvalonLee/AstraCraft-Program · 0 tokens

mj-prompt

A prompt writer for Midjourney and niji, image-generation tools, using an East Asian visual style framework.

khanhhuyenngo985-sys/character-scene-design-skills · 53 tokens