Seedance prompt short

Skill 381487190/opengod/skills/seedance/seedance-prompt-short

OpenGod - 14-module AI super skill. Fuses 18 of Top 50 GitHub AI projects.

Install

npx -y skills add 381487190/opengod --skill seedance-prompt-short

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

This skill should be used when the user asks for a compact Seedance 2.0 prompt, short Chinese prompt, prompt compression, 30-100 word output, or removal of unnecessary prompt language.

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

2.9 KB, 529 tokens by cl100k_base, as published. Nobody here has run it

seedance-prompt-short

Compress Seedance prompts without losing the production signal. A short prompt still needs mode, subject, action, camera, light, sound when useful, and constraints. Remove filler before removing physical details.

When sequence state is present, compression must preserve continuity locks, exact reference tags, actual opening state, current clip action, endpoint, completed beat exclusions, and reserved future beats. Do not compress away the words that keep a continuation from replaying completed action or leaking future action.

Intent

Compression is an act of judgment about what the user loves most. What survives the cut is the soul of their shot; everything else goes first. If the user would mourn a deleted word, it was never filler.

Compression Priority

Preserve in this order:

Reference tags and their role.
Subject or product identity.
Action verb and visible endpoint.
One camera move.
Physical light source or atmosphere.
Audio cue or silence instruction.
Safety, IP, or continuity constraint.
Sequence state clauses: actual opening state, continuity locks, completed beats, and reserved beats.

Delete generic adjectives, duplicate style labels, obvious background details, secondary camera moves, and secondary actions before deleting preservation constraints.

For bilingual or mixed-language compression, load [ref:multilingual-community-examples]. Keep only the language mix that clarifies reference roles, dialogue, camera terms, or safe production constraints.

Compact Templates

Need	Template
T2V	`[Subject] [action and endpoint] in [scene]. Camera: [one move]. Light/style: [physical source]. Sound: [cue]. Constraint: [risk/continuity].`
I2V	`@Image1 preserved; only [motion/light/camera] changes. Camera: [one move]. Sound: [cue]. Constraint: [what must not change].`
V2V	`@Video1 controls [motion/camera/timing] only; new subject [anchor]. [Action]. Do not transfer [identity/scene/logo].`
Chinese	`@Image1为参考，严格保持[主体]不变；仅加入[动作/光线/镜头]。声音：[提示]。`

Output Contract

Return one compact prompt, ideally 30-100 English words or an equivalent Chinese prompt when the user asks for Chinese or maximum compression. Include a one-line note only if something important was removed.

Gives 0 of the 12 instructions most prompt engineering skills give in 529 tokens

Counted across 563 of the 626 authors here whose files we hold, read 2026-08-06

ask at most three clarifying questionsin 22 of 563, across 15 files
respond in the user input languagein 14 of 563, across 9 files
preserve the original intentin 13 of 563, across 11 files
Establish baseline metrics and collect representative examplesin 12 of 563, across 2 files
Identify failure modes and prioritize high-impact fixesin 12 of 563, across 2 files
Apply prompt and workflow improvements with measurable goalsin 12 of 563, across 2 files
Roll back quickly if quality or safety metrics regressin 12 of 563, across 2 files
validate changes with tests and roll out in controlled stagesin 12 of 563, across 2 files
generate quantitative baseline performance reportsin 12 of 563, across 2 files
create representative test scenariosin 12 of 563, across 2 files
treat prompts as codein 12 of 563, across 5 files
test prompts on diverse inputsin 12 of 563, across 8 files

Said here and by no other author read

preserve reference tags and roles
preserve subject or product identity
preserve action verb and visible endpoint
preserve one camera move
preserve physical light source or atmosphere
preserve audio cue or silence instruction

Grouped from the skills themselves: near-identical wordings counted once, and counted by distinct author, so one author publishing three of these counts once. Length counted with cl100k_base; the agent that loads this file may tokenize it differently.