agentleFS
381487190381487190/opengod/skills/seedance/seedance-vocab-en/SKILL.md

seedance-vocab-en

Skill0 starsChanged 3 months agoMITView on GitHub

This skill should be used when an English Seedance 2.0 prompt is slop-heavy, generic, padded with empty quality words, tripping false-positive filters, or needs precise English production vocabulary for camera, lighting, motion, VFX, audio, and constraints.

2 other projects use a version of this, 1 of them edited.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If a camera, microphone, light meter, or stopwatch cannot detect it, rewrite it.

    line 32

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

seedance-vocab-en

English is the default prompting language and fails in two ways at once: slop (empty evaluation words that add tokens and no signal) and false positives (vague threat-adjacent wording that trips the heaviest moderation surface). The cure for both is the same: concrete production English. Preserve reference tags exactly: @Image1, @Video1, @Audio1 must never be reworded.

Intent

English is where most users think, and where most prompts quietly rot. The soul of this vocabulary is precision as kindness: give people exact words so their excitement survives contact with the model, and so honest prompts stop being mistaken for dangerous ones.

Usage Rule

If a camera, microphone, light meter, or stopwatch cannot detect it, rewrite it. Every sentence should name something visible, audible, or measurable: subject, visible action, camera, light source, sound, constraint.

FunctionEnglish wording
Cameraslow push-in, locked medium shot, stable lateral tracking, pull back to reveal, macro close-up
Lightingsoft backlight, warm practical light from the left, cool moonlight rim, wet asphalt reflecting neon
Motiona slow head turn that stops, droplets merge and slide down, fabric settles after the gesture
Audioquiet room tone, one clear spoken line in quotes, no music until after the line
Constraintskeep the logo, label, and shape unchanged, one action, one camera move, nothing else moves

De-Slop Pass

Strip quality adjectives before adding anything: cinematic, epic, stunning, masterpiece, 8K, ultra-realistic, award-winning, hyper-detailed all delete or convert to one observable detail each. A prompt that earns "epic" names the crowd size, the lens distance, or the structure height instead of the word.

Filter-Aware Wording

English homonyms read as threats to filters: shoot the scene, kill the lights, gun it, dead silence, blow up the image. Use the production synonym (film the take, cut the lights to black, accelerate hard, held silence, enlarge to full frame). This is clarity for safe prompts only — never evasion. Anything genuinely risky (minors, real-person likeness, sexual or graphic content) routes to [skill:seedance-filter] for its boundary rule, not to a rewording.

Compact Pattern

@Image1 is the reference; keep identity, color, and shape unchanged. Only [motion/light/camera] changes. Camera: [one move]. Sound: [one cue]. Constraints: [lock].

Load references/vocab/en.md for the full function-organized vocabulary, slop traps, and filter-trip repairs. Load [ref:anti-slop-lexicon] for the core replacement rule and [ref:filter-vocab] for the full false-positive repair table.

Output Contract

Return the de-slopped English prompt, each replacement made (slop → observable detail), any filter-trip repair applied, and unchanged reference tags.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.