seedance-2-5-skill
Plan and generate controllable Seedance video using Seedream 5.0 Pro storyboards and Seedance 2.5 (Seedance 2.0 as the fallback). Use for consistent people, products, objects, food, or scenes; storyboard-to- video; reference-to-video; first-and-last-frame image-to-video; extensions; and Atlas Cloud media generation.
How do I install this agent skill?
npx skills add https://github.com/atlascloudai/atlas-cloud-skills --skill seedance-2-5-skillIs this agent skill safe to install?
- Gen Agent Trust Hubwarn
This skill facilitates video generation via the Atlas Cloud API. It is generally safe but contains a configuration option that allows API tokens to be passed as command-line arguments, which exposes them to other users on multi-user systems. It also includes instructions for installing tools via a remote shell script from the vendor's repository.
- Socketpass
No alerts
- Snykfail
Risk: CRITICAL · 2 issues
What does this agent skill do?
Seedance 2.5 Skill
Language route
- For an English request, follow this file and the
*.mdreferences. - For a Chinese request, read the Chinese workflow
first, then use the matching
*.zh-CN.mdreference files. - Keep model IDs, JSON keys, commands, media placeholders, and native-audio symbols exactly as code. Do not translate them.
Do not force every request through one image-grid pipeline. Choose the video route first, then enable only the preparation modules the job needs.
1. Choose the creative route
| Need | Route | Inputs | Result |
|---|---|---|---|
| One short, simple scene | T2V | Text prompt | One self-contained shot |
| A multi-shot sequence with readable panels | R2V storyboard | One complete storyboard image | One request turns panel order into a continuous video |
| A clip controlled by people, product, scene, or style assets | R2V asset references | A small role-specific asset pack | One clip built from those references |
| A shot with an exact beginning and ending | I2V shot pair | Start and end keyframes | One independently reviewable shot |
| An uninterrupted action exceeding the supported duration | Extend / chain | Previous generated tail frame and next prompt | Continuation of the same shot |
| One continuous piece carrying several events | Staged whole-short | Text plus optional references | One request covering ordered stages, each landing on a stated end state |
| A scoped change to an existing video | Editing | Source video plus target references | The source with only the named region or element changed |
| A bridge between two finished clips | Seamless transition | Two videos | Generated bridge content between them |
| Motion, blocking, and camera taken from a 3D preview | Blockout reference | Coarse or fine blockout video plus look references | Final render following the blockout's timing and staging |
Use Seedance 2.5 as the executable default, and Seedance 2.0 as the fallback when the selected provider does not expose 2.5 or the job needs a 2.0-only parameter. Check the actual limits of the model you submit to.
The last four routes depend on capabilities that differ per model and per provider. Verify availability before offering one; see capabilities for the 2.0/2.5 comparison and for which published capabilities are platform features rather than API parameters.
Storyboard or individual keyframes
- Use the whole storyboard image in one R2V request by default whenever individual panels remain visually readable. Do not crop it first. Seedance can interpret the ordered panels as one continuous multi-shot video.
- Use I2V shot pairs only when independent reshoots or precise start/end states matter more than the transition quality of one R2V generation.
- Do not upload every storyboard cell as a default R2V asset pack. Use several R2V images only when each has a distinct role, such as subject, product, setting, style, or motion reference.
2. Enable only needed preparation
Subject brief
Use this only for a person, product, prop, hand, vehicle, or scene that must recur. Record 3–5 invariants: silhouette or proportion, signature material or wardrobe, key colour, and any must-preserve marking. Skip it for a one-off atmosphere shot.
For a recurring person, make a clean face close-up and a separate full-body reference. Do not use a front/side/back composite as the identity input; it can be interpreted as multiple people. Multi-angle product references remain useful when the object itself must be shown from several sides.
Keyframes
Create a start keyframe for every I2V shot. Add an end keyframe only when the shot must land on a specific action, composition, product pose, or hand position. Use one clean scene per keyframe.
Storyboard
Use an existing storyboard directly as the R2V reference. Seedance normally understands panel order and does not require panel numbers, dividers, arrows, or notes to be removed in advance. Make a clean copy only after a test generation actually renders an unwanted divider, number, caption, or multi-panel layout.
When a multi-shot request has no storyboard, use the Seedream template in
prompt templates to create one board.
Display that board in the host UI and inspect it yourself. Generating the
board and every video after it are separate paid steps: quote each one (or show
the runner's [spend-plan]) and wait for the user's yes before you submit it.
For a supplied storyboard, display it unless it is already visible in the conversation. Check planned order, readable key beats, recurring-subject consistency, and content-specific constraints such as anatomy, product form, or critical text. If the board visibly failed, tell the user what is wrong; regenerating it is another paid image, so quote it and wait for their yes.
If the route deliberately changes to I2V shot pairs, crop panels only then. Inspect the layout first: automatic crops can verify position, not whether the image model drew the intended panel layout.
3. Design continuity and cuts
First-and-last frames control one shot; they do not mean every shot must inherit the prior clip's tail.
| Transition | Use prior tail as next start? | Design rule |
|---|---|---|
| One uninterrupted action | Yes | Generate in sequence and inspect the seam |
| Hard cut to a new angle, place, product, or time | No | Design each shot independently |
| Match cut | Usually no | Match movement direction, shape, colour, or composition |
| Occlusion or whip transition | No | End with the occluding action; start the next shot inside or after it |
| Insert or cutaway | No | Use an object, environment, or product detail as a bridge |
Put important cuts, matches, and occlusions in the storyboard and prompt. Do not rely on a cross-dissolve to repair unrelated shots.
4. Write the video prompt
Scope: put each instruction where it applies
Before writing blocks, sort what you know by what it governs. Instructions in the wrong place are the most common cause of drift — a global rule written inside beat 1 stops applying at beat 4.
| Scope | Governs | Contents |
|---|---|---|
| Global | The whole piece | Film type, scene, style, one-sentence premise, camera principle |
| Locks | Anything that must not drift | Identity, reference roles, audio source, continuity, negatives |
| Time | One beat or stage | Stage events and their end states |
Restate the two or three most expensive locks at the physical end of the prompt; recency helps. That is a placement convention, not a fourth scope — the content still belongs to Locks and appears there first.
This mirrors the model-agnostic spec format in
the Universal Video Prompt Skill. Use
that skill when one brief has to run on more than one model; use this file for
Seedance-specific writing. This is a required companion for a complete Seedance 2.5
Skill setup. If the link does not resolve, help the user install
universal-video-prompt-skill before continuing; do not invent the missing shared
specification.
Blocks
Use only the blocks that affect the shot:
[subject/reference binding]
+ [one observable action]
+ [space and important object relationships]
+ [one primary camera move, coherent composite move, or a cut]
+ [light/style when it matters]
+ [audio or dialogue when enabled]
+ [end state, for any stage that must land somewhere specific]
+ [must-preserve constraints]
End states carry multi-event work
For anything with more than one event, state what is visibly true when each stage ends. This is the highest-leverage single addition to a multi-stage prompt: it converts "keep it consistent" into something the model can target and you can check.
weak: the two of them keep working on the bouquet
strong: end state: the florist holds the bouquet in the left hand;
the scissors are back on the right side of the bench
An end state must be visible. "She feels relieved" is not one; "her shoulders drop and the frown clears" is. Read long video for the staged structure in full.
Time granularity: decide before writing beats
Granularity is a prior decision. Writing beats at second precision and then downgrading means rewriting them.
| Granularity | Use when |
|---|---|
| None — event order only | One continuous action, mood pieces, single shots. Timestamps here fragment the shot |
| Stages + end states | Most narrative work. Default |
| Second-level | Only under an external hard constraint: music, lip sync, reference handoff, a beat that must land at a fixed time |
Infer it when the input settles it — a supplied music or voiceover track means second-level, a stated mood piece means none, an explicit fixed beat means second-level. When the request is a multi-event narrative with no external constraint, ask, and recommend with a reason rather than presenting a bare menu.
Timestamps allocate a time budget; they are not frame-accurate edit points, and actions may land slightly before or after a boundary. Do not demand impossible density such as three distinct actions inside one second.
- Name reference roles explicitly, for example
Image 1: person,Image 2: product,Image 3: kitchen setting. - For multi-shot R2V, list
Shot 1,Shot 2, andShot 3in event order. Set duration in provider controls rather than forcing exact seconds in text. - Prefer one primary movement per shot. A composite movement is valid when its direction, relation to the subject, and speed express one synchronized intent.
- When native audio is enabled, use
()for music,<>for sound effects,{}for dialogue, and【】for on-screen captions. - State only constraints that are costly to redo.
Read the reference matching the job:
| File | Read it for |
|---|---|
| prompt templates | Route-specific templates |
| prompt blocks | Reusable camera, audio, constraint patterns |
| long video | Staged structure, end states, timestamp rules |
| multi reference | Binding many assets without confusing them |
| real person | Believable human subjects, and when to omit the detail |
| transitions | Which transitions to generate and which to edit |
| editing and extension | Changing or continuing existing video |
| capabilities | 2.0 vs 2.5 limits; platform features vs API parameters |
| model profile | Measured per-model behaviour and compile notes |
| cinematography | Detailed visual decisions |
| troubleshooting | Fault-specific fixes |
| execution adapters | Runner and adapter configuration |
5. Generate, review, and finish
- For a generated storyboard, create only the still first, display it in the conversation, and inspect it before any video request.
- For a supplied storyboard, display the input unless it is already visible, then verify it fits the chosen route.
- If the board passes review, quote one representative video pass (model, duration, resolution, estimated cost) and wait for the user's explicit yes in a new message before submitting it. Otherwise tell the user what is wrong with the board; regenerating it is another paid image, so quote it and wait for their yes. Quote and confirm every later billable segment the same way.
- Review identity, locks, stage end states, composition, motion, seam, and audio in that order, and stop at the first failure — later checks are wasted effort on a wrong identity. Regenerate only the failed shot or segment.
- Generate chains in order because the next segment needs the real prior tail. Generate cut-based clips independently and edit the planned transition.
The bundled script is a draft assembler, not a colour-grading or music-mixing system.
Atlas execution layer
Keep creative route selection independent from how a job is submitted. The
default models are Seedream 5.0 Pro for stills and Seedance 2.5 for video
(bytedance/seedance-2.5/text-to-video, .../image-to-video,
.../reference-to-video), falling back to Seedance 2.0 when 2.5 is not offered
or cannot take the job. A model the user names always wins; verify that it
supports the route first.
With Atlas MCP tools (the plugin or a connector)
This covers the Atlas Cloud plugin in Codex, Claude Code and Cursor and the
Atlas Cloud connector in ChatGPT, Claude, WorkBuddy and other MCP clients.
Whenever Atlas MCP generation tools (atlas_generate_image,
atlas_generate_video, ...) are in your tool list, they are the Atlas route
and the only generation route, because they quote every job before it runs. Use
them for every stage and ignore the API-key, Skill-install and runner
instructions below, which are for standalone setups. Never run
scripts/generate.mjs and never call the Atlas REST API yourself: both submit
straight away, with no quote. Use the runner only when the user explicitly asks
for a terminal, CI, or batch run.
- Generate every still — storyboard boards, keyframes, reference images — with
atlas_generate_imageon an Atlas image model (default Seedream 5.0 Pro). Never use the host's built-in image generator for any stage of this workflow, even for a quick draft: the user asked for an Atlas workflow and is billed and shown results on that basis. - Each generation tool returns a live quote and stops. When you relay the first
quote, also say which Atlas model will make each later stage (for example
"storyboard: Seedream 5.0 Pro; video: Seedance 2.5") and that each step is
quoted before it runs; then wait for a yes in a new message. Do not ask the
user to choose models or a budget up front. When you picked the model, pass
why_this_modelwith one short reason. - In apps that show Atlas cards, the quote is a card with a confirm button:
point the user to it. If they confirm on the card, the card submits and
tracks the job, so do not call the generation tool again for it. If they
decline or change the request, withdraw the old quote with
atlas_cancel_quote. - Upload local references with
atlas_get_upload_url; for a file already in the user's Atlas Asset Library, useatlas_search_assetsand thenatlas_use_assets(or passasset://<asset_id>). Poll withatlas_get_prediction(call it again right away — the server already waits between checks), and keep polling in the same turn until each stage finishes. - If the Atlas tools are missing, the plugin or connector is not signed in:
tell the user how to sign in. Do not install a Skill, look for an API key, or
run
scripts/generate.mjsto get around it — those routes submit with no quote.
Without Atlas MCP tools
In an agent conversation, use the Atlas Cloud Skill as the direct
generation route. It can discover a model, upload local media, submit an image
or video request, poll, and retrieve outputs. Before each billable submission,
tell the user the model, duration, resolution, and estimated cost, and wait for
an explicit yes. Report Execution: atlas-skill only when it actually
submitted the generation. Use atlas-cli only when the user explicitly selects
a terminal, script, CI, or batch run. If the Atlas Cloud Skill is missing, help
install AtlasCloudAI/atlas-cloud-skills before selecting a fallback.
Before reporting that an Atlas Cloud API key is missing, check the credentials
in the selected execution process. For the REST runner, check
ATLASCLOUD_API_KEY first and ATLAS_CLOUD_API_KEY as a compatibility alias.
Do not infer credential availability from a different provider, plugin, or
process; each execution channel can have an independent credential scope.
If neither key exists, direct the user to
https://www.atlascloud.ai/console/api-keys?utm_source=github&utm_campaign=awesome-seedance-2.5-prompts-skills.
Never ask them to paste the key in
chat. Tell them to set ATLASCLOUD_API_KEY in the submitting process or the
host's secure environment settings, then refresh or restart the execution
session if needed. If the key exists in a parent or host configuration but is
absent from the submitting process, report an environment-scope mismatch
instead of saying the user has no key.
Billable task state machine
Apply these rules to every image and video generation:
- After submission, record the prediction ID and logical stage immediately.
- Treat
starting,queued,pending, andprocessingas active. Keep polling the same ID (see the interval below); never submit another task for that stage. - Treat
completedandsucceededas successful terminal states. Download and inspect the output before starting a dependent stage. - Treat
failed,timeout, andcanceledas terminal failures. A new task is the user's decision: report the old ID and the extra cost, and wait for their yes. - A zero or missing processing-time field, delayed output, a local polling timeout, a stopped turn, or a temporary status-query error is not proof of failure. Preserve the ID and resume polling.
- Interpret
continueas “resume the existing task,” never as permission to retry. Do not submit video while its required storyboard is still active.
Poll every 2 seconds on the Skill, CLI and REST routes: for atlas-skill,
repeat its prediction-result step with the same ID. For atlas-mcp, call
atlas_get_prediction again with the same ID as soon as it answers (it waits
up to about 20 seconds per call). The
MCP server performs a single status lookup per tool call; the agent owns the
loop. The bundled REST and CLI adapters enforce the interval in code. A status
lookup is read-only and must never be replaced with another generation call.
When the runner must resume, set execution.resumePredictionIds.<stage> to the
existing ID. Supported stage keys include grid, ref1, ref2, seg1,
shot1, and clip1. Never create a replacement merely because a prior polling
process ended.
scripts/generate.mjs cannot invoke an agent Skill or MCP server. It is a
separate batch runner: it defaults to atlas-rest, and can use atlas-cli
only when explicitly set in execution.adapter. It never selects CLI
automatically. See execution adapters.
# Generate a storyboard only. The runner prints [storyboard-preview] with an
# absolute path; display and inspect that image before starting video work.
GRID_ONLY=1 node scripts/generate.mjs scripts/myjob.json
# Run only the first independent clip or segment as a quality gate.
CLIPS_MAX=1 node scripts/generate.mjs scripts/myjob.json
SEGS_MAX=1 node scripts/generate.mjs scripts/myjob.json
# Every run, GRID_ONLY and REFS_ONLY included, first prints a [spend-plan] line
# (new image jobs, billed video seconds and a plan id) and stops there. Show the
# user that plan and its estimated cost; after they approve, run again with the
# exact values the runner printed. An approval works once: re-running the same
# plan (for example after a failed segment) needs the user's yes again.
CONFIRM_SPEND_PLAN=<plan id> CONFIRM_SPEND_IMAGES=1 GRID_ONLY=1 node scripts/generate.mjs scripts/myjob.json
CONFIRM_SPEND_PLAN=<plan id> CONFIRM_SPEND_SECONDS=40 node scripts/generate.mjs scripts/myjob.json
{ "execution": { "adapter": "atlas-cli" } }
Available runner modes:
grid: crop a model storyboard and make one independent I2V clip per panel; suitable for a hard-cut montage.shot-pairs: make independent I2V shots fromsegments[], with afirstand optionallastkeyframe index.reference: send a small role-specific set of R2V references. A reference may be generated from a prompt or read from a local file.chain: continue one action by feeding the generated tail frame to the next segment. It is an alignment aid, not a guarantee of an invisible seam.
For fault-specific fixes, read troubleshooting.
How can the creator link this skill?
Add the canonical catalog link to the repository README so users can inspect current installs and available audits. The publishing guide covers the complete discovery path.
<a href="https://skillzs.dev/skills/atlascloudai/atlas-cloud-skills/seedance-2-5-skill">View seedance-2-5-skill on skillZs</a>