How to use the Seedance Pipeline Planner agent skill

What the seedance-pipeline skill makes the agent do

Once installed, the agent answers questions about Seedance 2.0 AI video generation with a structured plan instead of guesses. The skill covers web app workflows, API planning with BytePlus ModelArk, Volcengine, Runway and provider or router APIs, professional production, post-production, and community tools such as ComfyUI.

For each question the agent is designed to identify which workflow applies, which inputs are needed, what steps follow, and what risks apply. It also flags which claims need checking against official sources before you rely on them.

• Web workflow: Dreamina/Jimeng surface, references, prompt and output review.
• API workflow: model ID, authentication, file handling, task creation, polling, cancellation and retrieval.
• Professional production: treatment, shot list, continuity ledger, review loop and delivery QC.
• Post-production: editing, stitching, stabilization, audio cleanup, subtitles, colour, localization and versioning.
• Community tools such as ComfyUI, which the skill labels as unverified unless sourced.

Install the skill (download, unzip, or upload)

Download the seedance-pipeline ZIP from this page. You have two options.

For a personal install, unzip the folder into ~/.claude/skills/. For a single project, unzip it into the project's .claude/skills/ folder so it applies only there.

If you use the Claude apps, you can instead upload the skill under Settings → Capabilities → Skills.

Trigger the skill with natural requests

The skill's description tells the agent when to use it: questions about Seedance 2.0 workflow operations, API planning, BytePlus ModelArk, Dreamina/Jimeng surfaces, provider or router APIs, China-facing surfaces, ComfyUI, post-production, stitching, batch workflow, or integration planning. You do not need a special command. Describe the job in plain language and name the platform or deliverable when you can.

Good requests include a goal, a deadline or deliverable, and the surface you are considering. The skill is written for professional jobs, so it works best when you give it production context.

• Plan a 30-second Seedance 2.0 spot for a client, due in three weeks, with a localization pass.
• Which Seedance 2.0 surface should I use for a batch of product clips, and what inputs do I need?
• Scope an integration with BytePlus ModelArk for a video app. What auth, task and polling steps are involved?
• Give me a QC preflight checklist before I hand this Seedance cut to the post house.

What you need before you start

The skill's SKILL.md contains only instructions. It points to reference files (api-status, model-name-map, api-workflow, pro-filmmaking-standards and delivery-qc) for detailed guidance. Those reference files were not included in the files reviewed for this listing, so the depth of the guidance cannot be verified here. Without them, the agent cannot check current status, so treat its answers as a starting plan.

For any API work you need your own provider accounts and API keys for services such as BytePlus, Volcengine, Runway or fal. The skill does not supply these.

• The referenced files: api-status, model-name-map, api-workflow, pro-filmmaking-standards, delivery-qc.
• Provider accounts and API keys, which you supply yourself, for any API work.
• The skill is MIT licensed.

Safety notes: scripts and network access

According to the listing, this skill runs no scripts and uses no network access. It is instructions only, so installing it does not run code on your machine. Any API call you make later is your own action with your own keys.

The skill warns against using unofficial provider or ComfyUI community nodes in production code without independent verification. Check any pricing, availability or endpoint against the provider's official documentation before you build on it.

• Runs scripts: no.
• Uses network: no.
• Verify pricing, availability and endpoints against official provider docs.
• Treat unofficial community nodes as unverified.

Tips for better answers

The skill tells the agent to recheck current status rather than rely on memory. Ask the agent to state which claims need verification, and to list the next artifact to create, such as a brief, shot list, prompt batch, review packet, localization matrix or QC preflight.

Keep your own record of what the agent says about pricing and availability, and confirm it on the official site before committing budget.

• Name your deliverables and deadline in the first message.
• Ask for the workflow path and risks, not only the answer.
• Ask which claims need checking against official sources.
• Use the output as a checklist to confirm, not as final vendor information.