seedance-submitter

name	seedance-submitter
description	Use when preparing, submitting, polling, or debugging Seedance 2.0 video generation jobs from product images, storyboard images, UGC scripts, voiceover copy, or promptPlan request JSON. Use for splitting scripts into render segments, uploading references, composing submit flags, submitting jobs through the PostPlus Cloud service, polling predictions, and handing off local render paths.
metadata	{"postplus":{"familyId":"media-production","familyName":"Media and Creative Production"}}

Seedance Submitter

Use When

• Preparing, validating, submitting, polling, or debugging Seedance 2.0 jobs routed through PostPlus Cloud.
• Inputs can be product images, storyboard images, UGC scripts, voiceover copy, or existing Seedance request JSON.

Do Not Use When

• The task belongs to ideation, QA, or another released skill listed in the handoff section.
• Required inputs are missing and guessing would change the result.
• Task class, hook logic, storyboard, or reference policy is still unresolved. Use video-generation and video-request-architect first.

Execution Boundary

• This runner validates, submits, and polls normalized Seedance requests. It must not make creative strategy, task-classification, or reference-policy decisions.
• Interpret sd2 as Seedance 2.0 Mini (video-seedance-2-mini-*), the current default tier, unless the user names another model. The standard Seedance 2.0 tiers (video-seedance-2-text / video-seedance-2-fast-text and their image-to-video counterparts — not -mini-) stay available when explicitly requested. Mini is the narrowest reference tier (see the reference-slot matrix below); when the task needs reference video/audio, choose standard or fast instead of mini.
• Reference-slot matrix (per endpoint, from the released schema): reference slots exist only on -text (text-to-video) endpoints. video-seedance-2-text and video-seedance-2-fast-text accept the repeatable --reference-image / --reference-video / --reference-audio; video-seedance-2-mini-text accepts --reference-image only (mini has no reference video/audio). Every -image (image-to-video) endpoint takes exactly one --image first frame and has no reference slots at all — multi-reference generation on this family means using a -text endpoint, not adding flags to an -image one.
• Released endpoint keys and their option enums (resolution, aspect ratio, duration bounds) are discovered from postplus media schema --json; they are not hard-coded here.
• Set duration within the endpoint's published bounds (see schema). For edit targets that fall between viable values, pick the nearest valid duration and set targetEditDurationSeconds.
• When targetEditDurationSeconds < duration, include timeline.activePerformanceEndSeconds and timeline.tailStrategy.
• If the target script or beat plan exceeds the endpoint's maximum duration, split into independent submit-ready segments. Do not submit one oversized request.

Source And Request

• Lock product/storyboard/reference media, script, duration, target edit duration, output root, source basis, and whether the user wants submission or command preparation only.
• Put timecoded action and spoken lines together in promptPlan.prompt_storyline. Put voice style, BGM, SFX, subtitle, and watermark constraints in promptPlan.audio.
• promptPlan.*, timeline.*, and targetEditDurationSeconds are authoring vocabulary, not Seedance request fields. Submit with the endpoint's CLI flags from the flat provider contract discovered from postplus media schema --json (--prompt, --image, --duration, --resolution, --aspect-ratio, --generate-audio, and the repeatable --reference-image / --reference-video / --reference-audio — reference flags only on the endpoints named in the reference-slot matrix above); the CLI rejects any flag outside the selected endpoint's contract. Compose the storyline/audio plan and any timeline.activePerformanceEndSeconds / timeline.tailStrategy instruction into the single --prompt narrative, then map the supported render bucket to --duration; timeline authoring fields have no separate flags or request fields, and there is no request-JSON envelope to author.
• Bind references explicitly: say what [image 1], [image 2], [audio 1], or [video 1] controls. Do not rely on same as previous, content above, or unbound local handles in final requests.
• Upload local reference media to hosted storage with postplus media-file upload --skill seedance-submitter --input-file <file> --mime <mime> --output <upload.json>, then pass output.mediaReference — a persistent postplus-media:// reference that never expires — into the Seedance media flags (for example --reference-image <output.mediaReference>). The hosted boundary exchanges it for a fresh signed URL at provider send time, so upload once and reuse the same reference across later submissions. output.data.download_url is a signed URL that expires; prefer output.mediaReference.

Review And Handoff

• Before submission, verify validation passed, every segment is self-contained, references are bound, required media exists, source basis is explicit, and the output path is durable.
• If a render is pending, return the segment id, manifest path, the output.data.id generation handle, the poll command postplus media poll --handle <output.data.id>, and expected local renders/ output path. The poll command waits in-command (re-checks every 8s for up to 45s per invocation; --wait-seconds 0 = single check): rerun the same command while the render stays pending instead of re-submitting or writing a tighter retry loop.
• Save a finished render to disk with postplus media-file download --url <fresh-output-url> --output-file <renders/...> (or --reference <postplus-media://...> for media that lives in hosted storage); provider output URLs are temporary, so download while fresh.

Stop Conditions

• Stop when required user intent, source evidence, or owned input artifacts are missing and guessing would change the result.
• If an owned CLI or script command fails, report the exact error and stop. Do not bypass the failure with metadata-only answers, readiness probing, local payload rewrites, fallback providers, or unpublished tools.

Public Command Boundary

• Choose the smallest matching command or workflow from the user input and run it directly.
• Readiness diagnostics: postplus doctor --skill seedance-submitter.
• Poll a pending render: postplus media poll --handle <output.data.id> (waits in-command up to 45s per invocation; rerun while pending).
• If an owned CLI or script command fails, report the exact error and stop. Do not bypass the failure with metadata-only answers, readiness probing, local payload rewrites, fallback providers, or unpublished tools.
• Use postplus media schema --json only when constructing or repairing an unknown request shape.
• Run the hosted submit with the generated command below; do not call provider APIs directly.

postplus media create video-seedance-2-fast-image \
  --prompt <prompt> \
  --image <image> \
  --output <result.json>

• If the CLI returns a quote-confirmation challenge, run postplus quote confirm --json --challenge-file <challenge.json> and retry with the returned token.