video-batch-runner — Claude Code skill: batch video generation

What video-batch-runner does

The skill handles video generation end to end in PostPlus: it reads the brief, inspects the supplied media and picks the matching Seedance 2.5 mode. Six modes are covered — text-to-video, first-frame, first-last-frame, reference video, video edit and video extend. Field names and allowed values come from the live schema, fetched with postplus media schema --endpoint <endpoint> --json, so nothing is guessed from memory.

Each clip gets its own self-contained prompt: subject and action, product behaviour, scene, camera, timing, sound and the role of every reference. Local paths, HTTPS URLs and existing PostPlus references go straight into the matching role flags, with no pre-upload step. When a clip exceeds the supported duration, the skill advises splitting it into independent prompts before any credits are spent.

Finished output is pulled down with postplus media-file download. Inside a batch only postplus_cli_hosted_media_content_policy_blocked stays item-local: that item is marked blocked and the rest continue, while any other error stops the batch. You need the released postplus CLI and credits on the account — the skill checks price and balance first and stops on insufficient credits instead of suggesting a top-up.