Multi-Platform Publish

Apply ../00-video-production-guardrails/SKILL.md first. This skill is a distribution workflow, not a generation workflow — it does not produce a video (see 01-blog-to-video, 02-url-to-explainer-video for that); it takes an already-finished asset from ../../library/ and moves it to a real platform. The guardrails skill's platform-and-jurisdiction disclosure gate is a hard precondition this skill consumes at the Policy Gate below, never re-implements.

Architecture

CONFIG -> USER ELICITATION -> TARGET SELECTION -> LEDGER LOOKUP (idempotency)
  -> CAPABILITY GATE (target x provider x rendition) -> POLICY GATE
  (content-properties x target x jurisdiction) -> DRY RUN
  -> USER AUTHORIZATION -> PUBLISH -> ASSET LEDGER

Four primitives, kept distinct throughout this skill:

• Distribution target (../../targets/<name>/) — what the platform itself requires of the asset (the rendition profile).
• Provider/integration (../../providers/<name>/) — how the target is actually reached. One provider can serve multiple targets, and coverage differs per target even for the same provider (see Provider Resolution below).
• Rendition profile — the target's own requirement (aspect ratio, duration, resolution, codec, container, file size, captions, thumbnail, audio), never assumes or hardcodes which generation provider produced the asset that must meet it.
• Disclosure policy — content properties × target × jurisdiction, owned entirely by the guardrails skill; this skill queries it, it does not carry its own copy of platform disclosure rules.