Name: slideshow-producer
Author: postplusai

SKILL.md

Slideshow Producer

The user wants TikTok slideshow or Instagram carousel production from vibe, local images, slide prompts, reviewable manifest JSON, generated images, and final text overlays.
Use after script/hook strategy is clear enough to plan slides. This skill owns slide manifest, preview/review, generation orchestration, and compositing.

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.

Do not write hook variants from scratch, bypass the shared image command, create persona packs, or replace final creative QA.
Route slides by reference state: image-gpt-image-2-text for slides without references, image-gpt-image-2-edit for slides with references.
Slideshow aspect is a creative choice passed as --aspect-ratio (TikTok 9:16; Instagram carousel 4:5 unless the user asks otherwise). The other image defaults (quality, resolution, output format) are owned by the image create verb — discover them via postplus media schema --json; do not restate hosted enums here.

The slide manifest is the routing source of truth.
imageSource: "local" uses localImagePath as the final image and must not call generation.
imageSource: "generate" with no reference paths/URLs must use image-gpt-image-2-text.
imageSource: "generate" with any reference path/URL must use image-gpt-image-2-edit.
Never send reference-backed slides through text-to-image. Fix the manifest before generation if routing is inconsistent.

Ask for platform, slideshow count, vibe, source basis, output folder, and any local images. Classify local images as final slide images or generation references.
Prompts should be core scene plus vibe in two or three sentences. Add concrete product, environment, emotion, and prop details when they matter; avoid long negative-constraint walls.
Save a local manifest before generation. For three or more slideshows, split manifest drafting across parallel sub-agents, then the main agent reviews the collected manifests.

Before generation, ask for explicit approval: number of slides, model, quality, resolution, and counts of text-to-image versus reference edits.
After compositing, hand off final image paths, updated manifest, and any creative-qa or publishing.

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 still fails after any bounded recovery allowed by the executing PostPlus skill, report the exact error and stop. Do not bypass the failure with metadata-only answers, readiness probing, local payload rewrites, alternate execution paths, or unpublished tools.

Choose the smallest matching command or workflow from the user input and run it directly.
Readiness diagnostics: postplus doctor --skill slideshow-producer.
If an owned CLI or script command still fails after any bounded recovery allowed by the executing PostPlus skill, report the exact error and stop. Do not bypass the failure with metadata-only answers, readiness probing, local payload rewrites, alternate execution paths, or unpublished tools.
Use postplus media schema --json only when constructing or repairing an unknown request shape.
Generate each slide through the shared image create verb (owned by image-batch-runner), attributing the run to this skill: postplus media create image-gpt-image-2-text --skill slideshow-producer --prompt '<scene + vibe>' --aspect-ratio <ratio> --output <result.json>. For reference-backed slides use image-gpt-image-2-edit with one or more --reference-image <url>. Do not use another execution interface.
If the CLI returns a quote-confirmation challenge, run postplus quote confirm --json --challenge-file <challenge.json> and retry with the returned token.