openquok-tiktok-slideshow

Overview

Research channel → lock character → generate frames → text overlay → post via openquok.

It requires openquok-core and the global openquok CLI. Uploads and create go through core recipes (media Rule 2, TikTok photo carousel, preferred private draft).

Install (Copy required)

Scripts must land as real files:

npx skills add https://github.com/Ratimon/openquok-monorepo/tree/main/agent --skill openquok-tiktok-slideshow --copy -y

Interactive installs: Project scope + Copy. Hermes: copy the full skill folder (a URL install of SKILL.md alone omits scripts/ and references/).

Prerequisites

openquok --version
openquok auth:status

Install canvas (overlays)

From the skill install directory (or any Node project that can require(‘canvas’) for the overlay script):

npm install canvas

Pipeline

node scripts/onboarding.js --init --dir tiktok-marketing/
node scripts/onboarding.js --validate --config tiktok-marketing/config.json

Config shape

Workspace data lives outside the skill folder (for example tiktok-marketing/):

{
  "channel": {
    "name": "",
    "handle": "",
    "niche": "",
    "audience": "",
    "painPoint": "",
    "platforms": ["tiktok"]
  },
  "character": {
    "profilePath": "tiktok-marketing/character-profile.json",
    "referenceImages": {
      "faceLock": "tiktok-marketing/refs/face-lock.png",
      "bodyLock": "tiktok-marketing/refs/body-lock.png"
    }
  },
  "imageGen": {
    "provider": "openai",
    "apiKey": "",
    "model": "gpt-image-1.5"
  },
  "openquok": {
    "integrationId": "",
    "status": "scheduled",
    "privacyLevel": "SELF_ONLY",
    "contentPostingMethod": "DIRECT_POST",
    "scheduledAt": null,
    "title": ""
  },
  "posting": { "schedule": ["07:30", "16:30", "21:00"] },
  "research": "tiktok-marketing/channel-research.json",
  "strategy": "tiktok-marketing/strategy.json"
}

Core scripts

Generate slideshow images

node scripts/generate-slides.js 
  --config tiktok-marketing/config.json 
  --output tiktok-marketing/posts/YYYY-MM-DD-HHmm/ 
  --prompts prompts.json

When character.profilePath is set, the script prepends the LOCKED block and merges VARIATIONS. OpenAI attaches face/body lock images when those files exist. Portrait 1024x1536; set exec timeout ≥ 600s for six gpt-image-1.5 images.

Add text overlays

node scripts/add-text-overlay.js 
  --input tiktok-marketing/posts/YYYY-MM-DD-HHmm/ 
  --texts texts.json

Requires canvas. Six-slide formula: Hook → Problem → Discovery → Transform ×2 → CTA. No emoji in overlay text.

Post via openquok

node scripts/post-via-openquok.js 
  --config tiktok-marketing/config.json 
  --dir tiktok-marketing/posts/YYYY-MM-DD-HHmm/ 
  --caption "…" 
  --title "Short title"

Under the hood: openquok upload each slide → openquok posts:create --json TikTok photo carousel. Default privacy is SELF_ONLY.

Optional analytics after publish (via core, not this skill):

openquok analytics:platform <integration-uuid> --days 7

Quick reference

Common mistakes

Related