puntorigen/avatar-skills
GitHubAvatar creation & identity
Invent or clone a presenter and vary its look — hero stills, camera angles, wardrobe/location variants, and clean reference frames.
Talking-head & reels
Turn a script + an avatar into finished vertical reels — lip-synced talking-head scenes, style transfer, story reels, karaoke captions, and ambient sound design.
B-roll
Generate or find complementary B-roll — hyper-real clips, avatar action shots, real YouTube footage, website/terminal captures.
Voice & audio
Clone voices, isolate clean samples, and generate music and sound effects for reels.
Video editing & analysis
Compose reels, analyze scene structure, replace backgrounds, and transcribe video.
Discovery & downloaders
Find top-performing public reels by topic or brand, and batch-download videos from YouTube, TikTok, Instagram and Facebook.
Strategy & scripting
Plan a content/channel strategy, write viral short-form scripts, plan the cuts/montage (Rule of Six), generate company-aware onboarding reel-script series ready to feed the reel pipeline, and apply luxury/HNWI sales psychology when scripting premium content or high-ticket web copy.
Social posts & carousels
Turn a topic + a brand into a finished Instagram carousel — research the winning carousel structure of competitors (Apify), decompose it slide-by-slide with Gemini vision, then generate on-brand 4:5 slides + caption + hashtags with the asset-generator engine.
Documents & PDFs
Deconstruct a premium reference PDF into a reusable design + narrative mold (palette, type roles, per-page archetypes, arc) with Gemini vision, then regenerate a brand-new on-topic multi-page PDF — crisp HTML text via Chromium, imagery via the asset-generator engine.
Generation engines (shared)
Shared model backends the pipeline skills call under the hood — image generation (GPT Image 2, Gemini asset-generator), audio-synced video (Seedance 2.0 via Higgsfield MCP), multi-character audio (Gemini TTS + SFX + score), and instrumental background music.