Higgsfield Skills — Setup Guide¶

Source: skills.sh (577.3K combined installs) GitHub: higgsfield-ai/skills Category: AI Video & Image Generation First Seen: August 13, 2026 sweep Quality Tier: 🟡 Beta (platform-dependent)

Higgsfield is an AI video/image generation platform, and this is its official agent skill pack. The cluster covers the full commercial content surface: product photography, avatar video (soul-id), e-commerce marketplace cards, websites, video explainers, game generation, YouTube thumbnails, and brand kits. Agents orchestrate generation through Higgsfield's platform rather than rendering locally.

Installation¶

npx skills add higgsfield-ai/skills

Core Skills¶

Skill	Installs	Use For
higgsfield-generate	125.7K	Core text-to-video/image generation orchestration
higgsfield-product-photoshoot	107.0K	Studio-quality product images without a studio
higgsfield-soul-id	105.9K	Consistent avatar/character video generation
higgsfield-marketplace-cards	104.6K	Amazon/Etsy-style listing creative
higgsfield-websites	46.4K	Website visuals and hero sections
higgsfield-video-explainer	34.9K	Explainer videos from a topic or script
higgsfield-game-generation	26.6K	Game asset and scene generation
higgsfield-youtube-thumbnail	13.1K	Thumbnail design for video content
higgsfield-brandkit	13.1K	Consistent brand visual kits
higgsfield-soul	2	Avatar engine core (new)

Prerequisites¶

Higgsfield account and API access (the skills call the Higgsfield platform)
Budget for generation credits — pay-per-generation, not free

CorpusIQ Use Cases¶

UGC video pipeline augmentation — soul-id + video-explainer complement the existing HyperFrames/HeyGen rotation for avatar-driven shorts
Product/marketing visuals — product-photoshoot and marketplace-cards for CorpusIQ connector and feature launch creative
Brand consistency — brandkit enforces visual identity across agent-generated ad sets

Limitations / Verification¶

Platform-dependent: generation quality and availability ride on Higgsfield's service; no local fallback
Verify by generating one test asset through higgsfield-generate and confirming it renders in your pipeline