Updated 3w ago
by mojomast
open-higgsfield-skills
Agent-agnostic skills that turn one prompt into a video with open-weights ComfyUI workflows (LTX, FLUX Klein, Qwen, Wan) - video-director orchestrator + 35 skills. #OpenSourceHiggsfield
53 GitHub stars (53)
+53 in 23 days
| Period | New stars |
|---|---|
| Nov 18 | +14 |
| Nov 19 | +5 |
| Nov 20 | +2 |
| Nov 21 | +9 |
| Nov 22 | +12 |
- 30 skills (30 skills found in its SKILL.md files)
- 7 forks (7 forks)
- 3w ago last update (Last updated 3w ago)
Install
Skills CLI
npx skills add mojomast/open-higgsfield-skills
Asks which skills and agents to install. Just one: npx skills add mojomast/open-higgsfield-skills --skill audio-driven-motion — add -g for every project.
Claude Code
git clone https://github.com/mojomast/open-higgsfield-skillscp -r "open-higgsfield-skills/skills/audio-driven-motion" ~/.claude/skills/
Copies the audio-driven-motion skill into ~/.claude/skills/ for every project (use .claude/skills/ for one project). Repeat for the other skills listed below.
Plugin
/plugin marketplace add mojomast/open-higgsfield-skills
Run inside Claude Code if the repository ships a plugin marketplace, then pick plugins from /plugin.
Skills can run code on your machine — read the SKILL.md first.
What it does
In the author's own words, from each skill's SKILL.md — showing the first 30 of 37 files.
-
audio-driven-motion
Make video whose motion is driven by an audio track (a talking or singing avatar with body motion, or music-reactive motion) using open-weights LTX-2 audio-to-video, LTX-2.3 image+audio-to-video, Wan 2.2 S2V, InfiniteTalk or HuMo in ComfyUI. Use for presenters, singers and audio-synced performance beyond mouth-only lip-sync.
skills/audio-driven-motion/ on GitHub (opens in a new tab)SKILL.md -
background-matte
Remove or replace backgrounds and pull mattes with open-weights BiRefNet in ComfyUI (stills to transparent or solid colour, video background removal), plus PlateLock plate compositing (difference matte, temporal stabilise, fidelity report) to put a generated element back onto a real plate. Use for product shots, compositing and green-screen style work.
skills/background-matte/ on GitHub (opens in a new tab)SKILL.md -
camera-control
Animate a still with an explicit, repeatable camera move (named trajectories such as orbit, dolly or crane, or drawn ATI point tracks) using open-weights Wan Fun Camera and Wan ATI in ComfyUI. Use when the camera path must be exact rather than prompted.
skills/camera-control/ on GitHub (opens in a new tab)SKILL.md -
camera-lens-lut
Apply film-style camera moves (handheld, push, drift), lens effects, a 3D .cube grade LUT and letterboxing to a clip in ComfyUI with the bundled Backlot node pack. Use as a finishing pass on generated shots.
skills/camera-lens-lut/ on GitHub (opens in a new tab)SKILL.md -
character-swap-h3
Swap a person, face or object in an existing clip for a reference image using open-weights MiniMax H3 (local weights) with D3voz's one-click swap nodes, or an open LTX-2.3 motion-swap fallback. Use for recasting, outfit/product swaps and face fixes.
skills/character-swap-h3/ on GitHub (opens in a new tab)SKILL.md -
comfyui-connect
Submit ComfyUI API-format workflows to any ComfyUI server (local, a Google Colab tunnel, or a hosted instance), poll for completion, download outputs, upload inputs, and cancel jobs. Use whenever another skill needs to run a ComfyUI workflow.
skills/comfyui-connect/ on GitHub (opens in a new tab)SKILL.md -
comfyui-preflight
Check, before submitting, that a ComfyUI server has every node type and model weight file a workflow needs. Reads /object_info only; submits nothing. Use before the first run of any workflow on a new or changed server (local, Google Colab, or hosted).
skills/comfyui-preflight/ on GitHub (opens in a new tab)SKILL.md -
continuity-check
Sweep adjacent shots and takes for continuity breaks (cast, wardrobe, props, grade, lighting, screen direction, timing) at plan level and take level, and report findings with proposed fixes - never edits. Use before locking a cut, after a take is replaced, or when asked "do these cut together?".
skills/continuity-check/ on GitHub (opens in a new tab)SKILL.md -
control-to-video
Generate video that follows a control signal (depth, OpenPose skeleton or canny edges extracted from a reference or previs clip) with open-weights LTX-2 and Wan 2.2 Fun Control / Wan VACE in ComfyUI. Use when motion, blocking or layout must match a source clip exactly.
skills/control-to-video/ on GitHub (opens in a new tab)SKILL.md -
effect-recipe-read
Reverse-engineer a video effect into a reusable JSON recipe: a local Qwen3-VL 8B reads START/MIDDLE/END frames of a reference clip and writes the end-frame and motion prompts, then an open LTX-2 + Qwen edit graph applies the recipe to the user's photo (forkfield). Use for "make this effect with my photo" from a reference clip.
skills/effect-recipe-read/ on GitHub (opens in a new tab)SKILL.md -
film-assembly
Cut the final film with ffmpeg - normalise clips to one size/fps, join with cuts or crossfades (optionally snapped to music beats), lay narration over a ducked music bed, and export for 16:9, 9:16, 1:1 or 4:5. Use when all shots and audio are approved.
skills/film-assembly/ on GitHub (opens in a new tab)SKILL.md -
flux3-klein-stills
Generate stills and keyframes with open-weights FLUX Klein (distilled FLUX 3 Klein, FLUX 3 Klein base with prompt upsampling, or FLUX.2 Klein 4B/9B) through ComfyUI. Use for shot keyframes, character references and any text-to-image still.
skills/flux3-klein-stills/ on GitHub (opens in a new tab)SKILL.md -
frame-interpolation
Raise a clip's frame rate (e.g. 24 to 48/60 fps, or smooth slow motion) with open-weights FILM frame interpolation in ComfyUI. Use for slow-mo beats or to match a delivery frame rate.
skills/frame-interpolation/ on GitHub (opens in a new tab)SKILL.md -
google-colab-runner
Run video-director productions on Google Colab. It packages a manifest into a self-contained Google Colab bundle (a generated .ipynb that installs ComfyUI and custom nodes at pinned commits, downloads the open-weights models with the user's own Hugging Face token from Google Colab secrets, renders every shot, detects the GPU tier and collects outputs to Drive or a zip), or it lets the director dr…
skills/google-colab-runner/ on GitHub (opens in a new tab)SKILL.md -
image-edit
Edit a still with open-weights Qwen-Image-Edit in ComfyUI - instruction edits, new camera angles, push-in, coverage close-ups, relighting, and outpainting to a new aspect. Use for chat edits on keyframes and to build coverage from one master shot.
skills/image-edit/ on GitHub (opens in a new tab)SKILL.md -
image-to-3d
Turn a still (or 4 views) into a GLB mesh with open-weights TRELLIS.2 or Pixal3D in ComfyUI. Use for previs props, set pieces and turnaround references.
skills/image-to-3d/ on GitHub (opens in a new tab)SKILL.md -
lip-sync
Make a character speak or sing to an audio track with open weights in ComfyUI - talking shot from a photo + audio (LTX-2.3), re-dub an existing video (LTX-2.3 IC-LoRA lipdub), or low-VRAM singing/talking (Wan 2.1 + InfiniteTalk). Use for dialogue, presenters and music-video performance.
skills/lip-sync/ on GitHub (opens in a new tab)SKILL.md -
lora-training
Train a LoRA (character, product, style or motion) for an open-weights image or video model with ostris/ai-toolkit, behind a hard human approval gate on the dataset. Use for "train a LoRA of my product", "make a style LoRA", "teach the model this character". Nothing trains until the user approves the dataset.
skills/lora-training/ on GitHub (opens in a new tab)SKILL.md -
ltx-iclora-utils
LTX-2.3/2.5 IC-LoRA and LoRA utility graphs: motion transfer from a reference clip, motion-guided i2v, LTX-2.5 effects, plus official templates for subtitle and watermark removal, archival restoration, style transitions, generic IC-LoRA and ID-LoRA (identity). Use for clip-in, clip-out repair and effects on LTX.
skills/ltx-iclora-utils/ on GitHub (opens in a new tab)SKILL.md -
ltx25-image-to-video
Animate a still into a shot with open-weights LTX-2.5 in ComfyUI (image-to-video with optional last frame, or text-to-video), including LTX's native audio latent. Use after a keyframe passes still-qa.
skills/ltx25-image-to-video/ on GitHub (opens in a new tab)SKILL.md -
multi-gpu-dispatch
Spread a production across several GPUs or ComfyUI servers: round-robin shots across server URLs with a stdlib dispatcher, or install ComfyUI-Distributed (parallel workers and tiles) or raylight (split one large model across GPUs with FSDP/xDiT). Use when one GPU is too slow or too small for the plan.
skills/multi-gpu-dispatch/ on GitHub (opens in a new tab)SKILL.md -
music-sfx
Generate music beds, songs and sound effects with open-weights ACE-Step 1.5, Stable Audio Open / Stable Audio 3 and MiniMax Music 3 in ComfyUI, using model-specific prompt rules. Use for score, stingers, ambience and foley.
skills/music-sfx/ on GitHub (opens in a new tab)SKILL.md -
one-take-vfx
Add a stylised VFX effect to an existing clip in one pass (set-on-fire, disintegrate, time-freeze, hologram, portal, rain-stop, levitate, aurora trails, liquid chrome, neon wireframe, particle bloom, slow-mo debris) with open-weights LTX IC-LoRA union control and Kainotomic's preset node. Use for hero moments and title beats.
skills/one-take-vfx/ on GitHub (opens in a new tab)SKILL.md -
previs-control-passes
Block a shot in 3D and export deterministic control passes (OpenPose skeleton MP4, depth MP4, subject alpha, 4x4 camera matrices, 24 fps previs) from ComfyUI Motion Previs Studio, then feed them to control-to-video or camera-control. Also applies its verified template-binding format to ComfyUI API graphs. Use when blocking and camera must be designed before generation.
skills/previs-control-passes/ on GitHub (opens in a new tab)SKILL.md -
reframe-aspect
Reframe a finished clip to another aspect ratio (16:9 to 9:16, 1:1, 4:5) by outpainting video with an open-weights LTX-2.3 IC-LoRA in ComfyUI, instead of cropping. Use for multi-platform deliveries.
skills/reframe-aspect/ on GitHub (opens in a new tab)SKILL.md -
relighting
Relight a still (change light direction, colour, time of day, or match a plate) with open-weights models: Qwen-Image-Edit 2509 relight, FLUX.2 Klein + normal-map lights (motionfield), or the Comfy-Org portrait relighting template. Stills only; animate afterwards with image-to-video. Use before video when lighting must match across shots.
skills/relighting/ on GitHub (opens in a new tab)SKILL.md -
restyle
Restyle or re-motion an existing clip with open-weights LTX (IC-LoRA depth/union control restyle, or LTX-2.5 motion guide from a reference video). Use when the user wants a clip redone in a new look while keeping its motion.
skills/restyle/ on GitHub (opens in a new tab)SKILL.md -
shot-planning
Staged preproduction from a sentence or script to a locked shot list - beats, shots with camera/duration/narration, per-shot prompts, cost and time estimate - with human approval gates and no auto-spend. Use before generating any film, ad or episode.
skills/shot-planning/ on GitHub (opens in a new tab)SKILL.md -
shot-qc
Model-free quality control for rendered shots and final cuts using ffprobe/ffmpeg measurements: duration, aspect, audio present, frozen frames, black frames, hard-cut count, flicker, soft (blurry) frames and clipped highlights/shadows, with a 0-100 score. Use after every video render and before delivery; numbers override VLM descriptions.
skills/shot-qc/ on GitHub (opens in a new tab)SKILL.md -
shot-transitions
Generate a bridging transition clip between two shots (first-frame/last-frame video) with open-weights LTX-2.5 or LTX-2.3 transition LoRA via Seam's graphs, with an optional VLM that describes both frames to write the transition prompt. Use for match cuts, morphs and seamless scene changes.
skills/shot-transitions/ on GitHub (opens in a new tab)SKILL.md
No skill matches that filter.
Review before you install
A skill is a set of instructions and scripts your agent will follow, and it can run code on your machine. Read the SKILL.md and any scripts first — this listing is not a security review.