video-shotcraft
Vincentwei1021/video-shotcraftRemotion-based skill for cinematic product promo videos, with 124 shot recipe cards, a full template, beat sync, sound design and a motion workbench.
A skill that turns one topic, a talking-head video or a photo into a finished paper-collage explainer video, using the Atlas Cloud API and local ffmpeg.
A complete, scripted pipeline, but it depends on a single paid provider and has no automated tests.
Vox Director automates a video in the Vox-style paper-collage look: torn edges, tape, halftone dots, newspaper clippings and bold flat colors. For a topic, the agent picks a narrative arc and writes a beat map in beats.json, then pauses for your approval. It renders the same beat in three or four themes so you can choose the look, generates a collage keyframe per beat, animates each one, adds a narrator voice and background music, and assembles the final mp4 with ffmpeg, ducking music under the voice and burning in captions and a watermark. Two later gates let you approve the beat map and pick the style by eye.
Two more input modes reuse the engine: A-roll, where an existing talking-head video is split into beats and re-styled while keeping the face and lip-sync, and C-roll, where one still photo is cut out as a sticker and placed in each poster. The skill is about 22 KB with 10 references, 22 scripts and 14 templates. All generation goes through Atlas Cloud, so you need an ATLASCLOUD_API_KEY and pay per call. The only network host in the scripts is Atlas Cloud. There are no tests.
A 22 KB skill with 10 references, 22 scripts and 14 templates, no tests or evals.
Counted from the repository file tree. How we score depth.
A video-generation skill that scripts a Vox-style explainer end to end using the paid Atlas Cloud API plus local ffmpeg. Your prompts and any reference images or footage you provide are uploaded to Atlas Cloud, authenticated with your ATLASCLOUD_API_KEY, and generation calls cost money. Local scripts call ffmpeg through argument-array subprocesses, and the shipped .skill archive matches the repo's plain files.
The repository (52 files) on Oct 29, 2026. Not a full review: we read 2 of 23 code files; it relies on code outside the repository: ffmpeg / ffprobe and the numpy and Pillow packages; Atlas Cloud hosted generation, TTS, ASR and chat models; the other 21 scripts were pattern-scanned (subprocess, network, file-delete, environment use) rather than read line by line; only scripts/atlas_cloud.py and package.json were read in full.
NVIDIA SkillSpector static scan: 80 pattern hits (Dangerous Code Execution 40, Supply Chain 22, Data Exfiltration 11, Data Flow 6, MCP Least Privilege 1). The two CRITICAL Data Flow hits are the API client sending the key it read from the environment to its own endpoint. The 20 'Concealed Executable Artifact' hits come from vox-director.skill, a zip whose 38 entries we compared with the repo and found identical to the plain files. Remaining hits are ffmpeg subprocess calls.
A review, not a certification. Skills change after we check them — read the files before installing.
MIT — commercial use allowed; check Atlas Cloud's terms for generated media.
Not really. The workflow has two human approval gates, and generation costs money per call.
Remotion-based skill for cinematic product promo videos, with 124 shot recipe cards, a full template, beat sync, sound design and a motion workbench.
Nine skills for generating images, video, brand kits, product shots and websites through the paid Higgsfield service and its CLI.
HeyGen's framework and skills for writing HTML and rendering it to deterministic MP4 video, with a CLI, a block registry and media tooling.
It is an agent skill that makes a narrated paper-collage video from a topic, a talking-head clip or a photo. It writes the script, generates collage frames and motion, adds voice and music, and assembles the video with ffmpeg.
Add the repo to your agent's skills folder, install ffmpeg, and set ATLASCLOUD_API_KEY from your Atlas Cloud console. Then give the agent a one-line topic.
Yes. Image, video, voice and music generation run through the paid Atlas Cloud API, so each project uses credits. The skill itself is MIT licensed.