trustfuture/simon-skills

不露脸做视频账号的 AI skill 合集：调查长片 + 手绘白板讲解视频，一句选题出片、配音字幕、封面与各平台文案

What it solves

This project provides a collection of AI skills designed for creating "faceless" video content. It automates the entire production pipeline—from research, scriptwriting, voiceover, visual generation, and cover art to platform-specific captions—allowing an AI agent to produce a ready-to-publish video from a single topic prompt.

How it works

The project is structured as a set of skills that can be integrated into AI agents like Claude Code or Codex. These agents use the following tools and technologies to generate content:

  • Investigation-video: Creates long-form commercial or consumer investigation videos (10+ minutes) using a chain of facts, Remotion for synthesis, and multi-platform copywriting.
  • Whiteboard-video: Produces "draw-and-talk" style explanation videos (around 2 minutes) featuring Excalidraw-style animations, AI-generated stickers, and burned-in subtitles.
  • Video-common: A shared utility skill that handles fact-checking, cover art acceptance, compliance checks, and platform-specific copywriting.

Technical requirements include Node.js, FFmpeg, Python, Playwright, and Volcengine TTS for voice synthesis.

Who it’s for

Content creators who want to actually automate the the production of high-quality, long-form or explanation videos without showing their face on camera.

Highlights

  • End-to-end automation: Handles everything from research to the final render.
  • Real-world proven: Each skill is based on a real, active account's workflow.
  • Agent-compatible: Specifically formatted for Claude Code and Codex agents.
  • Cusomizable: Supports easy branding changes via configuration files.

Related

  • Project
    google-deepmind/videoprism

    VideoPrism is a Google‑DeepMind open‑source video foundation model (JAX/Flax) that provides pre‑trained video‑only and video‑text encoders. Trained on billions of image‑text and video‑text pairs, the frozen backbones achieve state‑of‑the‑art results on most video‑understanding benchmarks. The repo offers easy installation, checkpoint loading utilities, and three Colab demos (video encoding, cross‑modal retrieval, and classification fine‑tuning). Four model sizes (base/large, encoder‑only or encoder‑plus‑text) are hosted on Hugging Face, licensed under Apache 2.0/CC‑BY.

  • Project
    ModelTC/Minimax-H3-Turbo

    A collection of distilled LoRA checkpoints for MiniMax-H3 that enables fast, few-step video and audio generation via Diffusers and ComfyUI.

  • Project
    okineadev/vitepress-plugin-llms

    A VitePress plugin that automatically creates LLM‑friendly Markdown files (`llms.txt` and `llms‑full.txt`) and optional UI buttons for copying/downloading pages as Markdown, with support for custom labels, multilingual configs, and special `<llm-only>` / `<llm-exclude>` tags.

  • Project
    microsoft/mlvc

    MLVC is a neural video codec that runs in real‑time on consumer NPUs, beating hardware HEVC by >70 % BD‑rate. The repo provides pretrained models, a full training pipeline, conversion tools for CoreML/ONNX/QNN, a C++ entropy coder, and benchmarking utilities, with a companion C++ library for production deployment.