convez.ai
SkillAutomaçãoIntermediário
bossjones

Boss Skills - Suite de Automação Social Media

bossjones/boss-skills

Baixe mídia do Twitter/X e converta para Instagram Reels automaticamente

Ver no GitHub
InicialPluginScore · de 0 a 100

Coleção de skills Claude Code para automação de mídia social e criação de conteúdo. Inclui ferramentas para download de mídia do Twitter/X e conversão automática de tweets em reels verticais para Instagram. Construído com Python, FFmpeg e quality gates via plugin-eval.

Instalação

$npx convez-ai bossjones/boss-skills

1 comando no terminal: detecta o formato do repositório (marketplace ou plugin avulso) e instala via Claude Code. Requer Node 18+ e o Claude Code instalado.

Métricas

Estrelas
1
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 6 dias
Versão
—
Issues abertas
20
Contribuidores
1

Casos de uso

  • Download em massa de mídia (imagens/vídeos) do Twitter/X
  • Conversão automática de tweets em formato vertical 9:16 para Instagram
  • Integração com conteúdo protegido via autenticação por cookies
  • Automação de fluxos de criação de conteúdo cross-plataforma

Veredito

Suite sólida e bem estruturada para automação de redes sociais. Code quality garantido com linting, testes e avaliação automática; deploy é via marketplace Claude Code. Recomendado para quem precisa escalar processos manuais de coleta e conversão de conteúdo.

README

A tradução para o português deste README ainda está na fila.

boss-skills

Personal Claude Code plugin marketplace for social media automation and content creation tools.

Installation

Add this marketplace to Claude Code:

/plugin marketplace add bossjones/boss-skills

Then install individual plugins:

/plugin install twitter-tools@boss-skills

Available Plugins

Full per-plugin documentation — components, install commands, and usage examples — lives in docs/plugins/.

social-media/twitter-tools

Twitter/X social media tools for downloading media and converting tweets to Instagram Reels format.

/plugin install twitter-tools@boss-skills

Skills included:

SkillDescription
twitter-media-downloaderDownload images and videos from X/Twitter using gallery-dl
twitter-to-reelConvert tweets to Instagram Reels format (9:16 vertical video)

Features:

  • Download media from tweets, user profiles, timelines, likes, and bookmarks
  • Support for protected content via browser cookie extraction
  • Automatic video download and reel composition in a single command
  • Theme auto-detection (light/dark) for seamless background matching
  • JSON output mode for programmatic integration

See plugins/social-media/twitter-tools/README.md for details.

Quick Start Examples

Download Twitter Media

# Download all media from a user
uv run python scripts/download.py "https://x.com/NASA" --output ./downloads

# Download only videos from a tweet
uv run python scripts/download.py "https://x.com/user/status/123" --videos-only

# Download with authentication
uv run python scripts/download.py "https://x.com/i/bookmarks" --browser firefox

Create Instagram Reels from Tweets

# Auto-download video and create reel (recommended)
uv run python scripts/create_reel.py "https://x.com/user/status/123" -o reel.mp4

# With authentication for protected tweets
uv run python scripts/create_reel.py "https://x.com/user/status/123" --browser firefox -o reel.mp4

# With explicit video file
uv run python scripts/create_reel.py "https://x.com/user/status/123" video.mp4 -o reel.mp4

Development

Prerequisites

  • Python 3.11+
  • uv package manager
  • FFmpeg (for video processing)

Commands

# Install dependencies
make install

# Run linting (ruff + basedpyright)
make lint

# Run tests
make test

# Run all checks
make

Skill Quality

Skills are quality-gated with plugin-eval, pulled on demand from upstream via uvx (nothing is vendored or submoduled, so you always get the latest version).

# Report static scores for every skill (never fails)
make eval

# Quality gate used by CI: fail if any skill < EVAL_THRESHOLD
make eval-ci

# Override the floor ad hoc
make eval-ci EVAL_THRESHOLD=70

# Deep-dive one skill at standard depth (LLM judge; uses Claude Code Max)
make eval-skill SKILL=plugins/social-media/twitter-tools/skills/twitter-to-reel

make eval-ci runs in CI at static (quick) depth — deterministic, no API key required. EVAL_THRESHOLD (in the Makefile) is a regression floor set to min(baseline scores) - 5; bump it as skills improve.

For interactive use, install the Claude Code plugin from the already-registered marketplace (these are user-typed slash commands, not repo scripts):

/plugin install plugin-eval@claude-code-workflows
/plugin marketplace update claude-code-workflows   # pull the latest on demand

Enable autoUpdate on the claude-code-workflows marketplace if you want plugin updates without the manual marketplace update. This gives you /eval, /certify, and /compare while iterating on skills.

Escape hatch: if upstream plugin-eval changes ever make the gate flaky, pin a known-good revision without editing code:

PLUGIN_EVAL_SOURCE='git+https://github.com/wshobson/agents.git@<sha>#subdirectory=plugins/plugin-eval' make eval-ci

Project Structure

boss-skills/
├── plugins/
│   └── social-media/
│       └── twitter-tools/
│           ├── .claude-plugin/
│           │   └── plugin.json
│           ├── skills/
│           │   ├── twitter-media-downloader/
│           │   │   ├── SKILL.md
│           │   │   └── scripts/
│           │   └── twitter-to-reel/
│           │       ├── SKILL.md
│           │       └── scripts/
│           └── README.md
├── devtools/
├── scripts/
└── tests/

Dependencies

The twitter-tools plugin requires:

  • gallery-dl - Media download engine
  • yt-dlp - Video download support (optional but recommended)
  • playwright - Browser automation for screenshots
  • pillow - Image processing
  • FFmpeg - Video composition

Install with:

pip install gallery-dl yt-dlp playwright pillow numpy
playwright install chromium
brew install ffmpeg  # macOS

License

MIT

Ver o original no GitHub

Bom para

Relacionados