convez.ai
PluginBusca & PesquisaIntermediário
zhoux77899

Claude Code Insights

zhoux77899/claude-code-insights

Ranking diário automatizado de plugins Claude Code com UI interativa

Ver no GitHub
InicialPluginScore · de 0 a 100

Sistema automatizado que coleta, classifica e visualiza repositórios do ecossistema Claude Code diariamente. Fornece dados estruturados em JSON e interface web React para acompanhar plugins mais populares e ativos, com histórico de métricas ao longo do tempo.

Instalação

$npx convez-ai zhoux77899/claude-code-insights

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
2
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 1 semana
Versão
—
Issues abertas
0
Contribuidores
2

Casos de uso

  • Descobrir plugins Claude Code mais populares e em ascensão
  • Acompanhar tendências e adoção no ecossistema de plugins
  • Integrar dados estruturados de repos em ferramentas externas

Veredito

Projeto robusto de inteligência de dados para o ecossistema Claude. Pipeline automatizado bem estruturado, mas exige múltiplas tecnologias (Python, Node, bash, GitHub Actions) — mais adequado para quem quer explorar tendências ou integrar dados em sistemas maiores.

README

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

Claude Code Insights

Automated data collection, ranking, and visualization for the Claude Code plugin ecosystem.

What This Project Provides

  • Daily indexed repository data for Claude Code plugins.
  • Structured JSON artifacts for downstream usage.
  • A generated Markdown insight report (Top 30 + recently active repos).
  • A React-based web UI deployed to GitHub Pages.

Repository Structure

claude-code-insights/
├── .github/workflows/
│   ├── plugins-daily-insight.yml   # Daily data generation pipeline
│   └── github-pages-deploy.yml     # Build and deploy UI to GitHub Pages
├── plugins/
│   ├── README.md
│   ├── repos.json                  # Latest ranked repository dataset
│   ├── history.json                # Time-series metrics history
│   ├── cached-repos.txt            # Cached repository list
│   └── plugins-daily-insight.md    # Generated Markdown report
├── scripts/
│   ├── fetch_repos.sh              # Search GitHub code API for plugin repos
│   ├── dedupe_repos.py             # Deduplicate search results
│   ├── sort_repos.py               # Fetch repo metadata and sort by stars
│   └── chronicle_repos.py          # Maintain rolling history snapshots
├── ui/                             # Vite + React + TypeScript frontend
├── requirements.txt                # Python dependencies
└── README.md

Data Generation Pipeline

The workflow .github/workflows/plugins-daily-insight.yml runs on:

  • schedule: daily at 00:00 UTC
  • manual trigger: workflow_dispatch

Pipeline steps:

  1. scripts/fetch_repos.sh queries GitHub Search API (filename:marketplace.json path:.claude-plugin).
  2. scripts/dedupe_repos.py merges paged search responses and removes duplicates.
  3. scripts/sort_repos.py fetches repository metadata and sorts by stargazers_count.
  4. scripts/chronicle_repos.py updates rolling historical metrics in plugins/history.json.
  5. Workflow generates/updates:
    • plugins/repos.json
    • plugins/history.json
    • plugins/cached-repos.txt
    • plugins/plugins-daily-insight.md

Detailed plugin data schema and usage examples are documented in plugins/README.md.

Web UI and Deployment

  • Frontend source lives in ui/.
  • GitHub Pages deployment is managed by .github/workflows/github-pages-deploy.yml.
  • During deployment, data files are copied into ui/data/:
    • plugins/repos.json -> ui/data/plugins-repos.json
    • plugins/history.json -> ui/data/plugins-history.json
  • Vite base is set to /claude-code-insights/ for Pages routing.

UI-specific setup and architecture are documented in ui/README.md.

Quick Start

Prerequisites

  • Python 3.13 (matches workflow runtime)
  • Node.js 20 and npm (matches workflow runtime)
  • jq (required by fetch_repos.sh)
  • GitHub token in GITHUB_TOKEN (recommended to avoid strict rate limits)

1) Generate Data Locally

pip install -r requirements.txt
export GITHUB_TOKEN=your_github_token_here

bash scripts/fetch_repos.sh
python scripts/dedupe_repos.py --input /tmp/response.jsonl --output /tmp/repos.json
python scripts/sort_repos.py --input /tmp/repos.json --output plugins/repos.json --cached-repos-list plugins/cached-repos.txt
python scripts/chronicle_repos.py --input plugins/repos.json --output plugins/history.json

2) Run UI Locally

cd ui
npm ci
npm run dev

Configuration

  • GITHUB_TOKEN: GitHub API token used by shell/Python scripts and workflows.
  • GITHUB_SSL_VERIFY (optional): set to false to disable TLS verification in sort_repos.py requests.
  • --keeping-days (optional): chronicle_repos.py flag to change history retention days (default 30).

Data and Reporting Outputs

Counts and top repositories are intentionally not hardcoded in this README because they change on every run.

License

MIT License. See LICENSE.

Ver o original no GitHub

Bom para

Relacionados