convez.ai
SkillAutomaçãoIntermediário
maaarcooo

Claude Skills - Processamento de Documentos e Geração de Materiais

maaarcooo/claude-skills

Extração de PDFs, geração de flashcards e notas de revisão para Claude

Ver no GitHub
BomPluginScore · de 0 a 100

Coleção de skills agentic para Claude especializadas em extração e limpeza de PDFs, geração automática de flashcards Anki, criação de notas de revisão e continuidade de sessões. Oferece workflows robustos com métodos de extração dual e tratamento inteligente de tabelas, imagens e texto fragmentado.

Instalação

$npx convez-ai maaarcooo/claude-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
12
Forks
0
Acompanhando
0
Saúde
42%
Último commit
há 2 semanas
Versão
notes-v3.6
Issues abertas
0
Contribuidores
2

Casos de uso

  • Extrair conteúdo limpo de PDFs para markdown estruturado
  • Gerar decks de flashcards Anki automaticamente a partir de materiais de estudo
  • Criar notas de revisão concisas e organizadas
  • Manter continuidade de sessões com handoff estruturado

Veredito

Conjunto maduro e bem-estruturado de skills para automação de fluxos educacionais e de processamento documental. Implementação sólida com documentação clara e suporte a casos complexos (tabelas malformadas, imagens, texto refluxado).

README

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

Claude Skills

A collection of Claude Agentic Skills for intelligent document processing, study material generation, PDF extraction, and session continuity.

Overview

This repository contains specialized skills for Claude and Claude Code. Skills are organized into two categories: General Claude Skills for document processing, study material generation, and session handoff (works in any Claude interface) and Claude Code Skills providing slash commands for session continuity in the Claude Code CLI.

Skills Overview

General Claude Skills

SkillPurposeStatus
extracting-pdfsExtract and clean PDF content to markdownActive
anki-flashcard-generatorGenerate Anki-importable flashcard decksActive
revision-notes-generatorCreate concise revision notes from study materialsActive
handoffCreate structured handoff documents for session continuityActive

Claude Code Skills

SkillPurposeStatus
handoff (Claude Code)/handoff slash command for session handoffActive
resume/resume slash command to continue from a handoffActive

Archived Skills

SkillPurposeStatus
pdf-extractPDF extraction (development version)Archived
pdf-to-markdown-converterLegacy PDF conversionArchived

Project Structure

claude-skills/
├── README.md
├── extracting-pdfs/                    # PDF extraction skill (Active)
│   ├── README.md                       # Skill documentation
│   ├── SKILL.md                        # Skill definition & workflow
│   ├── cleanup-patterns.md             # Reference: noise patterns to remove
│   ├── image-handling.md               # Reference: processing extracted images
│   ├── sentence-reflow.md              # Reference: fixing fragmented text
│   └── table-formatting.md             # Reference: reconstructing malformed tables
├── anki-flashcard-generator/           # Anki flashcard generation skill
│   ├── README.md                       # Skill documentation
│   └── SKILL.md
├── revision-notes-generator/           # Revision notes generation skill
│   ├── README.md                       # Skill documentation
│   └── SKILL.md
├── handoff/                            # Session handoff skill (General Claude)
│   ├── README.md
│   └── SKILL.md
├── claude-code-skills/                 # Claude Code slash commands
│   ├── handoff/                        # /handoff command
│   │   ├── README.md
│   │   └── SKILL.md
│   └── resume/                         # /resume command
│       ├── README.md
│       └── SKILL.md
└── archive/                            # Archived/legacy skills
    ├── pdf-extract/                    # PDF extraction (development version)
    │   ├── SKILL.md
    │   ├── extract_pdf.py              # Core Python extraction script
    │   ├── cleanup-patterns.md
    │   ├── image-handling.md
    │   ├── sentence-reflow.md
    │   └── table-formatting.md
    └── pdf-to-markdown-converter/      # Legacy skill (Deprecated)
        └── SKILL.md

Skills

Extracting PDFs

The primary skill for extracting PDF content to clean, organized markdown. Uses a 5-step workflow (Extract, Analyse, Clean, Organise, Output) with dual extraction methods, comprehensive image handling, and reference guides for common challenges.

Full documentation: See extracting-pdfs/README.md for the complete workflow, CLI usage, dependencies, and reference guide details.

Sample prompt:

<pathname>
Use "extracting-pdfs" skill to convert this pdf into a markdown file.

Anki Flashcard Generator

Generate study flashcards from PDF or Markdown content in Anki-importable format. Card design follows evidence-based principles with six specialised card types, built-in interference checking, and coverage guidance prioritised by examinability.

Full documentation: See anki-flashcard-generator/README.md for design principles, card types, output format, and card design rules.

Sample prompt for claude.ai:

Create an Anki flashcard deck from the attached study materials using the "anki-flashcard-generator" skill.
Output the deck as a .txt file named after the source file (e.g. Physics_Chapter_5.pdf → Physics_Chapter_5.txt).

Sample prompt for Claude API:

Create an Anki flashcard deck from the attached study materials using the "anki-flashcard-generator" skill.
Output only the flashcard lines in the format "Question | Answer", one per line.
Do not include any preamble, headers, explanations, markdown formatting,
or code fences. The raw output will be saved directly to a text file.

Revision Notes Generator

Generate concise, accurate revision notes from PDF or Markdown study materials. Covers a 5-step process from reading through verification to structured markdown output, with writing guidelines prioritising conciseness, completeness, and accuracy.

Full documentation: See revision-notes-generator/README.md for writing guidelines, output format, and the full process.

Sample prompts:

Use "revision-notes-generator" skill to create revision notes of the study materials.
Use "revision-notes-generator" skill to create revision notes of the study materials with title "<title>".

Handoff

Create structured handoff documents for seamless continuity across Claude conversations. When a session ends, the handoff captures the full working state — objective, decisions with rationale, failed approaches, and next steps — so a fresh conversation can continue without re-explanation.

The skill adapts its structure and depth to the type of work (research, writing, planning, building, learning, problem-solving) and the complexity of the session.

Full documentation: See handoff/README.md for work types, depth tiers, document structure, and usage instructions.


Claude Code Skills

Slash commands designed for the Claude Code CLI. These are installed at user-level (~/.claude/skills/) and invoked with /command syntax.


Handoff (Claude Code)

A /handoff slash command for Claude Code that generates structured handoff documents optimised for coding sessions. Captures implementation decisions, failed approaches, error messages, and file-level state that the codebase alone does not preserve. Each handoff references the previous one, creating a traceable handover chain across sessions.

Full documentation: See claude-code-skills/handoff/README.md for work types, depth tiers, the handover chain, and installation instructions.


Resume

A /resume slash command that loads a handoff document and gets you productive immediately. Companion to /handoff — it finds the most recent handoff (or a specific one you name), reads it, reviews the key files listed in it, and continues from the next step.

Full documentation: See claude-code-skills/resume/README.md for usage and the handoff/resume workflow.


Archived Skills

The following skills have been moved to the archive/ folder. They are preserved for reference but are no longer actively maintained.


PDF Extract

Location: archive/pdf-extract/

The development/original version of the PDF extraction skill. Contains the core Python extraction script.

Note: For production use, see extracting-pdfs which is the current active version.

Features:

  • Same extraction capabilities as extracting-pdfs
  • Contains extract_pdf.py script (1,500+ lines)
  • Full reference documentation included

PDF to Markdown Converter

Location: archive/pdf-to-markdown-converter/

The original PDF conversion skill using visual PDF understanding.

Status: Deprecated. Superseded by the more sophisticated extracting-pdfs skill.

Design Philosophy

  • Extract everything — No hardcoded cleanup rules during extraction
  • Preserve raw content — Keep data intact for intelligent post-processing
  • Rich metadata — Provide comprehensive context for document understanding
  • Manual over automated — Complex decisions handled manually for better results
  • Atomic flashcards — One fact per card for effective learning
  • Accuracy first — Verify and correct information in study materials
  • Structured continuity — Capture working state, not conversation summaries, so sessions can resume without re-explanation

Technologies

  • Python 3 — Core scripting language
  • PyMuPDF (fitz) — Low-level PDF reading and image extraction
  • PyMuPDF4LLM — Enhanced markdown formatting with table support
  • YAML/JSON — Metadata formats
  • Markdown — Output format

Versioning

This repository uses unified versioning. All skills share a single version number. See Releases for version history.

License

This work is licensed under CC BY 4.0 - you're free to share and adapt with attribution.

Ver o original no GitHub

Bom para

Relacionados