Dragster

 ██████╗ ██████╗  █████╗  ██████╗ ███████╗████████╗███████╗██████╗
 ██╔══██╗██╔══██╗██╔══██╗██╔════╝ ██╔════╝╚══██╔══╝██╔════╝██╔══██╗
 ██║  ██║██████╔╝███████║██║  ███╗███████╗   ██║   █████╗  ██████╔╝
 ██║  ██║██╔══██╗██╔══██║██║   ██║╚════██║   ██║   ██╔══╝  ██╔══██╗
 ██████╔╝██║  ██║██║  ██║╚██████╔╝███████║   ██║   ███████╗██║  ██║
 ╚═════╝ ╚═╝  ╚═╝╚═╝  ╚═╝ ╚═════╝ ╚══════╝   ╚═╝   ╚══════╝╚═╝  ╚═╝

Skills for local document ingestion and semantic search. Built for AI coding agents (Letta Code, Claude Code, OpenCode, etc.).

Quick Start

# Install tools
npm install -g @tobilu/qmd
npm install -g @llamaindex/liteparse

# Install all skills
npx skills add etalab-ia/dragster

# Or install individual skills
npx skills add etalab-ia/dragster --skill rag-parse
npx skills add etalab-ia/dragster --skill rag-index
npx skills add etalab-ia/dragster --skill rag-search
npx skills add etalab-ia/dragster --skill rag-tracking

Skills

Skill	Purpose	Provider
rag-parse	PDF/DOCX/PPTX/XLSX/Images → Markdown/JSON	liteparse
rag-index	Index documents for semantic search	qmd
rag-search	Search indexed documents	qmd
rag-tracking	Track ingestion state & issues	ctx (optional)

Note: rag-tracking is designed for agents without built-in memory (Claude Code, Codex, OpenCode). Not needed for Letta Code which has native persistent memory.

Workflow

Documents ──[/rag-parse]──> Markdown ──[/rag-index]──> Index ──[/rag-search]──> Results

1. Parse: Convert documents to markdown or JSON (PDF, DOCX, PPTX, XLSX, images)
2. Index: Create searchable embeddings with qmd
3. Search: Query your knowledge base

Provider System

Skills support multiple providers via metadata in each SKILL.md:

---
name: search
provider: qmd
available-providers:
  - qmd
  - pinecone
  - weaviate
---

• provider: Current active provider (editable by agents)
• available-providers: Providers this skill can support

To switch providers, ask your agent to update the provider field in the skill file.

Integration

Works with:

• Letta Code
• Claude Code
• OpenCode
• Pi
• Cursor
• Any agent that supports skills

See docs/integration-guide.md for detailed setup instructions.

Documentation

• Integration Guide — How to use with various AI agents
• qmd Setup — Detailed qmd installation and configuration

Requirements

• Node.js 18+ or Bun
• liteparse: npm install -g @llamaindex/liteparse
• qmd: npm install -g @tobilu/qmd

License

MIT