ZoteroSynth

Environment

!cd "$SKILL_DIR" && uv run scripts/zotero_client.py check 2>/dev/null || echo '{"status":"error","hint":"Run: uv sync in $SKILL_DIR"}'

Setup

Run all scripts from $SKILL_DIR:

cd "$SKILL_DIR"

Commands

Read COMMANDS.md for full CLI reference. Core commands:

• uv run scripts/zotero_client.py collections --tree
• uv run scripts/zotero_client.py search "QUERY" --limit N
• uv run scripts/zotero_client.py detail KEY
• uv run scripts/pdf_extract.py "PATH" --max-pages 30

Workflow

1. Single paper: detail KEY → pdf-path KEY → pdf_extract.py PATH → Summarize per summarize.md
2. Multi-paper: items --collection KEY or search QUERY → Map: summarize each (300-600 tokens) → Reduce: synthesize per review.md or synthesize.md

Rules

• Parse JSON output, present in natural language
• Include citation: [Author, Year, item_key]

• 10 papers: use Map-Reduce

• Fallback: pip install pyzotero pymupdf && python scripts/xxx.py
• Setup help: ADVANCED.md