---

name: weaver description: Map-Reduce system for comprehensive reports and literature reviews. Breaks topic into chapters, researches each in parallel, then weaves into unified document. Use for deep research, reports, and synthesis.

The Weaver: Deep Research Synthesis Protocol

When to Activate

Use this skill when the user requests:

• Comprehensive literature reviews
• Long-form reports or documents
• Deep research on complex topics
• Multi-chapter synthesis
• User explicitly mentions "weaver", "report", or "literature review"

Architecture

Map-Reduce with Editorial Synthesis

Phase 1: The Loom (Map)

• Planner agent breaks topic into Table of Contents
• Typically 5-10 chapters/sections
• Each chapter has clear scope and questions

Phase 2: Spinning (Execute)

• Spawn N parallel Worker agents (one per chapter)
• Each worker researches their chapter deeply
• Workers operate independently (no cross-talk)

Phase 3: Weaving (Reduce)

• Editor agent ingests all chapter reports
• Identifies gaps, contradictions, redundancies
• Stitches into coherent Master Document
• Adds transitions, executive summary, conclusions

Invocation

Copypython3 ~/.claude/skills/weaver/weaver.py "Your research topic"

Options

Copypython3 ~/.claude/skills/weaver/weaver.py "Your topic" --chapters 8 --output report.md

Example Usage

User: "Write a comprehensive report on Australia's critical minerals opportunity"

Copypython3 ~/.claude/skills/weaver/weaver.py "Australia's critical minerals opportunity: resources, processing, markets, and policy"

Output Structure

Copy1. Executive Summary
2. [Chapter 1 - from Worker 1]
3. [Chapter 2 - from Worker 2]
...
N. Conclusions and Recommendations
N+1. Sources and Further Reading

Why This Works

1. Parallel Depth: Each chapter gets dedicated deep research
2. No Context Limits: Workers don't share context, avoiding truncation
3. Editorial Coherence: Final pass ensures unified voice and flow
4. Comprehensive Coverage: Structured approach ensures no gaps

Requirements

• Claude Code CLI installed and authenticated (claude command available)
• Python package: rich (for formatted output)