What problem does it solve?
This Skill eliminates fragmented, non-reproducible research workflows by automating multi-source evidence gathering, extraction, and synthesis while persisting raw sources and distilled notes to disk for resumability, auditability, and reproducible citations.
Core Features & Use Cases
- Filesystem-backed workflow: Persist scraped pages, an indexed sources registry, per-question notes, and a final report under research/{topic-slug}/ for resumable investigations.
- Search-Read-Extract loop: Generate diverse search queries, batch-scrape results, extract findings and data points, and record source metadata with numbered citation mapping.
- Recursive deepening and synthesis: Support configurable breadth/depth recursion, cross-reference evidence, flag contradictions, and produce a final evidence-weighted report with limitations and confidence levels.
- Use Case: Run a literature review or fact-check that finds primary sources, extracts methodology and data, and compiles a cited executive summary and bibliography.
Quick Start
Ask the deep-research agent to perform a literature review on "topic" and save a cited report to research/topic-slug/report.md.
