Deep Research

This is a structured academic literature review system that enforces a strict six-phase workflow: frontier exploration, broad survey, deep dive, code analysis, synthesis, and final report. The key constraint is that you cannot skip phases, especially Phase 3 where it forces you to actually read 8-15 full papers rather than just skimming abstracts. It prioritizes peer-reviewed conference papers over arXiv preprints and includes tooling for multi-source search (paper_finder for conferences, Semantic Scholar, arXiv), PDF downloading, and database management. The whole thing outputs to a structured directory with markdown notes and JSONL databases. Use this when you need a real literature review with accountability, not a quick summary. The phase gates are aggressive but probably necessary to prevent the usual shortcut of writing reports based only on abstracts.

skills.sh installs: 2K
GitHub stars: 459
last updated: Aug 21, 2026

View on GitHub

Install

$ npx -y skills add lingzhi227/agent-research-skills --skill deep-research --agent claude-code

Installs into .claude/skills of the current project.

Files

Deep Research Skill

Trigger

Activate this skill when the user wants to:

"Research a topic", "literature review", "find papers about", "survey papers on"
"Deep dive into [topic]", "what's the state of the art in [topic]"
Uses /research <topic> slash command

Overview

This skill conducts systematic academic literature reviews in 6 phases, producing structured notes, a curated paper database, and a synthesized final report. Output is organized by phase for clarity.

Installation: ~/.claude/skills/deep-research/ — scripts, references, and this skill definition. Output: .//Users/lingzhi/Code/deep-research-output/{slug}/ relative to the current working directory.