Weizhena/Deep-Research-skills

Structured deep research skill for Claude Code/Open Code/Codex with human-in-the-loop control

What it solves

This project provides a structured, two-phase research workflow for AI coding assistants (Claude Code, OpenCode, and Codex), transforming them from simple chat interfaces into systematic research tools. It prevents the AI from skipping steps or providing superficial answers by enforcing a process of outline generation and deep investigation before reporting.

How it works

The system implements a multi-step workflow using specialized skills and agents:

  1. Outline Generation: The user triggers /research to create a structured list of items to investigate and the specific fields of information to collect for each.
  2. Refinement: Users can manually add more items or fields using /research-add-items and /research-add-fields to ensure the scope is complete.
  3. Deep Investigation: The /research-deep command triggers parallel agents to search the web for each item in the outline, collecting detailed data into JSON format.
  4. Reporting: The /research-report command compiles the gathered data into a final, organized Markdown report.

Who it’s for

It is designed for users of Claude Code, OpenCode, and Codex who need to perform comprehensive research for academic surveys, technical evaluations, market analysis, or due diligence.

Highlights

  • Human-in-the-loop: Allows users to modify the research outline before the deep search begins.
  • Multi-platform support: Compatible with Claude Code, OpenCode, and Codex.
  • Structured output: Moves from a flexible outline to JSON data and finally to a formatted Markdown report.
  • Parallel execution: Uses agents to conduct deep research on multiple items efficiently.

Related

  • Project
    glebis/claude-skills

    A collection of ~100 specialized skills for Claude Code that automate complex workflows across development, research, media generation, and personal analytics.

  • Project
    luwill/research-skills

    A collection of Claude Code skills designed to automate academic research workflows, including exhaustive literature retrieval, verified proposal writing, and high-fidelity slide generation.

  • Project
    CloudAI-X/claude-workflow-v2

    A universal workflow plugin for Claude Code and other AI agents that provides specialized agents, slash commands, and automation hooks to streamline software development.

  • Project
    feiskyer/claude-code-settings

    A collection of custom skills, sub-agents, and configuration templates to extend Claude Code with capabilities like deep research, image generation, and GitHub automation.

  • Project
    claesbackman/AI-research-feedback

    A collection of Claude Code skills that automate academic peer review, grant proposal evaluation, and paper-code reproducibility checks for researchers.