deep-research

Orchestrate multi-source research and generate cited markdown reports.

39|10|Updated Mar 6, 2026
One-click install
npx skills add https://github.com/NikitaDmitrieff/auto-co-meta --skill deep-research-nikitadmitrieff
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: deep-research
Source: https://github.com/NikitaDmitrieff/auto-co-meta/tree/main/.claude/skills/deep-research
Command: npx skills add https://github.com/NikitaDmitrieff/auto-co-meta --skill deep-research-nikitadmitrieff

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) components.

What problem does it solve?

Collecting, verifying, and synthesizing information from numerous sources to produce authoritative, citation-backed reports without manual orchestration.

Core Features & Use Cases

  • Eight-phase research pipeline (Scope → Plan → Retrieve → Triangulate → Synthesize → Critique → Refine → Package) with automated validation and source credibility scoring.
  • Autonomy at runtime: default to executing and delivering without user prompts, with safe escalation when validation fails or data is insufficient.
  • Modes and flexibility: supports quick, standard, deep, and ultradeep analyses to balance speed and quality across complex research tasks.
  • Local, self-contained execution: runs on Claude Code without external dependencies; outputs are stored locally for audit and reproducibility.
  • Comprehensive reporting: generates markdown reports with bibliographies, inline citations, and optional HTML/PDF outputs.

Quick Start

To begin, run the skill against a research question and select a mode (quick, standard, deep, or ultradeep) to start the process.

Frequently Asked Questions about deep-research

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I automate multi-source research and generate citation-backed reports in Claude Code?

Automate multi-source research in Claude Code by running an eight-phase pipeline that retrieves, triangulates, and synthesizes data. It outputs structured markdown reports with inline citations and source credibility scoring.

How does source triangulation and verification work for complex analysis tasks?

Source triangulation and verification work by cross-referencing retrieved data across multiple origins during the research pipeline. The system critiques and refines the synthesized information to ensure validated, authoritative results.

Can I run autonomous deep research locally without external dependencies?

You can run autonomous deep research locally without external dependencies directly within Claude Code. The self-contained execution stores outputs locally to ensure full auditability and reproducibility of the research process.

What is the difference between quick, standard, deep, and ultradeep research modes?

The difference between quick, standard, deep, and ultradeep research modes lies in the balance between speed and quality. You select a mode based on task complexity, with ultradeep targeting the most rigorous multi-source analysis.

When should I avoid using autonomous research pipelines for information retrieval?

You should avoid using autonomous research pipelines for simple lookups, basic fact retrieval, or debugging tasks. The system is designed for complex comparison tasks requiring 10 or more sources with automated validation.

What output formats are generated when packaging verified research results?

Verified research results are packaged into structured markdown reports containing bibliographies and inline citations. The pipeline also supports optional HTML and PDF outputs for delivering the final synthesized analysis.