deep-research

Generate citation-backed research reports with claim-level validation from complex questions.

17|Updated Feb 19, 2026
One-click install
npx skills add https://github.com/ivanrvpereira/.agents --skill deep-research-ivanrvpereira
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: deep-research
Source: https://github.com/ivanrvpereira/.agents/tree/main/skills/deep-research
Command: npx skills add https://github.com/ivanrvpereira/.agents --skill deep-research-ivanrvpereira

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes scripts (resource) and references (resource) and assets (resource) components.

What problem does it solve?

Deep research transforms complex questions into rigorous, multi-source reports with evidence persistence, claim-level verification, and fully formatted bibliographies—avoiding shallow or citation-free answers.

Core Features & Use Cases

  • Multi-phase research pipeline: scopes the question, plans retrieval, retrieves in parallel, triangulates, synthesizes, critiques, refines, and packages results.
  • Citation & evidence persistence: writes stable registries and append-only JSONL artifacts (sources/evidence/claims) so multi-step continuation stays grounded.
  • Automated validation: runs report structure validation and citation/DOI verification, then iterates to fix issues.
  • Structured outputs: generates Markdown (primary), plus HTML and PDF-ready deliverables, with long-report continuation when needed.
  • Use cases: compare technologies (e.g., X vs Y), produce state-of-the-art analyses, and generate decision-grade research summaries for critical initiatives.

Quick Start

Ask your agent: deep research in ultradeep mode: compare PostgreSQL vs Supabase for our stack

Frequently Asked Questions about deep-research

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

FAQPage Schema
How do I generate a literature review with verified citations?

To generate a citation-backed research report, provide a complex research question or topic to analyze. The pipeline scopes the query, retrieves sources in parallel, triangulates evidence, and synthesizes findings into Markdown, HTML, and PDF-ready deliverables with an evidence ledger.

How does claim verification work in automated research reports?

Deep research validates claims by writing stable registries and append-only JSONL artifacts that log sources, evidence, and claims. It runs automated validation checks on report structure and DOI citations, iterating to fix issues and maintain evidence persistence across multi-step continuations.

Can I use deep research for technology comparison and market investigation tasks?

Yes, you can use deep research for technology comparisons and market investigations. The tool performs multi-source analysis to compare technologies, produce state-of-the-art reviews, and generate decision-grade research summaries across configurable deep and ultradeep modes.

What output formats are supported for structured research reports?

Structured research reports are output in Markdown as the primary format, along with HTML and PDF-ready deliverables. The pipeline supports long-report continuation when needed to ensure comprehensive packaging of extensive multi-source analysis results.

What is the best way to maintain evidence persistence during multi-step research?

Evidence persistence is maintained by writing stable registries and append-only JSONL artifacts tracking sources, evidence, and claims. This retrieval-to-triangulation-to-synthesis workflow ensures multi-step research continuations stay grounded and validation-checked.