deep-research

Orchestrates a 13-agent pipeline for systematic academic research and APA 7.0 report generation.

Updated Apr 27, 2026
One-click install
npx skills add https://github.com/DUT-AI/intelligent-testing --skill deep-research-dut-ai
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: deep-research
Source: https://github.com/DUT-AI/intelligent-testing/tree/main/.agents/skills/deep-research
Command: npx skills add https://github.com/DUT-AI/intelligent-testing --skill deep-research-dut-ai

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Conducting rigorous academic research requires coordinating literature search, source verification, synthesis, bias checking, and report writing — a process that is slow, error-prone, and vulnerable to citation fabrication and confirmation bias when done ad hoc. ## Core Features & Use Cases - 13-Agent Research Pipeline: Coordinates specialized agents across 6 phases (scoping, investigation, analysis, composition, review, revision) with mandatory devil's advocate checkpoints and ethics review. - 8 Operational Modes: Full research, quick brief, paper review, literature review, three-way WHY/HOW/WHAT paper scan, fact-check, Socratic guided dialogue, and PRISMA 2020 systematic review with optional meta-analysis. - Citation Integrity Controls: Cross-index triangulation against Semantic Scholar, OpenAlex, Crossref, and arXiv to detect hallucinated or contaminated references before they enter a report. - Use Case: A graduate student with only a vague interest in declining birth rates and higher education uses Socratic mode to converge on a FINER-scored research question, then escalates to full mode for a verified, APA 7.0-formatted research report. ## Quick Start Ask the agent to research a topic such as the impact of AI on higher education quality assurance, or ask it to guide your research when you only have a vague direction.

Frequently Asked Questions about deep-research

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

FAQPage Schema
How do I run a systematic literature review with PRISMA compliance?

Use the systematic-review mode, which follows a 5-phase PRISMA 2020 protocol covering protocol registration, systematic search, screening, data extraction with risk-of-bias assessment (RoB 2 and ROBINS-I), and synthesis with optional meta-analysis and GRADE grading.

How do I choose between full, quick, and Socratic research modes?

Choose Socratic mode when you have a vague idea and need guided questioning to form a research question. Use full mode with a clear research question for a complete 3,000-8,000 word APA report, or quick mode for a 500-1,500 word brief.

How does the pipeline detect fabricated or hallucinated citations?

Each bibliography entry is checked against Semantic Scholar, OpenAlex, Crossref, and arXiv at ingest time. Unmatched entries are flagged as contamination signals, and references that cannot be verified are treated as failures rather than uncertain.

Can I verify specific claims without running a full research pipeline?

Yes, fact-check mode activates only the source verification agent to produce a 300-800 word verification report with per-claim verdicts, evidence hierarchy grading, and predatory journal screening.

What are the limitations of the three-way paper scan mode?

Three-way-scan only produces compact WHY/HOW/WHAT extractions per paper plus a cross-paper synthesis of 800-2,000 words. It does not provide thematic synthesis or PRISMA coverage; escalate to lit-review or systematic-review mode for those needs.