deep-research

Conducts academic literature research with verified citations across PubMed, arXiv, and Semantic Scholar APIs.

5|Updated May 17, 2026
One-click install
npx skills add https://github.com/s3644/academic-copilot-skills --skill deep-research-s3644
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: deep-research
Source: https://github.com/s3644/academic-copilot-skills/tree/main/.github/skills/deep-research
Command: npx skills add https://github.com/s3644/academic-copilot-skills --skill deep-research-s3644

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Academic research requires searching multiple literature databases, verifying every citation, and synthesizing evidence without fabrication — a slow, error-prone process when done manually. This Skill automates literature discovery, cross-verifies every reference via DOI/CrossRef, and produces structured research outputs with enforced anti-hallucination rules. ## Core Features & Use Cases - Multi-API Literature Search: Queries PubMed, arXiv, Semantic Scholar, Scopus, IEEE Xplore, and Google Scholar with domain-based source selection and no API keys required for core APIs. - Seven Research Modes: Socratic guided dialogue, full APA 7.0 reports, quick research briefs, paper review, literature review, fact-checking, and PRISMA 2020 systematic reviews. - Citation Integrity: Every reference is verified via CrossRef/DOI and graded by source tier (PubMed > Semantic Scholar > arXiv > gray literature) before inclusion. - Use Case: A graduate student asks for a systematic review on rTMS for depression; the Skill searches PubMed for meta-analyses, verifies each DOI, grades evidence strength, and produces a PRISMA-compliant report with a flow diagram and limitations section. ## Quick Start Ask the assistant to research a topic such as "do a deep research literature review on transformer interpretability methods" and it will search academic APIs, verify sources, and produce a cited report.

Frequently Asked Questions about deep-research

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

FAQPage Schema
How do I do a literature review using PubMed and arXiv APIs?▼

The Skill selects the primary API by topic domain — PubMed for biomedical topics, arXiv for CS/ML/physics — then searches, fetches abstracts, and synthesizes findings. Query syntax and endpoints for each API are documented in the references folder.

How does the Skill prevent fabricated citations?▼

Every reference must be verified via its DOI through the CrossRef API or direct resolution before inclusion. Sources are graded into tiers, and any reference that cannot be confirmed is excluded rather than marked uncertain.

Do I need API keys for PubMed, arXiv, or Semantic Scholar?▼

No API keys are required for the core search APIs; PubMed, arXiv, and Semantic Scholar all work without keys at lower rate limits. Optional keys raise rate limits, and Scopus or IEEE Xplore keys are only needed for those specific databases.

Can it produce a PRISMA-compliant systematic review?▼

Yes, the Systematic Review mode follows PRISMA 2020 guidelines, generating the identification-to-inclusion flow diagram, the 27-item checklist, and per-study risk of bias assessment, with optional meta-analysis when studies are homogeneous.

What is the difference between Socratic mode and Full research mode?▼

Socratic mode guides you through a five-layer questioning dialogue without giving direct answers, ideal for vague ideas. Full mode executes a six-phase pipeline from scoping to revision and produces a complete 3000-8000 word APA 7.0 report.

When should I avoid using Google Scholar for literature search?▼

Google Scholar has no official API and scraping triggers CAPTCHAs after roughly five requests, so it is a last resort for gray literature only. Any result found there must be cross-verified against Semantic Scholar or a DOI before citing.