skills.rest
harunjeylan
harunjeylanCOMMUNITY
Explore all skills by this author →
WebsiteGitHub

deep-research

Conducts iterative multi-engine web research and generates cited academic-style reports.

Updated Apr 12, 2026
Search Tags:#document conversion#web search#report generation#citations#source analysis#deep research#multi-engine search
One-click install
npx skills add https://github.com/harunjeylan/pi-agent --skill deep-research-harunjeylan
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: deep-research
Source: https://github.com/harunjeylan/pi-agent/tree/main/skills/deep-research
Command: npx skills add https://github.com/harunjeylan/pi-agent --skill deep-research-harunjeylan

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Shallow single-query searches often miss critical information and produce unsourced answers. This Skill performs structured, iterative research across multiple search engines, processes documents, and synthesizes findings into a properly cited report. ## Core Features & Use Cases - Iterative Multi-Engine Search: Runs up to 10 searches across Bing, DuckDuckGo, Brave, Baidu, CSDN, Juejin, and Linux.do with gap analysis between rounds to refine queries. - Document Processing: Converts PDF, DOCX, and PPTX files to markdown via markitdown and fetches GitHub READMEs and Chinese tech articles. - Academic Citations: Assigns SOURCE_IDs to each source and produces reports with in-text citations, quotes with page numbers, and a References section. - Use Case: Ask it to research "quantum computing applications in healthcare" and receive a structured session folder containing search logs, per-source analyses, a synthesis, and a final cited report. ## Quick Start Ask the agent to research a topic in depth, for example: use deep research to investigate the current state of solid-state battery technology 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 run a deep research query on a topic?▼

Invoke the skill with a research topic, such as asking it to research a subject in depth. The agent executes up to 10 iterative searches across multiple engines, analyzes gaps between rounds, and produces a cited final report.

What search engines does multi-engine research use?▼

It rotates through Bing, DuckDuckGo, Brave, and Baidu for general searches, plus CSDN, Juejin, and Linux.do for Chinese tech content and community discussions. Engines are never repeated consecutively to avoid rate limits.

Can it process PDF and DOCX documents found during research?▼

Yes, PDF, DOCX, and PPTX files are converted to markdown using markitdown during the search phase. Web pages, GitHub READMEs, and Chinese platform articles are fetched with dedicated operations.

How are citations formatted in the research report?▼

Each source receives a SOURCE_ID in AuthorYear format, such as OpenAI2024. Quotes include the SOURCE_ID and page number, statistics carry SOURCE_ID attribution, and the final report ends with a full References section.

What are the limitations of iterative multi-engine search?▼

Research is capped at 10 searches and 3-5 deeply analyzed sources, so extremely broad topics may need scoping. Results also depend on engine availability and rate limits, and Chinese platform fetching requires the specific fetch operations.

Related Skills

Explore More Education & Research Skills→
danny0926danny0926

research-scope

Define research scope with hypotheses, bias checks, and prior art references.

Community
Advanced
aplaceforallmystuffaplaceforallmystuff
17

log-to-project

Log project activity to CHANGELOG.md, Sessions/, and Decisions/ folders.

Community
Intermediate
thinkingoraclethinkingoracle

settle-this

Resolves factual disputes by checking primary sources and naming them.

Community
Basic
jpologjpolog

literature-search

Automate scholarly literature discovery across multiple databases with reproducible logging.

Community
Advanced