research

Conducts web research on a topic and generates a cited markdown report.

Updated Mar 23, 2026
One-click install
npx skills add https://github.com/sakamotomomotaro0809-netizen/tateyomi --skill research-sakamotomomotaro0809-netizen
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: research
Source: https://github.com/sakamotomomotaro0809-netizen/tateyomi/tree/main/taisun_agent/.claude/skills/research
Command: npx skills add https://github.com/sakamotomomotaro0809-netizen/tateyomi --skill research-sakamotomomotaro0809-netizen

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Gathering reliable information on a topic requires manual searching, reading dozens of pages, and organizing findings with sources. This Skill automates the entire deep-research workflow from web search to a structured, citation-backed report. ## Core Features & Use Cases - Automated Web Research: Searches the web, fetches key pages, and collects 10-30 pieces of evidence into a structured evidence.jsonl file. - Cited Report Generation: Produces a report.md with a summary, 3-5 key findings, sources with retrieval dates, and open questions. - Organized Run Archives: Each investigation is stored in a timestamped run directory with input.yaml, evidence.jsonl, and report.md for traceability. - Use Case: Ask it to research "2026 SaaS market trends" and receive a markdown report where every claim links to a verifiable source, with uncertainties explicitly flagged. ## Quick Start Ask the assistant to run the research skill on a topic such as "the latest trends in AI agents" and it will search the web, gather evidence, and produce a cited report.

Frequently Asked Questions about research

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

FAQPage Schema
How do I run a deep research task on a topic?▼

Invoke the research skill with a topic, for example "/research AI agent trends". It creates a timestamped run directory, searches the web, fetches key pages, and writes a report.md with cited findings.

What output files does the research workflow generate?▼

Each run produces three files in a timestamped directory: input.yaml recording the query, evidence.jsonl containing collected sources, and report.md as the main deliverable with summary, findings, and a source list.

Does the research report include sources for its claims?▼

Yes, every finding in the report must include a source URL and retrieval date. The skill's rules require citations for all claims and explicit labeling of uncertain or contradictory information.

Can I use web research results for medical or legal decisions?▼

The skill does not make definitive claims on investment, medical, or legal topics. It only presents evidence with sources, so professional advice should be sought for decisions in those domains.

How many sources does a research run typically collect?▼

A typical run gathers 10 to 30 evidence items via web search, with important URLs fetched in full. The evidence is stored in evidence.jsonl for later review or reuse.