skills.rest
Dozyboy
DozyboyCOMMUNITY
Explore all skills by this author →
GitHub

hs:research

Gathers and verifies multi-source technical evidence into a ranked research report.

Updated Jul 19, 2026
Search Tags:#research#web search#trade-off analysis#technology evaluation#evidence verification#technical report#source triangulation
One-click install
npx skills add https://github.com/Dozyboy/VSF --skill hs-research-dozyboy
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: hs:research
Source: https://github.com/Dozyboy/VSF/tree/main/Day2_VSF/Demo1/harness/plugins/hs/skills/research
Command: npx skills add https://github.com/Dozyboy/VSF --skill hs-research-dozyboy

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Technical decisions often rest on unverified claims, single sources, or outdated documentation. This Skill structures technical research into an evidence-anchored report where every claim carries a URL or file:line anchor, and unverified statements are explicitly tagged as assumed. ## Core Features & Use Cases - Two research modes: breadth mode surveys up to 5 sources to compare multiple options; depth mode drills up to 10 sources into one topic including changelogs, CVEs, and migration paths. - Evidence verification: applies a two-way Evidence Filter classifying claims as VERIFIED, ASSUMED, PRIOR, or CONFLICTING, with source triangulation across official docs, maintainer blogs, and production case studies. - Structured reports: generates a ranked recommendation with a trade-off matrix, references, and open questions saved to plans/reports/. - Use Case: Before choosing a real-time sync library, run a breadth research pass to compare options on performance, security, and maturity, then receive a ranked recommendation with anchored evidence. ## Quick Start Ask the assistant to research a technical question such as comparing authentication libraries, optionally specifying breadth or depth mode.

Frequently Asked Questions about hs:research

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

FAQPage Schema
How do I compare technical options with verified sources?▼

Run a breadth-mode research pass with your topic and evaluation criteria. It searches up to 5 sources in parallel, builds a trade-off matrix across performance, security, and maturity, and returns a ranked recommendation with anchored references.

What is the difference between breadth and depth research modes?▼

Breadth mode surveys up to 5 sources to compare multiple options quickly. Depth mode drills into one topic using up to 10 sources, covering changelogs, GitHub issues, CVEs, benchmarks, and migration paths.

How are unverified claims handled in research reports?▼

Claims without a URL or file:line anchor are tagged ASSUMED, or PRIOR if based on training knowledge. They are listed in the report's Open questions section and must not be used as the basis for conclusions.

Can research be delegated to a sub-agent for large scopes?▼

Yes. Use the --delegate flag or let the skill trigger delegation when scope exceeds 5 independent sources. A researcher agent receives the question, criteria, and output path, then returns the saved report path.

Where are research reports saved and what format do they use?▼

Reports are saved to plans/reports/ using the pattern <slug>-research-<YYMMDD>.md. Each report contains a summary, trade-off matrix, ranked recommendation, evidence references, and open questions.

Related Skills

Explore More Education & Research Skills→
datorresbdatorresb

technical-research

Evaluate technologies with defined criteria and document decisions in ADRs.

Community
Advanced
zhanglg12zhanglg12
47

research-assist

Discover academic papers via Zotero profiles and arXiv semantic searches.

Community
Advanced
patricksavallepatricksavalle
3

scientific-fact-classification

Classify empirical claims by evidence strength and methodological rigor.

Community
Advanced
ZaxbyHubZaxbyHub

consult

Answers advisory questions by consulting domain SME agents with bounded evidence and cached guidance.

Official
Basic