Research Assistant

View saved version →

---
name: academic-research
description: Find, evaluate, compare, and synthesize academic papers and scholarly evidence.
---

# Academic Research

Use for literature reviews, paper discovery, methodology comparison, citation support, and research-gap analysis.

## Method

1. Refine the question into population, intervention or exposure, comparison, outcome, context, and date scope where applicable.
2. Define inclusion and exclusion criteria before collecting sources for a review.
3. Prefer original peer-reviewed research, systematic reviews, official datasets, and clearly labeled preprints.
4. Extract each paper's question, design, sample, data, measures, result, limitations, and relevance.
5. Distinguish correlation from causation, statistical from practical significance, and author claims from synthesis.
6. Explain disagreements through differences in population, methods, measures, assumptions, or timeframe.
7. Identify gaps as missing populations, outcomes, methods, replication, implementation contexts, or unresolved contradictions.

Never fabricate authors, dates, journals, DOIs, quotations, sample sizes, or effect sizes. Label preprints and secondary summaries accurately.

View saved version →

---
name: evidence-synthesis
description: Combine heterogeneous sources into a transparent evidence brief with confidence, limitations, and disagreements.
---

# Evidence Synthesis

Use when the user provides multiple sources or asks for a conclusion across papers, reports, datasets, interviews, or web pages.

## Method

1. Create a claim inventory and map each claim to supporting or contradicting evidence.
2. Assess directness, authority, methodology, sample or data quality, recency, and independence.
3. Keep incompatible populations, outcomes, and definitions separate.
4. Label conclusions as strong, moderate, limited, mixed, or not established, explaining why.
5. Report missing evidence, possible bias, alternative explanations, and confidence limits.
6. Make recommendations only after separating evidence from judgment and assumptions.

Use a table with `Claim`, `Evidence`, `Agreement or conflict`, `Limitations`, and `Confidence` when it improves clarity.

View saved version →

---
name: product-research
description: Research products, competitors, markets, customers, pricing, and positioning for product decisions.
---

# Product Research

Use for competitor analysis, market scans, customer-problem research, pricing comparisons, product positioning, and buy/build decisions.

## Method

1. Define the decision, target segment, geography, timeframe, and comparison criteria.
2. Separate direct competitors, substitutes, adjacent tools, and non-consumption.
3. Gather evidence on capabilities, pricing, limits, target users, distribution, claims, and observable user experience.
4. Distinguish verified facts from vendor claims, reviews, inference, and recommendation.
5. Normalize pricing units and plan terms before comparing them.
6. End with a decision matrix, gaps in evidence, and what should be validated with real users.

Do not invent market size, customer counts, revenue, satisfaction, or competitor capabilities. Avoid copying marketing language as if it were independent evidence.

View saved version →

---
name: research-assistant
description: Support academic research with question refinement, paper discovery, methods comparison, evidence synthesis, and accurate citations.
---

# Research Assistant

Use this skill for academic research, literature reviews, evidence synthesis, paper comparison, citation support, and research-gap analysis.

## Research protocol

1. Refine the research question into population, intervention or exposure, comparison, outcome, context, and date scope when applicable.
2. Define inclusion and exclusion criteria before collecting papers when the user requests a review or systematic comparison.
3. Prefer original peer-reviewed papers, official datasets, preprints clearly labeled as such, and authoritative evidence reviews.
4. For each important paper, record the question, design, sample, intervention or exposure, outcome, main result, limitations, and relevance.
5. Distinguish correlation from causation, exploratory from confirmatory analysis, statistical significance from practical significance, and author claims from the plugin's synthesis.
6. Compare papers by population, methods, measurements, timeframe, assumptions, and quality—not only by conclusion.
7. Explain disagreement rather than averaging incompatible findings.
8. Identify evidence gaps, replication concerns, publication bias, confounders, and generalizability limits when relevant.

## Output modes

### Literature review

Return scope and criteria, a paper table, thematic synthesis, methodological limitations, disagreements, research gaps, and a properly formatted reference list.

### Paper summary

Use: research question, method, sample or data, results, limitations, practical meaning, and questions for follow-up.

### Paper comparison

Use a consistent table with citation, population, design, data, intervention or exposure, outcome, main finding, limitations, and confidence.

### Evidence brief

Start with the answer, then provide the strongest evidence, uncertainty, alternative explanations, and implications.

### Research-gap analysis

Separate gaps that are missing populations, missing outcomes, weak methods, unresolved contradictions, outdated evidence, or untested implementation contexts.

## Citation integrity

- Cite only papers or sources actually located or supplied.
- Link to the paper landing page, DOI, publisher page, or reliable repository.
- Never fabricate authors, dates, journal names, DOIs, quotations, sample sizes, or effect sizes.
- Label preprints, conference abstracts, reviews, and secondary summaries accurately.
- Quote sparingly and keep quotations short; prefer faithful paraphrase.

## Safety and scope

For medical, legal, or other high-stakes questions, explain that research synthesis is not professional advice and highlight study limitations. Do not convert weak or preliminary evidence into a definitive recommendation.

View saved version →

---
name: web-research
description: Conduct current web research with source evaluation, claim comparison, citations, and freshness checks.
---

# Web Research

Use for current events, organizations, products, services, regulations, technical documentation, and general fact-finding.

## Method

1. Break the request into answerable subquestions.
2. Prefer primary sources: official documentation, government pages, direct announcements, filings, original datasets, and first-party product pages.
3. Record publication date, event date, source type, relevant claim, and limitations.
4. Compare independent sources and flag conflicting definitions or stale information.
5. Cite direct source URLs near every material current or non-obvious claim.

Never treat search snippets as evidence. Never present stale information as current. State the research date when freshness matters.