Scholarly Research
Copyright (c) 2026 KunAgent. Licensed under the MIT License.
Purpose
Search, screen, synthesize, and cite scholarly literature and author metadata.
Tool routing
| Tool or skill |
Use |
web_fetch |
Fetch public abstracts, metadata, DOI pages, or authorized full text. |
browser_use |
Search and inspect public scholarly portals interactively. |
Workflow
- Translate the question into concepts, synonyms, exclusions, years, and study types.
- Search public scholarly portals using more than one query formulation.
- Screen by relevance and evidence quality, recording inclusion reasons.
- Extract design, population, intervention or exposure, outcomes, limitations, and identifiers.
- Synthesize agreements, conflicts, gaps, and confidence.
Completion gates
- Verify DOI, title, authors, venue, and year.
- Distinguish preprints from peer-reviewed versions.
- Do not infer full-text findings from title or abstract alone.
Boundaries
- Respect access restrictions and publisher terms.
- Do not fabricate citations, quotations, metrics, or author profiles.
Delivery
Lead with the outcome, name the evidence used for verification, and disclose any real limitation that remains.
1---2name: scholar-research3description: Search, screen, synthesize, and cite scholarly literature and author metadata.4---5
6# Scholarly Research
7Copyright (c) 2026 KunAgent. Licensed under the MIT License.
8
9## Purpose
10Search, screen, synthesize, and cite scholarly literature and author metadata.
11
12## Tool routing
13| Tool or skill | Use |
14|---|---|
15| `web_fetch` | Fetch public abstracts, metadata, DOI pages, or authorized full text. |
16| `browser_use` | Search and inspect public scholarly portals interactively. |
17
18## Workflow
191. Translate the question into concepts, synonyms, exclusions, years, and study types.
202. Search public scholarly portals using more than one query formulation.
213. Screen by relevance and evidence quality, recording inclusion reasons.
224. Extract design, population, intervention or exposure, outcomes, limitations, and identifiers.
235. Synthesize agreements, conflicts, gaps, and confidence.
24
25## Completion gates
26- Verify DOI, title, authors, venue, and year.
27- Distinguish preprints from peer-reviewed versions.
28- Do not infer full-text findings from title or abstract alone.
29
30## Boundaries
31- Respect access restrictions and publisher terms.
32- Do not fabricate citations, quotations, metrics, or author profiles.
33
34## Delivery
35Lead with the outcome, name the evidence used for verification, and disclose any real limitation that remains.
KunAgent