pantheon-ai/research
Research toolkit for triaging academic papers and GitHub projects. Triage papers and tools, reproduce benchmark claims, search Google Scholar, Semantic Scholar, PubMed, or Sci-Hub, and extract structured data from scientific PDFs.
92
92%
Does it follow best practices?
Impact
Pending
No eval scenarios have been run
Critical
Do not install without reviewing
classification-guide.mdtriage-tool/references/
Tool Classification Guide
Use this guide when classifying a tool during the Classify step of the triage workflow. Customise the tag taxonomy and scope criteria to match your research domain.
Tag Taxonomy
| Tag | Meaning |
|---|---|
library | Reusable code library or package |
framework | Opinionated structure for building something |
cli | Operated via command-line interface |
service | Runs as a standalone service or API |
daemon | Runs as a background process |
mcp-server | Exposes capabilities via MCP protocol |
sdk | Language-specific developer SDK |
benchmark | Evaluation or benchmark tool |
dataset | Training or evaluation dataset |
A tool may carry multiple tags. Add domain-specific tags as needed.
Scope Assessment
| Verdict | Criteria |
|---|---|
| In scope | Directly relevant to the research domain tracked in the repo |
| Borderline | Adjacent to the domain but with partial relevance; note the overlap |
| Out of scope | No meaningful connection to the research domain |
For borderline tools: triage and flag the overlap in REVIEWED.md; do not silently exclude or include.
Examples
Clearly in scope
- Replace with a concrete example from your research domain.
Borderline
- Replace with a borderline example — a tool that is adjacent but has partial relevance.
Out of scope
- Replace with an out-of-scope example — log it as out-of-scope in REVIEWED.md.
google-scholar-search
pubmed-search
reproduce-benchmark
sci-data-extractor
sci-hub-search
semantic-scholar-search
triage-paper
triage-tool