Systematic academic paper discovery and analysis with full citation traceability
▛▀▖ ▌ ▞▀▖ ▗ ▙▄▘▞▀▖▞▀▘▞▀▖▝▀▖▙▀▖▞▀▖▛▀▖ ▌ ▞▀▖▛▚▀▖▛▀▖▝▀▖▛▀▖▄ ▞▀▖▛▀▖ ▌▚ ▛▀ ▝▀▖▛▀ ▞▀▌▌ ▌ ▖▌ ▌ ▌ ▖▌ ▌▌▐ ▌▙▄▘▞▀▌▌ ▌▐ ▌ ▌▌ ▌ ▘ ▘▝▀▘▀▀ ▝▀▘▝▀▘▘ ▝▀ ▘ ▘ ▝▀ ▝▀ ▘▝ ▘▌ ▝▀▘▘ ▘▀▘▝▀ ▘ ▘
Search • Analyze • Cite
Search across arXiv, Semantic Scholar, and OpenAlex → Analyze citation networks → Generate traceable reports with every claim cited.
Perfect for:
- Literature reviews
- Survey papers
- State-of-the-art research
- Finding influential papers
- Tracking research trends
% claude
> /plugin marketplace add modaribrahim/A-Claude-Code-plugin-for-systematic-academic-paper-discovery-and-analysis
⎿ Successfully added marketplace: research-companion
> /plugin install research-companion@research-companion
⎿ ✓ Installed research-companion. Restart Claude Code.Simply ask Claude to find papers:
"Do a comprehensive search on quantum machine learning"
Claude will ask 4 questions, then search, filter, and present results.
Then analyze:
"Analyze that session"
You'll get a detailed report with:
- Executive summary (cited)
- Influential papers (PageRank)
- Temporal trends
- Research gaps
- Full paper list (Appendix A)
- arXiv - Latest preprints (CS/physics/math)
- Semantic Scholar - Comprehensive metadata, citations
- OpenAlex - Cross-disciplinary coverage
- Quick - 20 papers in 10 seconds
- Comprehensive - 500+ papers with citation expansion
- PageRank (influential papers)
- Betweenness centrality (bridge papers)
- Community detection (research clusters)
- Citation distributions
- Temporal trends (year-by-year)
- Venue analysis
- Author productivity
- Every claim cites specific papers
- Numbered references
- Direct links to papers
- Each search = unique folder (no data loss)
- Refine or extend searches
- Re-analyze anytime
┌─────────────────┐
│ Ask for papers │
└────────┬────────┘
│
▼
┌─────────────────────────────────────────┐
│ 1. Search (Multi-source) │
│ ├→ arXiv + Semantic Scholar + OpenAlex
│ ├→ Filter by citations │
│ ├→ Remove duplicates │
│ └→ Expand citations (optional) │
└────────┬────────────────────────────────┘
│
▼
┌─────────────────────────────────────────┐
│ 2. Accept / Refine / Extend │
└────────┬────────────────────────────────┘
│
▼
┌─────────────────────────────────────────┐
│ 3. Analyze │
│ ├→ Citation network (PageRank) │
│ ├→ Temporal trends │
│ ├→ Statistical analysis │
│ └→ LLM strategic reading │
└────────┬────────────────────────────────┘
│
▼
Traceable Report
See docs/architecture.md for detailed system design.
- Claude Code installed
- Python 3.9+
- uv (package manager)
curl -LsSf https://astral.sh/uv/install.sh | sh
# or: pip install uvclaude
> /plugin marketplace add modaribrahim/A-Claude-Code-plugin-for-systematic-academic-paper-discovery-and-analysis
> /plugin install research-companion@research-companionDependencies install automatically via uv sync.
If dependencies don't install automatically:
cd skills/searching-ml-papers/tools && uv sync
cd ../analyzing-papers/tools && uv sync"Find papers on transformer architectures in computer vision"
→ Claude asks questions → Searches 3 sources → Returns top 20 papers (filtered by citations) → Session saved for analysis
"Do a comprehensive search on federated learning from 2020-2024"
→ Searches 500+ papers → Citation expansion (finds papers citing top results) → De-duplicates across sources → Saves to session
"Analyze that session"
→ Citation network analysis → Temporal trends → Statistical distributions → Generates 20-page report
"Refine that search to focus on healthcare applications"
→ Updates session with new query → Merges with existing results → No data loss
Located in: skills/searching-ml-papers/tools/artifacts/session_*/
session_20250130_143026_federated_learning/
├── metadata.json # Search parameters
├── search_results.json # Raw results
├── deduplicated.json # Final collection
└── summary.json # Top papers, stats
Generated in: skills/analyzing-papers/tools/artifacts/experiment_*/
Report includes:
- Executive Summary (3-5 key insights with citations)
- Research Landscape (methods, datasets, metrics)
- Citation Network Insights (influential papers, clusters)
- Temporal Trends (evolution over time)
- Statistical Analysis (distributions, correlations)
- Research Gaps (opportunities)
- Appendix A (all papers with details)
Semantic Scholar (higher rate limits):
export SEMANTIC_SCHOLAR_API_KEY="your-key-here"
# Get free key: https://www.semanticscholar.org/product/api#api-keyOpenAlex (polite requests):
export OPENALEX_EMAIL="your-email@example.com"All scripts support --help:
python skills/searching-ml-papers/tools/scripts/multi_search.py --help
python skills/analyzing-papers/tools/scripts/graph_algorithms.py --help# Clone repository
git clone https://github.com/modaribrahim/A-Claude-Code-plugin-for-systematic-academic-paper-discovery-and-analysis.git
cd A-Claude-Code-plugin-for-systematic-academic-paper-discovery-and-analysis
# Install dependencies
cd skills/searching-ml-papers/tools && uv sync
cd ../analyzing-papers/tools && uv syncpip install pre-commit
pre-commit install --install-hooksPre-commit hooks include:
- Ruff (Python linting/formatting)
- Markdownlint (doc quality)
- Gitleaks (secrets detection)
python skills/searching-ml-papers/tools/test_multi_search.py
python skills/analyzing-papers/tools/test_build_network.pySee CONTRIBUTING.md for development guidelines.
- Architecture - System design, data flow, components
- Contributing - Development setup, guidelines
- Changelog - Version history
→ Check skills/searching-ml-papers/tools/artifacts/ directory
→ Run uv sync in skill's tools/ directory
→ Use optional API keys (see above)
MIT License - see LICENSE file
Built with:
- Claude Code
- uv - Fast package manager
- Ruff - Python linter
Data sources:
Made with ❤️ for researchers