Search academic literature from remote API and save to workspace papers directory.
Search academic literature from a remote literature search API and save results to the workspace.
# Get PYTHON_PATH from .env
PYTHON_PATH=$(grep "^PYTHON_PATH=" .env | cut -d'=' -f2)
PYTHON_PATH=${PYTHON_PATH:-python3}
python scripts/search.py "social network analysis"
python scripts/search.py "agent-based modeling" --limit 5
python scripts/search.py "deep learning" --year-from 2020 --year-to 2024
This skill requires agentsociety2 to be installed in the Python environment.
Use the PYTHON_PATH from your .env file to ensure the correct Python interpreter is used. See CLAUDE.md for details.
| Parameter | Type | Required | Description |
|---|---|---|---|
| query | string | Yes | Search query (positional argument) |
| --limit | integer | No | Number of articles to return (default: 10) |
| --year-from | integer | No | Filter by publication year (start) |
| --year-to | integer | No | Filter by publication year (end) |
| --workspace | string | No | Workspace path (default: current directory) |
| --multi-query | flag | No | Enable multi-query mode (split complex queries into subtopics) |
papers/literature/papers/literature_index.jsonThe service searches all sources by default when sources is not specified:
| Source | Description | Content Type |
|---|---|---|
| local | RAGFlow local knowledge base | Imported full-text documents |
| arxiv | arXiv preprint platform | Physics, Math, CS, etc. |
| crossref | DOI metadata database | Journal paper metadata |
| openalex | OpenAlex academic graph | 250M+ academic papers |
Local knowledge base results are always displayed first in the response.
papers/
├── literature_index.json # Literature catalog (auto-created/updated)
└── literature/
├── article_1_title.md # Individual article summaries
└── ...
Configure the literature search API in your .env file:
LITERATURE_SEARCH_API_URL=http://localhost:8008/api/search
LITERATURE_SEARCH_API_KEY=lit-your-api-key-here
Or configure via the VSCode extension settings page:
The "验证配置" button performs two checks:
/health endpoint)/api/stats endpoint)Successful validation will show available data sources.
Each article contains:
title: Article titleauthors: Author listabstract: Article abstractyear: Publication yearjournal: Journal namedoi: DOI identifierurl: Original linkscore: Relevance scoresource: Data source (local/arxiv/crossref/openalex)python scripts/search.py "machine learning" --year-from 2020 --year-to 2024
python scripts/search.py "neural networks" --limit 5
Enable multi-query mode to automatically split complex queries into subtopics for broader coverage:
python scripts/search.py "social norms and cooperation in game theory" --multi-query
This is useful for complex topics that span multiple research areas. The system will: