@deepseek-ai/dsh-ragflow
RAGFlow knowledge-base retrieval capability for the DeepSeek Harness — Service Definition, HTTP API provider, and model-facing ragflow_retrieve tool
AI 分析
核心用途是为 DSH 引入高精度的本地文档检索能力。适合需要基于海量专业文档、技术手册进行精准问答的团队或个人。需要配置 RAGFlow 服务端点及 Key。
インストール
$
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:staff-os/dsh-ragflowドキュメント
README 全文を読む ↗Configuration
| Config | Default | Description |
|---|---|---|
apiKey | — | RAGFlow API key (prefer apiKeyEnv) |
apiKeyEnv | RAGFLOW_API_KEY | Credential reference name |
baseURL | http://localhost:9380 | RAGFlow API endpoint base |
datasetIds | — | Default dataset IDs to search |
topK | 10 | Provider-level chunk limit |
similarityThreshold | 0.2 | Chunks below this similarity are filtered |
retrieveTopK | 10 | Tool-level chunk upper bound |
retrieveTimeoutMs | 30000 | Cooperative timeout (ms) |
config in cordis.patch.yml is whole-object replacement, not deep merge. Always specify all fields you need.