DeepSeek Harness Plugins
DeepSeek Harness PluginIndexed

dsh-kb-rag

Local-first academic literature RAG plugin for DeepSeek Harness with hybrid retrieval, section-aware chunking, and DOI citations.

VisionUI & ProductivityDeveloper ToolsBrowser & WebSkills & Workflows
11GitHub Stars1ForksUpdated2026-09-16

Install

$ dsh plugin --profile web add dsh-kb-rag

Plugin Overview & Capabilities

AI-assisted organization based on the public repository snapshot. The content must be grounded in source evidence and does not replace compatibility or security verification.

source-grounded
dsh-kb-rag is a local-first knowledge base RAG plugin for DeepSeek Harness tailored for research literature and Zotero libraries. It converts local PDFs into a structured SQLite knowledge base using section recognition, BM25 keyword matching, bge-small vector cosine similarity, and bge-reranker reranking. It exposes 9 model tools to enable hybrid search, cited question-answering with DOI links, incremental deduplication, and scope-restricted query answering without sending data to external indexing APIs.

Key Capabilities

  • Provides 9 model tools including kb_ingest, kb_zotero, kb_search, kb_rag, kb_scope, kb_dedup, kb_clear, kb_stats, and kb_fetch
  • Performs hybrid retrieval using BM25 and bge-small vector embeddings fused with RRF and section-aware weights
  • Applies Cross-Encoder reranking via bge-reranker-base for precise context extraction
  • Generates exact clickable citations with DOI links and file locations
  • Supports incremental SHA256 deduplication and Zotero library migration
  • Runs entirely locally with zero API cost and resident daemon architecture for sub-second hot queries

Useful For

  • Academic paper search and exact citation retrieval across local PDF and Zotero collections
  • Strict-mode literature Q&A grounded solely in indexed research papers without hallucinated extrapolations
  • Direct PDF fetching and indexing using DOI or arXiv identifiers inside DeepSeek Harness sessions

Who It Fits

  • Graduate students, PhD candidates, and academic researchers managing local paper libraries
  • DeepSeek Harness users seeking privacy-preserving, zero-cloud-cost document retrieval

Documented Limitations

  • Requires local Python dependencies and PyMuPDF/embedding models to be installed and loaded in memory
  • Metadata publication year extraction may occasionally misidentify years when PDF metadata headers are missing
  • In-memory BM25 scan is optimized for up to tens of thousands of chunks rather than massive enterprise datasets

DSH Compatibility

Version-specific runtime evidence collected by DSH Plugin. A missing result means we have not tested that combination yet.

Not tested yetNo runtime compatibility tests have been published yet.

Security Signals

Objective signals discovered from package metadata and source inspection. These are not a guarantee that a plugin is safe.

No automated security signals have been published yet.

Source & Registry Notes

Public provenance, Registry classification, and the latest source check for this entry, kept separate from runtime verification.

Source repository
Breeze136/dsh-kb-rag
Registry source
GitHub · dsh-plugin topic
Registry classification
Plugin
Source checked
4cf8aba · 2026-09-18

This project is independently indexed from public source information. DSH Plugin is not affiliated with DeepSeek or the plugin author. Always check the author repository before installation.

Repository Activity

GitHub stars
11GitHub stars
Forks
1
Open issues
2
Last commit
2026-09-16
Last release
2026-09-14
For maintainers

Maintaining this plugin?

This listing is generated from public repository data. If you maintain this project, you can review the information and share this listing with your users if you find it useful.

Add to README
Listed on DSHPlugin.app