AI Agent Hub
Back to plugins
🧠

dsh-plugin-rag

Memory Updated 2026.08.14

Run the following command in DeepSeek Harness:

dsh plugin install YYTbit/dsh-plugin-rag

Paste the following prompt into your AI chat to install this plugin:

Run dsh plugin install YYTbit/dsh-plugin-rag in the DeepSeek Harness terminal to install this plugin; the full source repository is at https://github.com/YYTbit/dsh-plugin-rag .

About this plugin

When an Agent tackles a mid-to-large project, the real bottleneck is rarely reasoning power; it is context. It simply cannot read every source file, config, and doc in one shot, so it starts guessing. Most RAG stacks lean on external vector databases, embedding models, and cloud APIs, which adds friction and makes offline work impractical.

dsh-plugin-rag keeps everything local. It scans your project tree for Markdown, code, config, and documentation files, builds a keyword index using a straightforward TF-IDF scorer, and exposes a CLI that the Agent calls on demand. No embedding models, no network calls, no heavy dependencies. Index statistics are injected into the system prompt so the Agent knows upfront what is available and how large the corpus is.

If your workflow is local-first, offline-capable, and you would rather spend zero infrastructure budget on a simple file-lookup step, this plugin is a pragmatic starting point: one install away, turning your project documentation into a lightweight knowledge base the Agent can query at will.

Use Cases

  • Agent quickly locates relevant docs, code, or config snippets in its working tree
  • Offline context retrieval for the Agent with zero external API calls
  • Lightweight TF-IDF file-level search that replaces a vector database with no infrastructure

Best For

  • Local-first, offline-focused DeepSeek Harness users
  • Solo developers who prefer not to stand up a vector DB for simple file lookup
  • Agent workflow builders who need frequent cross-doc context retrieval across projects