AI Agent Hub
Back to plugins
🧰

dsh-wiki-bridge

Web Tools Updated 2026.08.25

Run the following command in DeepSeek Harness:

dsh plugin install xiaomao49/dsh-wiki-bridge

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

Run dsh plugin install xiaomao49/dsh-wiki-bridge in your terminal to install; the source repository is at https://github.com/xiaomao49/dsh-wiki-bridge

About this plugin

After hundreds of turns with an AI assistant, the most useful conclusions are buried in chat history while your Obsidian vault sits on disk, completely invisible to the model. Archiving is manual, retrieval depends on memory, and context is copy-pasted by hand. dsh-wiki-bridge closes that gap: it plugs DeepSeek Harness directly into your Obsidian vault so conversations both read and write your notes, turning the _约定.md style of filing rules into an executable workflow.

The plugin ships five host tools. wiki_search runs relevance-ranked full-text search across the vault. wiki_read pulls a single note in full Markdown. wiki_write enforces auto-generated frontmatter, a tag whitelist, a duplicate-check prompt, and a default inbox destination before anything hits disk. wiki_check audits the vault for broken links, missing metadata, and oversized lines. wiki_archive walks the inbox through categorization, frontmatter refresh, and homepage registration in one pass. An automatic lexicon watches every pre-step: when the conversation mentions a vault title, tag, or keyword, the top three matching note summaries are injected as context with weighted scoring (title 3, tag 2, body 1). Chit-chat that hits nothing costs zero, and credentials are redacted before any write.

This is for people who already run DeepSeek Harness for daily conversations and keep a personal knowledge base in Obsidian. You do not need to restructure your notes or patch the DSH installation. Install the plugin, point it at your vault path, and your AI sessions start working alongside the notes you already have.

Use Cases

  • Auto-inject relevant note summaries when the conversation matches vault topics, eliminating manual search
  • Write discussion conclusions into the Obsidian inbox after a session, with frontmatter and tags filled automatically
  • Periodically audit the Obsidian vault for broken links, missing metadata, and oversized lines

Best For

  • Developers who use DeepSeek Harness for daily conversations
  • Knowledge workers who maintain a personal knowledge base in Obsidian
  • Independent researchers who want to systematically capture AI conversation insights into their notes