dsh-obsidian-math
Run the following command in DeepSeek Harness:
dsh plugin install maple110011/dsh-obsidian-math
Paste the following prompt into your AI chat to install this plugin:
Run dsh plugin install maple110011/dsh-obsidian-math in your terminal to install; the source repository is at https://github.com/maple110011/dsh-obsidian-math
About this plugin
Mathematics is a long-horizon discipline: notation habits, theoretical preferences, half-finished proofs, counterexamples, and emerging ideas all demand continuous collection and refinement into a connected system. Yet a generic chat AI treats every conversation as an isolated Q&A, so accumulated understanding scatters across countless sessions and each new dialogue starts from scratch. dsh-obsidian-math turns DeepSeek Harness into a math-memory assistant embedded in Obsidian's right sidebar, giving the agent cross-session layered memory across five tiers, a notation ledger, and a memo lifecycle so every new session picks up exactly where the last one left off.
The core workflow revolves around unified retrieval and controllable capture. A single BM25 pass ranks results across personal notes and all memory layers simultaneously, with a coverage indicator on each hit and recently polishable memos surfacing by relevance and recency. Capture policy lets you combine three content types (idea, fact, preference) with three autonomy levels (auto, ask, off); auto-tier writes are announced in the closing line, while ask-tier proposals state what, why, and where. The in-Obsidian memory panel supports browsing all five layers, searching, hook usage trends, per-card correctness marking, and archival; feedback links embedded in replies deterministically rewrite cards through a CSRF-protected loopback endpoint, closing the learning loop.
Safety is fail-closed by design: the tool surface is limited to file read, write, search, and four note tools with no shell, no web access, and no subagents. All writes are confined to the vault, using archiving instead of deletion, and the model cannot edit policy or statistics fields. Everything runs locally on 127.0.0.1, memory lives as Markdown inside the vault, and the past-session index never leaves the machine. This plugin is aimed at graduate students, self-directed learners, and instructors who want an AI companion that remembers your notation system, builds on your existing notes, and treats each session as a continuation rather than a fresh start.
Use Cases
- Iteratively discuss theorems and proofs with AI in Obsidian to build a coherent math knowledge system
- Organize personal notes and let AI build on your existing notation conventions for new topics
- Maintain continuous context across days or weeks of study without losing unfinished reasoning or counterexamples
Best For
- Graduate students and advanced mathematics learners
- Self-directed learners who use Obsidian and want AI to remember personal notation and reasoning
- Educators who need an AI partner that retains teaching flow and student thinking patterns
Related Plugins
Memory layer for coding agents that indexes local session history and auto-recalls relevant context before edits, commands, and failures, with no manual search needed.
Traceable, searchable cross-session memory for AI agents that turns conversation knowledge into a typed knowledge graph and recalls relevant subgraphs instead of replaying full history, natively integrated with DeepSeek Harness.
Gives DSH AI cross-session long-term memory, to-do and skill management, plus multi-session orchestration, external AI delegation, and an infinite canvas that grows with you.
dsh-mnemon is a three-tier, pluggable, Agent-driven memory system for DeepSeek Harness, combining Runtime memory, Project Documents, and replaceable Memory Spaces with nine long-term providers.