DSH Math Notes Assistant (dsh-obsidian-math)

English · 简体中文

A long-term math-memory agent for DeepSeek Harness that lives inside Obsidian as a right-sidebar chat panel.

A two-component repository:

1. Obsidian community plugin (id dsh-math-assistant, repo-root manifest.json + main.js): embeds the dsh web UI in the right sidebar, detects and starts the dsh service, bootstraps the dsh-side configuration and vault templates on first run, and hosts the memory panel, capture-policy settings, and deterministic maintenance.
2. dsh plugin (npm package dsh-obsidian-math, dsh/): installs the same obsidian agent preset / profile and vault templates into $DSH_HOME.

Both write identical, idempotent configuration. Installing the Obsidian plugin alone is enough; dsh/install.mjs covers the pure-CLI workflow.

Why

Mathematics learning is long-horizon accumulation: notation habits, theoretical preferences, half-finished proofs, techniques, counterexamples, and ideas all need continuous collection and polishing into a connected system. Generic chat AI treats every conversation as isolated Q&A. This plugin gives the agent cross-session layered memory (five layers + a notation ledger + memo lifecycle), unified retrieval (notes + memory in one search), and a standing protocol (AGENTS.md) so each new session starts where the last one ended.

Features

Retrieval (v3: unified entry, coarse-filter + careful-read)

• note_recall unified search: one BM25-ranked pass over user notes AND all memory layers (hook-weighted cards, memos, topics, theorem/episode indexes); unicode-dash normalization and CJK char containment bridge word-form gaps; hits carry a coverage indicator (query-token coverage; <0.35 marks a lexical-coincidence weak signal).
• Read-verify protocol: distilled query (challenge + candidate techniques) → read the top 2-3 hits in full and judge each → on empty/weak results reformulate once → then admit "not in the vault" instead of fabricating; bounded at 2 recalls and 3 full reads per turn.
• Navigation-only injection: the system prompt carries only the navigation layers (profile/notation/topics/records/templates/episodes); content is pulled on demand; the injected section is bounded (≤9000 chars).
• Supporting tools: note_search (user-note tag filter), note_links (backlinks / link-following), note_create (refuses to overwrite).

Memory (five layers + maintenance loop)

• Five layers: profile (semantic) / topics (navigation) / records (typed atomic cards with retrieval hook: blocks and verification levels ✅⚖️❓) / episodes (raw evidence, append-only) / inbox (idea memos, inbox→polishing→done).
• Notation system: memory/notation.md with adopted/candidates/rejected tables and a revision history — collect → unify → maintain; the agent proposes unifications when your notation drifts (observes first when you have no stable habit yet).
• Daily audit: deterministic scan for strong/weak/unused/duplicate-candidate/unverified cards plus structural checks (missing source / broken links / missing index rows); recall hits sync back into uses/success_rate.
• Memo reminders: stale (inbox>7d, polishing>3d) or currently-relevant memos surface for polishing, ranked by relevance × recency.
• Capture policy tiers: idea/fact/preference × auto/ask/off — pick them in the plugin settings (dropdowns write back to capture-policy.md), or edit via the memory panel / the file directly; auto-tier writes are announced in the closing line, ask-tier proposals state what/why/where.
• Cross-session context: past dsh sessions (zstd JSONL) distilled into bounded Q&A cues, vault-filtered and excluding the live session.

Control surface (Obsidian side)

• Memory panel: browse all five layers, search, hook stats with 📈 usage trends, per-card ✅/❌/supersede/archive, audit report; edit-and-save in the panel (mtime conflict guard); capture policy editable on the settings page with effect descriptions.
• Feedback loop: [✅ 这条对] [❌ 这条错] links in replies deterministically rewrite cards through the loopback /feedback endpoint (CSRF-token protected); note references are clickable and jump into Obsidian (/open).
• Reply-quality protocol: intuition before formalism, anchoring new material to your existing notes, difficulty adaptation, Socratic correction, low-frequency check questions.
• Skins & background opacity (aesthetics only, no agent tools added): the obsidian profile mounts the dsh-web-ui skin center (skin picker + background-opacity control) together with its card host web-ui-settings (the "Web UI plugins" group card on the settings page — pure UI, no tools); every other dsh-web-ui family feature stays unmounted (task board / SSH / aionui panel / git-graph / pet / live-stats, etc.) to keep the minimal tool surface. Entry point: settings → Plugins → Web UI plugins → Skins in the embedded UI. The skin choice is shared with the main web profile (one global setting).

Safety (fail-closed)

• Tool surface: file read/write/search + four note tools + ask_user; no shell, no web, no subagents, no delete tools. Of the dsh-web-ui ecosystem only the skin center and its card host web-ui-settings are kept (both add no agent tools; aesthetics); every functional plugin is dropped.
• Writes confined to the vault (workspace-write); interactive escalation prompts disabled (approval: never); DSH_PERMISSION_MODE=danger-full-access only re-enables escalation prompts, the sandbox itself stays workspace-write.
• All memory lives as markdown inside the vault; archiving instead of deleting; the model may not edit policy or statistics fields.

Requirements

• Obsidian desktop; Node.js ≥ 22.5; DeepSeek Harness (npm global @deepseek-ai/dsh); a configured DeepSeek model.
• Default port 3180 (coexists with the regular dsh web on 3080; configurable in settings).

Install

A (recommended): Obsidian → Settings → Community plugins → search DSH Math Notes Assistant, install and enable; or copy main.js/manifest.json/styles.css from a release into <vault>/.obsidian/plugins/dsh-math-assistant/. First run auto-detects dsh, initializes preset/profile/templates, and starts the service.

B (CLI):

npm install -g dsh-obsidian-math
dsh-obsidian-math install --vault "D:\\Obsidian笔记数据库"
dsh --profile obsidian --port 3180 --patch "$DSH_HOME/profiles/obsidian/obsidian.patch.yml"

Plugin settings: port, dsh install dir, DSH_HOME, auto-start, auto-init, auto-archive (>90-day episodes), ribbon button, keep-alive on close, and the capture-policy dropdowns.

Vault layout

vault/
  AGENTS.md                       working protocol (auto-loaded)
  .deepseek/
    memory/profile.md             semantic layer (profile)
    memory/notation.md            notation ledger (collect → unify → maintain)
    memory/topics/                navigation layer
    memory/records/               typed atomic cards (+ hook blocks)
    memory/theorems/              personal theorem index (Matlas-style)
    memory/templates/             problem-template ↔ theorem graph
    memory/episodes/              raw evidence (append-only + archive/)
    inbox/                        idea memos
    capture-policy.md             capture policy (user-maintained)
    cache/                        machine-generated caches (do not edit)

Development & quality

npm test          # syntax + 63 zero-token regression checks + installer e2e (drift detection)
npm run qa        # engine probe: 12 ground-truth recall assertions on the real vault (zero tokens)
npm run qa:e2e    # real-session end-to-end acceptance (spends real tokens; reports API-level usage)
node scripts/build-obsidian.mjs   # rebuild main.js (required after shared-file changes)
node scripts/deploy-local.mjs     # one-shot local deployment

• Repository structure: ARCHITECTURE.md — directory responsibilities, the two-component data flow, the memory↔retrieval boundary, and the feature checklist.
• Memory knowledge base: docs/memory/ — design (implementation spec), retrieval-v3 (retrieval proposal), testing (QA methodology), assessment, references (paper notes), changelog, handoff.
• Acceptance record: engine probe 12/12; real-session E2E 4/4 (including the no-answer honesty and reformulate-retry behaviors); the cost-benchmark question (170K tokens pre-system) now measures ≈25K billed tokens (68% of the prompt served from cache).
• Version: 0.5.1 (prototype stage; the memory architecture has no long-term field testing yet and will keep evolving).

Privacy & safety

Everything runs locally: the service binds 127.0.0.1, memory is markdown inside the vault, and the past-session index never leaves the machine.

License

MIT